[2014] HCJAC 19

Lord Justice Clerk

Lord Drummond Young

Lord Clarke

Appeal Nos: XC224/13, XC258/13, XC292/13 and XC385/13



in the Appeals against Sentence by







Act (1): JDM Macara QC (sol adv), NG Allan, Solicitor Advocate; BCKM

Act (2): J Keenan, Solicitor Advocate; Capital Defence (for Keegan Smith, Livingston)

Act (3): SM McCall, Gilchrist; John Pryde & Co

Act (4): A Ogg, Solicitor Advocate; Drummond Miller (for JC Hughes, Glasgow)

Alt: Fairley QC AD; the Crown Agent

28 February 2014

Legislative History
[1] For many years, Parliament has sought to augment the powers of the courts in relation to the range of penalties which might be imposed on convicted persons. These efforts have been designed, in large part, to provide a variety of disposals which might be thought appropriate within the general context of selecting a penalty which reflects societal needs for punishment, deterrence and the protection of the public. Rehabilitation of the offender may play a part in this exercise. In relation to offences meriting a custodial disposal, until 1967 there was no provision for any formal post release supervision. With the introduction of parole in that year (Criminal Justice Act 1967, Part III), social work supervision became a standard condition for all released on licence. It was also possible to add conditions in relation to drink or drugs counselling and residence. The parole system was regarded as successful, judging by the limited number of persons recalled, but its use, in the case of sentences of 5 years or more, came to be restricted to periods of a few months as a result of Government policy in the mid-1980s.

[2] The system was reviewed by the Kincraig Committee, which was established in December 1987 and had, as part of its remit, the examination of the role of social work supervision on release. Their report in February 1989 (Parole and Related Issues in Scotland, Cmnd 598) led to the introduction of the conditional release and parole system in the Prisoners and Criminal Proceedings (Scotland) Act 1993. There were several salient features to the new system. First, as a generality, there would be no supervision of prisoners conditionally released early (all short term prisoners after one-half of their sentences and all long term prisoners after two-thirds). In addition, the possibility of parole would be available for all long term prisoners (those serving 4 years or more) after they had served one-half of their sentences. Parole would also be available, but with no minimum period specified, for prisoners serving mandatory life sentences. The court was given the option of specifying a minimum term in respect of discretionary life prisoners (see 1993 Act, ss 1 and 2). The specification of the minimum "punishment part" became mandatory for all life prisoners following upon amendments to the 1993 Act in the Convention Rights (Compliance) (Scotland) Act 2001 (s 1).

[3] It was the 1993 Act (s 14) which introduced the supervised release order (SRO) in respect of short term prisoners. This innovation (Criminal Procedure (Scotland) Act 1995, s 209) allowed the court to specify a period of compulsory supervision upon release if it considered that:

"it is necessary to do so to protect the public from serious harm on his release" (emphasis added).

Notwithstanding that any supervision, which might include conditions of various types, was to be imposed in respect of persons who, it was anticipated, might cause the public "serious harm", the maximum period of supervision for these prisoners was (and is) 12 months.

[4] The next significant development came 5 years later with the introduction of the extended sentence (1995 Act, s 210A) by the Crime and Disorder Act 1998 (s 86)). This was a United Kingdom statute, in terms of which a system of extended sentences was introduced for England and Wales (s 58) for "sexual or violent offenders". A similar provision was devised for Scotland. The court was given a discretion to impose a period of post release licence in respect of any sex offender convicted on indictment and any person convicted of a violent offence for which the court intended to impose a custodial term of 4 years or more. The extension period could be up to 10 years for a sex offender and 5 years for a violent one (3 years in the sheriff court). For present purposes, the significant features are that the imposition of an extended sentence is discretionary ("may"; s 210A(1)) and it can be selected if the court:

"210A(1)(b) considers that the period (if any) for which the offender would, apart from this section, be subject to a licence would not be adequate for the purpose of protecting the public from serious harm from the offender".

The use of the words "serious harm" is identical to that for a SRO. The practical utility of the extended sentence was (and is) to allow the court to increase the period during which the offender may otherwise have been on parole, and thus subject to supervision, with the prospect of his licence being recalled.

[5] That then is the background against which the MacLean Committee took up the challenge. This Committee, which reported in June 2000, had as its remit:

"... to make proposals for the sentencing disposals for, and the future management and treatment of serious violent and sexual offenders who may present a continuing danger to the public in particular:

... whether the current legislative framework ... provides the courts with an appropriate range of options and affords the general public adequate protection from these offenders..." (Report p 1).

The Committee adopted an offender-focussed approach, whereby the gravity of the offence was only one factor to be taken into account in the assessment of any future risk to the public. They considered that they should concentrate on "those who may present a particularly high risk" (para 1.8). Thus, it was recommended that:

"Special sentencing considerations are necessary for persons convicted on indictment of a violent or sexual offence, or exceptionally another category of crime, whose offence(s) or antecedents or personal characteristics indicate that they are likely to present particularly high risks to the safety of the public. We refer to them henceforward, in the context of this report, as 'high risk offenders'" (Recommendation 1).

[6] The Committee expressed the view that:

"... while for many such offenders [ie those offenders within the Committee's remit] the present range [of custodial disposals] is satisfactory, for a small number of others the current sentencing provisions are deficient since they do not require the courts to impose on exceptional individuals an exceptional sentence which both marks the gravity of what they have done and provides an appropriate level of public protection, having regard to the risk that such individuals pose" (para 5.1, emphasis in original).

On that basis, the Committee proposed that:

"For this latter group of offenders ... [a] new and separate provision requires to be made so that ... they are subject to the control of the State for the remainder of their lives" (ibid).

[7] The Committee recommended (Recommendation 12) that legislative provision should be made "for the lifetime control of serious violent and sexual offenders who present a high and continuing risk to the public"; that is, the "order for lifelong restriction" (OLR). Although the Committee recommended that only the High Court could impose an OLR, they did consider that such imposition could be in respect of a relatively minor offence, when coupled with past behaviour, and thus remits from the Sheriff to the High Court for the purpose of imposing an OLR must have been in contemplation.

[8] The Committee explained that:

"The OLR is intended for cases where the offender presents a serious risk to the public, and the Court should only impose this sentence if it is satisfied that there are reasonable grounds for believing that the offender presents a substantial and continuing risk" (para 6.5).

This test is reflected, albeit not in identical terms, in the Committee's recommendation (Recommendation 17) that an OLR may only be imposed where:

"...the offender presents a substantial and continuing risk to the safety of the public such as requires his lifelong restriction".

Having also observed that OLRs were intended to address "fundamental concerns when dealing with offenders who present a high and enduring risk" (para 4.7) and to meet the needs of "the very high risk" group of offenders concerned (para 9.6), it is clear that the Committee's use of language cannot be relied upon for the formulation of a precise test for the level of risk justifying the imposition of an OLR. Ultimately, notwithstanding the Committee's reference to the "legal test" (Chapter 6(i)) for imposing OLRs, no clear test can be derived from the terms of the Report. The Committee anticipated, however, that, upon adoption of the proposals to introduce OLRs:

"... the extended sentence will be an important bridge between the new disposal and the normal determinate sentence, in cases where the proposed statutory criteria for the new disposal are not met" (para 4.33).

Thus, the extended sentence was not envisaged to be an alternative disposal where the statutory criteria for the imposition of an OLR were met.

[9] The Executive (as it then was) produced a White Paper (Serious Violent and Sexual Offenders, 2001) noting the MacLean Committee's use of terminology such as "high risk" and "high risk offenders" and observing that:

"The [MacLean] Report started by identifying problems with the current system of sentencing and managing high risk offenders, specifically that:

There is uncertainty about the numbers of offenders that should be considered high risk..." (p 3, para 3).

The phrase "high risk" had not featured in the Committee's remit (supra) but the White Paper adopted this terminology and offered the following guidance:

"'Risk', in the context of this Paper, means the risk of personal harm by an offender to members of the public. 'High risk' is used to mean that an offender poses a significant risk of causing personal harm. The terms 'risk' and 'high risk' ... are limited to violent and sexual offending, and such offending behaviour as is manifestly related to violent and sexual offending" (para 10).

Notwithstanding this reference to "high risk" meaning a "significant risk", the terminology is not consistent throughout the paper.

[10] The stated intention of the Executive was to introduce a "fundamental link" between the OLR and assessment according to Risk Management Authority (RMA) guidelines to ensure that sentencing was contingent upon the outcome of a professional risk assessment which would comply with the RMA's standards framework (p 14, para 32). Specifically, it was envisaged (p 16) that:

"37 ... the RMA will promulgate best practice in the management of all offenders who may pose a risk. However, there will be a smaller sub-group of offenders, the RMA's 'client group', in whom the RMA will have a specific interest. These will be high risk offenders who will need concerted life long efforts to minimise and manage their risk."

It would appear, therefore, that it was not in contemplation that all "high risk offenders" would require an OLR; only those requiring life long attention.

[11] The paper introduced OLRs as being "for serious violent and sexual offenders who are a high risk" (p 21, para 1). It was explained (p 23, para 8) that:

"An OLR will be imposed only if the court considers that there are reasonable grounds for believing that the offender presents a substantial and continuing risk to public safety. In reaching a decision, the court will have regard to statutory criteria. These statutory criteria have not yet been finalised."

Illustrative examples then followed.

[12] As originally framed for consideration by the Justice 2 Committee (8th Report, 2002, SP Paper 645), the risk criteria were broadly defined (para 51) to the effect that:

"... the nature of, or the circumstances of the commission of, the offence of which the convicted person has been found guilty either in themselves or as part of a pattern of behaviour are such as to demonstrate that (either or both) -

(a) there is a likelihood that he, if at liberty, will seriously endanger the lives, or physical or psychological well-being, of members of the public at large;

(b) he is indifferent to the consequences, for members of the public at large of the commission of such offences by him and is unlikely, if at liberty, to accord with such standards of behavioural restraint as ordinarily prevail within society."

Principally due to concerns over the interpretation of "likelihood" and the impression that part (b) would significantly extend the scope of the provision beyond that envisaged by the MacLean Committee, the Justice 2 Committee recommended that the criteria be "significantly tightened" (para 55).

[13] The Criminal Justice (Scotland) Bill was introduced (Stage 1 Debate, Official Report of the SP, 18 September 2002) with the Executive's stated intention being "to ensure that high-risk offenders are identified at the time of sentence". This was notwithstanding that the words "high risk" do not appear in either the Bill or the Explanatory Notes (col 10793).

[14] Whilst emphasising that the mandatory nature of the OLR was intended to increase uniformity in sentencing for the purposes of public protection, the Justice Minister said that an OLR would be imposed:

"where, assisted by the findings of a rigorous evidence-based risk assessment, the court concludes that the statutory risk criteria are met and that the offender is indeed high risk" (col 10795).

He went on to clarify that, in drafting the risk criteria:

"... we have tried to strike a balance between ensuring on the one hand that they are sufficient to provide adequate public protection, are workable under law and are not so tightly drawn that high risk offenders slip through the net; and on the other hand that they are not so wide as to catch lower-level risk offenders" (col 10795).

[15] At the Stage 2 debate, an amendment was proposed whereby the making of a Risk Assessment Order (RAO), as a necessary procedural step prior to the imposition of an OLR, would be discretionary (amendment 89). There were various proposals in respect of the alternative limb (b) (supra) of the risk criteria. It was proposed that the making of the OLR itself would also be discretionary rather than, as phrased in the Bill, mandatory (amendment 94).

[16] At Stage 3 (19 February 2003), the Deputy Justice Minister said that the existing provisions would achieve the MacLean Committee's objective of "providing an understandable and workable measure that the court can use to establish whether an offender is high risk" (col 15325). It was suggested that the precise wording of the Committee should be adopted to "narrow and clarify" the risk criteria (col 15326). It was also suggested that an OLR would not be proportionate for "someone who might pose a serious threat at one point, but would not do so for the rest of their life" (col 15327). None of the suggestions in the various amendments was adopted.

[17] The statutory provisions (s 210B et seq of the 1995 Act, as introduced by the Criminal Justice (Scotland) Act 2003, s 1(1)) do not reflect the terms contemplated by either the MacLean Committee or the White Paper. The statutory language is not the same as that used to explain the provisions in the course of the Parliamentary debates.

[18] The structure of the provisions involves first a definition of the circumstances in which a court can commence the OLR procedure by making a RAO. The power to do so rests only with the High Court, when a person has been convicted of a sexual or violent offence, but the precise terms are that:

"210B(2) ... the court ..., if it considers that the risk criteria may be met, shall make ... a 'risk assessment order'...".

It can be seen immediately that there is no discretion left to the court if it considers that the risk criteria may be met.

[19] The risk criteria are that:

"210E ... the nature of, or the circumstances of the commission of, the offence of which the convicted person has been found guilty either in themselves or as part of a pattern of behaviour are such as to demonstrate that there is a likelihood that he, if at liberty, will seriously endanger the lives, or physical or psychological well-being, of members of the public at large".

The threshold then is whether there is a "likelihood" of "serious" endangerment. There is no mention of the level of any risk beyond "likelihood".

[20] The making of a RAO prompts the compilation of a risk assessment report (RAR) by the RMA. A RAR is defined as being "a report as to what risk [the offender] being at liberty presents to the safety of the public at large" (s 210B(3)(a)). It is specifically provided that there "shall be no appeal against" the making of a RAO (s 210B(6)).

[21] The provisions detail certain matters which can be taken into account by the risk assessor when compiling the RAR. These include not only previous convictions but also the existence of allegations of criminal conduct which have not resulted in a conviction (s 210C(1)). The risk assessor is specifically required to state whether, in accordance with guidelines (infra) issued by the RMA, the risk of the offender being at liberty is classified as "high, medium, or low" (s 210C(3)). The offender is able to instruct his own RAR (s 210C(5)). He can object to the terms of the RAR obtained by the court and, if he does so, he can lead evidence in support of his position (s 210C(7)).

[22] The test for the making of an OLR is set out in section 210F. This provides that:

"(1) The High Court ... if it is satisfied, having regard to -

(a) any risk assessment report ...

(c) any evidence given under section 210C(7); and

(d) any other information before it,

that, on a balance of probabilities, the risk criteria are met ... [shall] make an order for lifelong restriction...".

The word "shall" was contained in the original provision, deleted in error by a provision in the Management of Offenders etc (Scotland) Act 2005, and deemed still to exist by Johnstone v HM Advocate 2011 SCCR 470.

RMA Guidelines
[23] The RMA issued "Standards and Guidelines for Risk Assessment". In the first version (April 2006; Appendix B: Glossary of Terms) they defined the categories of risk as follows:

Risk Rating -

High Risk

This offender presents an ongoing risk of committing an offence causing serious harm. The identified scenarios involve pervasive risk and there are few if any protective factors to mitigate that risk. The offender requires long-term risk management, including supervision, and where the offender has the capacity to respond, ongoing treatment.

Risk Rating -

Medium Risk

This offender is capable of causing serious harm, but in the most probable future scenarios there are sufficient protective factors to moderate that risk. The offender evidences the capacity to engage with risk management strategies and may respond to treatment. This offender may become a high risk in the absence of the protective factors identified in this report.

Risk Rating -

Low Risk

This offender may have caused serious harm in the past, but a repeat of such behaviour is not probable. There are likely to co-operate well with risk management strategies and they may respond to treatment. All probable scenarios for this offender have sufficient protective factors to support ongoing desistance from offending.

A "protective factor" is defined as "Any circumstance, event, factor or consideration with the capacity to prevent or reduce future offending or minimise the harm caused". Risk is said to be the "Nature, likelihood, frequency, duration, seriousness and imminence of an offence". "Risk of Serious Harm" is defined as:

"... a risk of harmful behaviour which is life threatening and/or traumatic and from which the victim's recovery, whether physical or psychological, can be expected to be difficult or impossible. That is, sexual or violent behaviour, for example murder, serious assault, rape, all sexual offences against children, all violent robbery, kidnapping, holding hostage, terrorism and fire raising (where there was a clear intent to harm persons). The likelihood of this occurring is just as important as the fact that the person has caused serious harm in the past. They must be regarded as having the potential to inflict such harm again."

The Guidelines state (section 9) that the definitions are:

"essentially a combination of the following parameters:

Capacity to commit an offence causing serious harm;

The pervasiveness of risk factors in the future

The presence or absence of protective factors in the future; and

The need for ongoing risk management including treatment to mitigate risk."

They stress that the category assigned by the assessor does not automatically determine whether the court will make an OLR but that "it is reasonable to expect that an RAR indicating a high risk is much more likely to result in the Court making an OLR than a report indicating a low risk".

[24] In March 2013, the RMA made significant changes to the risk definitions. These considerably expanded the definitions, perhaps in part as a result of concerns about the approach of the courts (infra), as follows:

The nature, seriousness and pattern of this individual's behaviour indicates an enduring propensity to seriously endanger the lives, or physical or psychological well being of the public at large. The individual has problematic, persistent, pervasive characteristics that are relevant to risk and

· not likely to be amenable to change; or

· the potential for change with time and/or treatment is significantly limited.

Without changes in these characteristics the individual will continue to pose a risk of serious harm.


· there are few if any protective factors to counterbalance these characteristics;

· concerted long-term measures are indicated to manage the risk, including restriction, monitoring, supervision, and where the offender has the capacity to respond, treatment;

· the nature of the difficulties with which the individual presents are such that treatment is unlikely to mitigate the need for long term monitoring and supervision; and

· in the absence of such measures, this individual is likely to continue to seriously endanger the lives, or physical well being of the public at large.

The nature, seriousness and pattern of this individual's behaviour indicates a propensity to seriously endanger the lives, physical or psychological well being of the public at large. The individual may have characteristics that are problematic and/or persistent and/or pervasive but

· there is reason to believe that they may be amenable to change or are manageable with appropriate measures;

· there is some evidence of protective factors;

· the individual has the capacity and willingness to engage in appropriate treatment;

· may be sufficiently amenable to supervision; or

· has other characteristics that indicate that measures short of lifelong restriction may be sufficient to minimise the risk of serious harm to others.

The nature, seriousness and pattern of this individual's behaviour suggests a capacity to seriously endanger the lives, physical or psychological well being of the public at large, but there is no apparent long term or persistent motivation or propensity to do so.

The individual may have caused serious harm to others in the past, but

· it is unlikely that they will cause further serious harm;

· there is clear evidence of protective factors which will mitigate such risk;

· they are likely to respond to treatment;

· they are amenable to supervision;

· they do not require long-term restrictions in order to minimise the risk of serious harm of others.

While the individual may have or have had characteristics that are problematic and/or persistent and/or pervasive, they can be adequately addressed by existing or available routine services or measures."

Statistical Material
[25] In terms of a letter dated 13 November 2013 from the RMA to agents for the first appellant, there are said to have been 110 OLRs made since their inception. At the date of the hearing of the appeal, no offender subject to an OLR had been released (although this is not now the case). Further statistical material was provided by the first appellant, which was said to have been drawn from RMA annual reports and which was not contradicted. It is not easy to reconcile this material with the Government's Statistical Bulletin (Crime and Justice Series) information. However, it was to the effect that there had been 143 RAOs made prior to April 2013. Some of these may not have resulted in the imposition of a sentence as at the date of the hearing. The breakdown of the figures on an annual basis was given as follows:


Number of RAOs

Number of OLRs






















Total to 31.3.2013






Total OLRs to date


In relation to the 99 OLRs made prior to April 2013, 60 involved punishment parts of less than 5 years and 26 had expired. A further 16 were due to expire in the current fiscal year.

[26] Putting this into some kind of perspective, according to the Government Statistics (Table 10(a)) in 2012-2013 (2011-2012)), there were about 1,000 (1,350) persons convicted of crimes of violence attracting a determinate custodial sentence, of which some 160 (240) were sentenced to imprisonment for 4 years or more. There were almost 300 (200) persons convicted of sexual crimes attracting a determinate prison sentence, of which about 100 (74) received sentences of 4 years or more. A very rough estimate would suggest that, of offenders convicted of serious sexual or violent crimes (other than murder), between 5 and 7% might have received an OLR. Statistical information produced by the Crown in relation to the number of sexual and violent (non murder) offences prosecuted in the High Court indicated that there were about 500 cases per annum which might attract an OLR, thus suggesting an OLR rate of between 3 and 4% following upon a conviction for a serious crime.

Circumstances of the Appellants

[27] On 7 December 2012, at the High Court at Edinburgh, this appellant pleaded guilty to charges of assault, assault to injury, assault and robbery, theft by housebreaking and attempted theft using a bank card stolen from the house. On 16 April 2013, the sentencing judge imposed an OLR in respect of the four charges of violence with a punishment part of two years and three months.

[28] The circumstances of the four charges relative to the OLR were as follows. On 2 June 2012, DH was sitting at a bus stop in Wallyford when the appellant ran towards him. The appellant started to punch him on the body and struck him on the back four times with a screwdriver before he was able to run off. The appellant was known to Mr H, who told him at the time of the offence "I'm meant to be your mate and you've f... stabbed me".

[29] On 13 June 2012, Mr H again encountered the appellant, this time in Tranent, after reporting him to the police for entering a nearby house. The appellant accused Mr H of "grassing". A short time later he emerged from a nearby house with a kitchen knife. Mr H ran off and shouted to his friend, WM, who tried to calm things down. The appellant swiped at Mr M with the knife and hit his hand. Mr M knocked the appellant to the ground and ran off to catch up with Mr H. The appellant chased both men, still brandishing the knife, and shouting "I'm easy, I'll stab the two of you".

[30] On 25 June 2012, elsewhere in Tranent, the appellant approached LS, with whom he had gone to school. Suddenly, the appellant produced a Stanley knife, pointed it at Mr S and began shouting at him. Mr S struggled with the appellant, who seized him by the neck and repeatedly attempted to stab him on the body.

[31] Finally, on 5 July 2012, the appellant, YM, and C McA were at Ms M's home in Tranent. Ms McA had been invited there under the pretence that she would receive repayment of a debt. Ms M proceeded to assault her, for reasons unknown to Ms McA, by punching her on the face and body. The appellant entered the room with a kitchen knife, which he held to Ms McA's head. The appellant made repeated threats to Ms McA to "cut" her up. He stole Ms McA's mobile phone and house keys, entered her home address and stole her bank card and prescription diazepam tablets. He issued further threats to her through a third party by which he obtained her PIN and attempted to withdraw money from her bank account.

[32] The appellant was almost 30 at the time of sentencing. He had never worked and his life had been one characterised by alcohol and drug misuse, confrontational behaviour and offending. His record stretched back to 1996, initially through the children's hearing system, and culminated in his release from prison on 30 May 2012. He had committed the present first offence within a matter of days of that release. He remained liable to have a section 16 order (1993 Act) made in respect of the unexpired portion of the prior custodial sentence, which was for assaults and drug offences. The appellant's early offending consisted mainly of public order and road traffic offences, but shifting in 2006 to dishonesty. Offences of violence appear throughout his record. He had appeared for sentencing on 28 occasions between 2001 and 2010, having been convicted of various assaults, assaults to injury, assault by the use of a knife and assault and robbery. He had convictions for carrying knives. He had received various fines and short-term periods of custody. Previous community disposals, in the form of a probation order in 2004 and a drug treatment and testing order in 2007, had been unsuccessful. In 2008 he had been sentenced to 2 years imprisonment for assault and robbery and in 2010 he had been given a custodial sentence of almost 4 years for drugs offences.

[33] Following the making of the RAO in January 2013, a RAR was forthcoming; as was a Criminal Justice Social Work Report (CJSWR). The CJSWR identified 12 risk factors, including the appellant's offending history, his anger control, impulsivity, "pro-criminal" attitude and use of weapons. No significant protective factors were identified. The appellant had told the social worker that he would react in the same way if he were faced with similar situations. The report expressed doubts about the appellant's motivation to comply with community supervision. Overall, the appellant demonstrated a propensity for continuing violent behaviour exacerbated by the use of weapons, a high risk of harm to others through physical violence, and a high risk of imminent offending on release.

[34] The RAR considered re-offending to be likely and imminent on release, without significant changes. However, the appellant was assessed as presenting only a "medium" risk to the safety of the public. The nature, seriousness and pattern of the appellant's behaviour indicated a propensity seriously to endanger the public. He was likely to return to a chaotic lifestyle, feeling the need to carry a weapon and resorting to violence either impulsively if threatened or deliberately if he felt it was justified. The risk assessor expressed cautious optimism on the appellant's capacity for change, due to his expressed commitment to change his lifestyle in the community, his wish to address his substance misuse and his desire to be a good father to his 7 year old son. The risk assessor recognised that previous expressions of optimism by professionals had been followed by further offending. The appellant was regarded as unmanageable in the community, but he could develop insight into his risk factors, and become more manageable, with long-term intervention and robust risk management.

[35] The sentencing judge did not share the risk assessor's optimism. He observed that almost everything had been done to stop the appellant from offending, but "regrettably without success".


[36] On 26 October 2012, this appellant pleaded guilty to an attempted murder on 24 July 2012. He had been drinking heavily and behaving aggressively towards his partner following a funeral. The appellant's mother had gone to meet him in a local chip shop, because of concerns about his conduct, in order to walk him the short distance to his home. When he became aware that his partner and her mother were standing outside his home, the appellant walked off. The appellant's mother remained at his home address with his partner and her mother. Shortly afterwards, the three of them were joined by the complainer, who was a close family friend. The group were talking in the garden when the appellant re-appeared.

[37] The complainer kissed the appellant's partner and shook her hand. The appellant approached the complainer and, without any provocation or warning, head-butted him twice on the face, knocking him to the ground. He repeatedly kicked and punched him on the head and body. He sat astride him and continued to punch him repeatedly on the head using both fists. The appellant stood up and began to kick and stamp on the complainer's head. He walked away but turned, and ran back towards the complainer, before kicking, stamping and jumping on his head with both feet.

[38] The complainer sustained multiple fractures to his cheekbones and jaw, severe facial swelling and bruising, a very deep wound on his top lip and the loss of several upper teeth. He continued to suffer pain and discomfort, but it was anticipated that this would resolve.

[39] At the time of the offence, the appellant was 28 years old and lived with his partner and their two young children. He was employed in the building industry. He had 9 previous convictions between 2002 and 2011, notably several for assault, including one assault to the danger of life in 2008 involving kicking and jumping on the head of the victim. This had attracted a sentence of 18 months. In August 2011 he was convicted of possession of an offensive weapon, namely a hammer, and made the subject of a restriction of liberty order (which he breached).

[40] The appellant tendered a report from a consultant forensic psychiatrist. The sentencing judge made a RAO and called for a CJSWR in light of the possibility of an extended sentence. The psychiatrist considered it likely that the appellant had an autistic spectrum disorder, perhaps Asperger's Syndrome, but that the offence did not seem to be so much related to such a disorder as the appellant's extreme intoxication. The CJSWR concluded that the appellant was intellectually and emotionally limited. In view of the escalating seriousness of his violence, it was concluded that he would present an obvious risk to the general public if at liberty. His excessive use of alcohol was a recurring feature of his violence and he would benefit from an extended period of supervision if at liberty.

[41] On 16 April 2013, the sentencing judge imposed an OLR with a punishment part of 3 years and 4 months' imprisonment. He took account of the three reports available to him but noted, in particular, the view in the RAR that the appellant accepted violence as a normal feature of his everyday life. The assessor had identified a number of risk factors for future violence with no accompanying salient protective factors. He assessed the appellant as presenting a "medium" risk to the safety of the public. In this respect, the risk assessor considered that the nature of the risk could be managed by measures short of lifelong restriction. Although the appellant retained the propensity seriously to endanger the public, the assessor considered that he might be amenable to change if a suitably rigorous risk management plan were put in place; specifically treatment within prison for his addictions, anger management and cognitive difficulties coupled with intensive supervision on his release.

[42] The risk assessor had himself been a member of the MacLean Committee and he had felt constrained to make the following comments in his report:

"I take into account what I believe to be the function of the Order for Lifelong Restriction based on both my experience as a member of the MacLean Committee and my experience as an accredited assessor. It is my understanding the OLRs were designed for managing rare and unusual adult offenders."

He quoted the Chief Executive Officer of the RMA as commenting that, because an OLR replaced the discretionary life sentence, the expectation was that it would be used "rarely". It was "an extraordinary sentence intended to deal with the extraordinary offender". The risk assessor concluded that, although the appellant's violence was extreme, his psychological profile could not be regarded as either rare or unusual within the general prison population.

[43] On 2 August 2012, at the High Court in Edinburgh, this appellant pleaded guilty to 5 charges of repeated assaults against 5 different women. All were ex-partners of the appellant and 2 of them had had a child with him; the appellant having assaulted them before, during and after their pregnancies. Each victim had been assaulted on various occasions during the course of their relationships with the appellant. The assaults took place over the course of some 17 years, with some overlapping of the relationships, between January 1994 and March 2011. By the time of sentencing, the appellant was 40 years old.

[44] The first victim, NW, had met the appellant in November 1993, a week before her 18th birthday. The appellant moved in with her shortly thereafter. They lived together until November 1996. The relationship had been volatile from the outset and, after 6 months, became violent. In June 1994, following an argument, the appellant head-butted her on the face, breaking her nose. On another occasion, he banged NW's head onto a kitchen worktop, injuring her eye. In early 1996, when NW was around 7 or 8 months pregnant, the appellant returned home under the influence of alcohol and began to assault her without warning. He seized her by the hair and pulled her into the living room causing her to fall to the floor. He sat astride her, with his full body weight on her stomach, grabbed her by the throat and repeatedly punched her on the head. In April 1996, NW gave birth to a daughter. The relationship ended 6 months later, when NW moved out.

[45] The second victim, SW, had met the appellant in May 2003, when she was 15 years old. She was initially in foster care but later living alone under social work supervision. On at least three occasions, during arguments at her home, the appellant threw SW against a wall and furniture. Ultimately, the relationship ended in August 2003, when SW moved out.

[46] The third victim, LM, had met the appellant when she was 20 years old. Their relationship lasted approximately 10 years, from 2000 to 2010. They had a daughter, who was born in January 2005. From early 2004, LM was regularly physically abused by the appellant. The assaults were less frequent after the couple moved to a different address in 2006, when LM saw the appellant less frequently, but he would often grab her by the hair and strike her on the head and body, occasionally with implements such as remote controls and lighters. On an occasion in November 2008, he dragged her out of bed, breaking two bones in her arm, struck her on the forehead with a wine bottle and punched and head-butted her on the face. On another occasion, he struck her on the face with a metal walking pole in front of their young daughter. In August 2010, the appellant dragged LM across the floor by her hair, prompting his young daughter to try to intervene. In March 2011, LM moved out, but the appellant continued to visit her at her new address. In September 2011, at this address, the appellant seized LM by the face and pinned her down on a sofa, covering her mouth and nose so that she had to struggle to breathe.

[47] Meantime, the fourth victim, LB, had met the appellant in 1996/1997. She began a volatile relationship with him, during which he would abuse her verbally, spit at her, stand nose to nose with her, prod her on the shoulder and pull her by the hair. In December 2006, she became pregnant. The appellant was worried that his partner, LM, would find out. When LB was around 4 or 5 months pregnant, the appellant appeared at her home and accused her of telling people that he was the father of her child. He seized her by the hair and threw her onto the floor. Afterwards, he had asked her to drive him to a friend's house but, during the drive and without warning, he punched her on the eye whilst shouting and screaming that the baby was not his. Their daughter was born in September 2007. In September 2011, LB visited a friend's house at the appellant's request, whereupon he locked her in, shouted and swore at her, slapped her on the face and grabbed her by the neck, restricting her breathing.

[48] The fifth victim, HR, had met the appellant in September 2009. They moved in together within days. Their relationship became violent after about 3 weeks, with assaults occurring on almost a daily basis until October 2011. The appellant variously shouted and swore at HR, punched her on the body, head-butted her on the face, spat at her face and stamped on her leg. On one occasion, he threw HR around a bedroom at a friend's house, sat astride her and pinned down her arms before striking her on the neck with a piece of a mirror which had broken during the incident. He then punched her on the face, breaking a tooth. On another occasion, in late 2010, the appellant seized her by the neck at a friend's house and punched her twice on the face. On yet another occasion in the garden of this house, he punched her on the face and body. On numerous occasions, he grabbed HR around the throat, compressing her neck and restricting her breathing to the point of occasional unconsciousness. On one such occasion, he wrapped a seatbelt around her neck whilst parked at a shopping centre. On another, the appellant pushed her onto her bed and compressed her throat. In September 2011, when visiting the appellant at LM's home, he stood over HR and tried to hold a knife to her throat.

[49] The appellant had 14 previous convictions between 1993 and 2011, including assault, breach of the peace and breach of bail. In light of these, and having regard to the catalogue of serious violence persistently inflicted by the appellant on women throughout his whole adult life, the sentencing judge considered that the risk criteria may be met. A RAO was made and a RAR was produced in 6 December 2012. A CJSWR was also available. After various adjournments granted to enable the appellant to consider the terms of the RAR and the possibility of lodging objections or instructing his own psychologist's report, the case finally called for sentencing on 9 May 2013. No objections were ultimately lodged.

[50] The appellant was assessed as at "high" risk in the RAR, as he was suffering from a dissocial personality disorder co-existing with paranoid, narcissistic and psychopathic personality traits. Any new relationship was likely to result in repeat or escalating physical violence towards female partners. The severity could be high, including death. His five victims were particularly at risk, but their children, male adults and other family members could also be at risk. No protective factors were identified. Notwithstanding the risk assessor's recommendations in relation to future community risk management, the sentencing judge considered that these were not made with any confidence. He determined that, without long-term treatment for his personality disorders, it would not be safe to release the appellant into the community. Accordingly, he considered that the risk criteria were met and an OLR was imposed. The judge fixed the punishment part at 4 years.

[51] On 2 April 2013, at the High Court in Glasgow, this appellant pleaded guilty to a charge of attempted murder. On 19 October 2012, in Glasgow, the appellant assaulted the complainer by repeatedly stabbing him on the body with a knife to his severe injury, permanent disfigurement and impairment, and to the danger of his life. The appellant and the complainer had been in a homosexual relationship. They had been consuming alcohol and cocaine at a friend's house. The appellant shouted that the relationship was over. The complainer was holding a knife and lunged at the appellant, who grabbed the knife. After a struggle, the appellant stabbed the complainer 5 times to the chest, stomach, shoulder and back. The complainer suffered a collapsed lung, injuries to his bowel and bladder, for which he required surgery, including removal of the small bowel. The incident had an horrific impact on the complainer. His life had been profoundly affected. He struggled to sit up, could walk only to the toilet, suffered from pain and his voice had been badly affected.

[52] The appellant was 42 years old at the time of sentencing. He had some 27 previous convictions, including 3 convictions for possessing offensive weapons, various offences of assault and robbery, an assault to severe injury and the danger of life by stabbing and striking with a baseball bat in 2004, and an attempted murder by setting fire to a doormat at a flat occupied by a couple and their 13 year old son in 2007. He had been on licence under a 10 year extended sentence for the latter offence when he committed this attempted murder.

[53] The RAR assessed the appellant as presenting a "medium" risk to the public. The nature, seriousness and pattern of his behaviour indicated a "clear propensity" seriously to endanger the lives of those around him. He had committed the offence despite previously undertaking course work in prison, being subject to social work supervision and attending addiction services. Although he had persistent problematic characteristics, there was no evidence of major psychopathology. The risk assessor considered that the appellant had a willingness to co-operate with supervision and treatment and the capacity to develop protective factors. The appellant was thought to be possibly amenable to change. The risk assessor concluded that measures short of lifelong restriction may be sufficient to minimise the risk of serious harm to others.

[54] The sentencing judge did not, however, share the assessor's view of the appellant's willingness and capacity to engage in appropriate treatment. Notwithstanding that he was categorised as a "medium" risk, which the judge found "very surprising", the appellant satisfied the risk criteria. The judge said that he was "in no doubt" that the appellant was a "dangerously violent man" and that it was "perfectly plain" that, if at liberty, he would seriously endanger the lives or physical or psychological well-being of members of the public. Accordingly, on 11 June 2013, an OLR was imposed with a punishment part of 4 years and 8 months.


[55] It was accepted that there was no ambiguity or absurdity arising from the wording of section 210E of the 1995 Act. Nevertheless, it was relevant to consider the background to the legislation in order to identify the types of offender which it had been intended to cover. The MacLean Committee recommendations, which had been accepted by the Executive, had been designed to apply to high risk offenders (Recommendation 1). These were those offenders posing the "highest risk". There had been a concern before the Justice 2 Committee that the risk criteria had been too broadly drafted and that offenders who were not "high risk" would be drawn into the OLR regime. That may be what the statistics were indicating and it is what had happened to the appellant.

[56] The court ought to substitute an extended sentence with a custodial element of 4 years and 6 months (ie double the punishment part). The sentencing judge had erred in holding that the high test of serious endangerment of the public at large had been met. He had erred in failing to give sufficient weight to the analysis in the RAR. Had he done so, he would have recognised that an extended sentence would have sufficed. Despite the appellant's lengthy and significant record, he had not previously appeared in the High Court or been made the subject of an extended sentence.

[57] The CJSWR had reported that the appellant was at high risk of imminent re-offending and posed a high risk of physical violence, but it had been based only on one interview with the appellant. The RAR on the other hand, which had been compiled after 4 meetings with the appellant and upon the basis of significant other information, had assessed the appellant as being only at "medium risk". In that respect, the alteration in the category of "medium risk" in March 2013 meant that this category now involved a "propensity" (tendency) to cause harm. This had heightened the previous test.

[58] The sentencing judge had expressly rejected the RAR's conclusion in favour of that in the CJSWR. He had applied a counsel of despair when he had said in his sentencing statement that "almost everything has been attempted to stop you committing crime, but regrettably without success". In fact, the RAR had said that there was some evidence of protective factors in the appellant's case, notably his "willingness and potentially the capacity to engage in appropriate treatment". There was "reason to believe that he may be amenable to change". Any process of change was, it was recognised, one which would be "slow and will take time and effort".

[59] The previous sentences imposed upon the appellant had not enabled him to engage in appropriately focused work within a prison setting. Courses had now been identified which he could carry out, notably: (1) a substance abuse course, lasting 6 months, involving the transfer of the appellant from Saughton to Shotts; and (2) a violence prevention programme, which could be undertaken thereafter. In all these circumstances an extended sentence, which could involve the imposition of onerous conditions should the appellant be returned to his community, had been the appropriate sentence to impose.

[60] The appellant adopted the general submissions which had been made on behalf of the first appellant, to the effect that it had been anticipated by Parliament that the legislation was to apply only to those displaying a high and continuous risk to the public at large. Notwithstanding the apparent awareness of Parliament that this anticipation was not reflected in the statutory wording, it was nevertheless intended that OLRs apply only to those posing the "highest risk". The wording could be compared with the similar terms of section 210A, introducing extended sentences. It was envisaged that OLRs would apply to a different category of offenders, with the pre-existing extended sentence regime acting as a "bridge" between the ordinary offender and those requiring OLRs.

[61] The appropriate sentence for the appellant had been an extended one, with the custodial element being 6 years and 8 months (ie double the punishment part). It was accepted that the offence committed by the appellant had been a serious one and required the imposition of a significant custodial sentence, with post-release supervision. However the appellant's record, whilst including analogous offending, related largely to summary matters. The present offence, although one involving serious violence, had not involved any weapons and fortunately it was anticipated that the complainer would fully recover from his injuries.

[62] The RAR had stated that the appellant presented a medium risk of future violence, in the sense that he had a "propensity" seriously to endanger the lives of the public at large. Nevertheless, it was reported that there were reasons to consider that he may be amenable to change if "a suitably rigorous risk management plan can be put in place". The RAR had specifically stated that the nature of the risk posed was one capable of being managed by measures short of an OLR.

[63] There had been no proper or adequate explanation given by the sentencing judge as to why the very clear conclusions of the full and detailed RAR, which had taken several months to prepare, should have been departed from.

[64] Once again it was accepted that the appropriate sentence was one of an extended term with the custodial element being double the punishment part (ie 8 years).

[65] The appellant first argued his ground of appeal 2, which was directed towards the making of the RAO itself, under section 210B of the 1995 Act. It was understood by the appellant that section 210B(6) provided that there was no appeal against a RAO. Nevertheless the court had, at an earlier procedural hearing and for reasons which were not now entirely clear, asked to be addressed on this point. The "sequential difficulty" that the prohibition on appeal created was appreciated. It was still argued, however, that an error had been made when the sentencing judge had determined that the risk criteria "may" have been met in terms of section 210B(2). In deciding this, the judge had failed to take account of, or attach appropriate weight to, the gap in the appellant's offending between 1996 and 2001. He had also failed to attach sufficient weight to the absence of a previous sentence of custody and the summary nature of the appellant's record. He had erred in having regard to the offences themselves, without considering whether, in the whole circumstances, the risk criteria may have been met. His approach in proceeding solely on the basis of the offences and then considering whether any other factor undermined his conclusion had also been erroneous. The analysis had to take place in the context of the offender's whole history and not simply isolated to the charges of which he had been convicted. In taking this approach the judge had effectively placed an onus on the appellant to demonstrate that his offending did not indicate that the risk criteria may be met.

[66] The second error advanced under this ground was that the judge had made the RAO before obtaining a CJSWR. In that respect, the appellant's case differed from those of the first and second appellants. The judge ought first to have considered whether a determinate sentence would suffice. It was only if there was a risk to the public, which could not be met by a determinate sentence, that a judge should move on to consider an extended sentence. Such a sentence was intended as a bridge between an ordinary determinate sentence and an OLR. Only when an extended sentence was not capable of meeting the aim of protecting the public, because of the nature of the risk involved, should a judge consider whether the criteria for an OLR were met. It was all the more important to take this incremental approach where there was no judicial discretion if the criteria were met.

[67] The need to ensure that there was a genuine correlation between the aim of detention and the detention itself became all the more compelling (James v United Kingdom (2013) 56 EHRR 12, para 204). With an extended sentence, the prisoner had to be released after a specified period in custody. He would thereafter be subject to supervision and defined conditions, including curfew, participation in rehabilitation programmes and restrictions on lifestyle and relationships. If the offender breached these conditions, he could be returned to custody. The emphasis was, therefore, on protecting the public by controlling the offender whilst at liberty. An extended sentence was appropriate in cases where it was envisaged that the need for public protection would either come to an end as a result of, for example, rehabilitation or where it could not be justified beyond a certain period because of its disproportionate nature. With an OLR, after the defined minimum custody period, the offender was not released until such time as his level of risk had diminished to the extent that it was manageable in the community. The emphasis was on control in the custodial environment. In order to justify an OLR, it had to be established that the offender presented such a danger that he required lifelong control (whether through custody or monitoring on life licence).

[68] A sentencing judge therefore had first to rule out the adequacy of a finite sentence (Robertson v HM Advocate 2004 SCCR 180, at para [32]). He therefore ought to have obtained a CJSWR before obtaining a RAR. In this case, once the CJSWR had been obtained, it made reference to a management plan. If the judge had waited for such a report before seeking a RAR, its terms could have had a real impact on whether the judge considered that the risk criteria may be met. In this regard, although Article 5 proportionality was not pressed, the judge had to assume that the Scottish Ministers would put in place suitable rehabilitation programmes and treatments for personality disorders of the type that the appellant had. He was a candidate for domestic violence programmes, although these were not available in prison as distinct from the community.

[69] Grounds of appeal 1 and 4 were taken together. The first point was that the making of an OLR had been excessive and inappropriate and the second was that the sentencing judge had erred in his interpretation of the RAR and, in particular, that the risk criteria had been met. The first error was in not recognising that the RAR had made an assessment of current risk and not future risk as required by statute. The judge had interpreted section 210C(3) as requiring the risk assessor to express an opinion on the level of risk currently presented by the offender, if at liberty. The RAR had expressed an opinion that the appellant's risk was a high level one, but there was no discussion about future, as distinct from present, risk. The judge ought to have directed himself, not to section 210C(3), but to the terms of 210E, where it is said that there must be a likelihood that the offender, if at liberty, "will" seriously endanger lives; ie in the future (cf Robertson v HM Advocate (supra), at para [27]). Because the Order was one for lifelong control, the judge ought to be able to say that the risk was indefinite. That meant considering what would happen in the future (cf R v Smith (Nicholas) [2011] 1 WLR 1795, Lord Phillips at paras 14-15; R (on the application of Sturnham) v The Parole Board [2013] 2 AC 254, Lord Mance at paras 32-38, referring to Secretary of State for Justice v James [2010] 1 AC 553, Lord Judge at paras 102, 103 and 108 and R v Pedley [2009] 1 WLR 2517, Hughes LJ at paras 16, 20 and 21).

[70] The second error of the sentencing judge had been to read a lack of confidence into the risk assessor's recommendations about supervision, monitoring and victim safety. The test was not whether the measures would prevent future offending, but whether the risk would be reduced to a tolerable level. In determining otherwise, the judge had set too high a benchmark. There had been no indication that supervision with appropriate conditions would not be successful. The judge had not addressed this question in any detail. The appellant may have breached probation in the past, but that was not for failing to co-operate with his supervision. He had attended his appointments in a sober state. None of this was redolent of an inability to engage with the conditions of probation.

[71] The sentencing judge's third error had been his failure to notice that the RAR had not taken account of the conviction free period between 1996 and 2001 and the lack of any previous sentences of custody. The fourth error had been his failure to determine the impact of the sexual violence allegations, which the risk assessor had taken into account despite their unproved nature.

[72] There had then been no articulation by the sentencing judge as to why he had ruled out a finite sentence and why any risk could not have been met by an extended sentence (cf Ross v HM Advocate [2013] HCJAC 111, paras [6]-[9] and [11]).

[73] It was accepted that an extended sentence was appropriate and that 9 years 4 months (double the punishment part) would be appropriate as the custodial element. This would run consecutively to a section 16 (of the 1993 Act) return period of in excess of 6 years.

[74] OLRs dealt with situations in which a determinate sentence, whether extended or not, was not appropriate (eg Baillie v HM Advocate 1993 SCCR 1084). They were intended for that narrow group of offenders whose offences, antecedents or personal characteristics indicated that they presented particularly high levels of risk to public safety (MacLean Committee Report, para 1.8). An OLR was a sentence of last resort (McFadyen v HM Advocate 2011 SCCR 759). The MacLean Committee had distinguished cases, where violence had been perpetrated by persons who had ingested excessive amounts of alcohol or drugs, from those involving persons who had been identified as part of a small group of "high risk offenders" (para 1.15). Conviction of a serious offence was not enough. There needed to be reasonable grounds for believing that the offender presented a "substantial and continuing risk" (Recommendation 17).

[75] Prior to calling for a RAR, the sentencing judge should have called for a CJSWR, which would have enabled consideration to be given to the imposition of an extended sentence (see eg Hamilton v HM Advocate 2005 SCCR 316). Such a report was necessary before imposing an extended sentence (1995 Act, s 210A(4)).

[76] Section 210E required that the offence, or pattern of offending, demonstrated a likelihood that the offender would seriously endanger the lives etc. of the public. "Likelihood" meant more than just probability, but rather a "real and substantial risk" which need not be more probable than not, but which was more than a bare possibility (R v Piri [1987] 1 NZLR 66; see also "no likelihood" and "no real risk" in Ludriecus v Thomson 2008 SCCR 996, Lord Reed at para [22]). In the specific context of OLRs, the court had previously adopted the Oxford English definition of likelihood as "something that is likely; a probability" (Liddell v HM Advocate 2013 SCCR 541), but the court had not had the advantage of a full citation of authority. Likelihood meant more than a mere probability but a real risk. Such an interpretation would be in keeping with the MacLean Committee's recommendations that an OLR should be applicable only to a narrow group of offenders.

[77] In LBM v HM Advocate [2011] HCJAC 96, the court appeared to hold that, if someone had been assessed as at medium risk in a RAR, he would almost automatically fit the risk criteria. However, that case had been decided prior to the change in the categorisation of "medium risk" in March 2013, whereby the offending behaviour had to indicate "a propensity to seriously endanger" the lives of the public etc. The new definition did not concur with the approach in LBM v HM Advocate (supra) to the effect that an offender, at medium risk, required an OLR.

[78] The RAR had assessed the appellant as a medium risk to the public. Although the offence had occurred when the appellant was already subject to an extended sentence, the risk assessor had considered that his propensities might be amenable to change and might be manageable with appropriate measures. There was a potential for protective factors to develop and the appellant had a capacity and willingness to engage with appropriate treatment. There was an absence of any major psychopathology in the appellant's personality. The risk assessor considered that, if the appellant could gain a greater insight into the problems which substance abuse had caused him and if he returned to the community via an open prison and demonstrated that he could be trusted and maintained a more productive and stable lifestyle, his level of risk would be reduced. The appellant required to achieve the same level of control which he was capable of exhibiting in the prison setting.

[79] The sentencing judge had explained that he considered that the appellant had satisfied the risk criteria, having regard to his two previous High Court convictions for violence, including one for attempted murder upon which he was on licence at the time of the present offence. The judge did not share the optimistic view of the risk assessor in relation to the appellant's capacity and willingness to reform. However, he ought to have attached greater weight to the risk assessor's views. Although the appellant had previous convictions for violence, at least one of those had occurred against a background of drink, drugs and a volatile relationship. The present offence had occurred in similar circumstances. Insufficient weight had been attached to the specific and particular circumstances of this offence, which did not indicate that the appellant was a danger to the public at large.

[80] The appellant, whilst at liberty, had been the carer for his mother and was described as a model prisoner. He had a supportive family. He had been at liberty for some 10 months before the offence had occurred and, prior to that, had co-operated with the supervision conditions. In all these circumstances, the "likelihood" test and the risk criteria were not met and an OLR was both inappropriate and excessive.

[81] The purpose of an OLR was the protection of the public. The appellants' submissions contained, as a consistent theme, the notion that an extended sentence would be adequate in each case. Having regard to the terms of the RARs, that was demonstrably incorrect. Section 2(4) of the 1993 Act provided that, where an OLR had been pronounced, the prisoner would only be released by the Parole Board if "the Board is satisfied that it is no longer necessary for the protection of the public that the prisoner should be confined". This was to be contrasted with the extended sentence regime under section 1(2) whereby, after two-thirds of the custodial element of the sentence, there was an obligation to release the prisoner. The effect was therefore to make release mandatory under that regime (see Robertson v HM Advocate (supra) at para [35]). There was no power to deal with the offender differently if, during the custodial regime, he declined to attend any courses for his rehabilitation or it became clear that these had not worked. All that could be done was to impose certain conditions on his release.

[82] Taking the first appellant as an example, although the same considerations applied in the other appeals, he required to address certain matters before the protective factors, as defined, would come into play. He was described as unmanageable in the community and the sentencing judge had been entitled to conclude that the statutory criteria had been met. That much was apparent. The only factor mitigating against this was that he may be amenable to change. It was not clear that the cautious optimism of the risk assessor would be realised. An extended sentence was not a way of managing risk. Given the powers of the Parole Board, it was easy to understand why the judge would have preferred an OLR over an extended sentence. The theme throughout all of the appeals, that an extended sentence was adequate to manage risk, was therefore misconceived.

[83] The third appellant's case stood out as involving a particularly high risk, with the picture painted being a very bleak one of an offender with a psychopathic and narcissistic personality. There were no protective factors identified in the RAR in relation to him, and nothing indicated which would be mitigatory of the risk which he posed. It would be wholly inappropriate to conclude that an extended sentence would manage the risk in his case.

[84] In order to quash the orders, the court would have to be satisfied that the sentencing judges had fallen into error. Once a judge was satisfied that the risk criteria were met, he was bound to make an OLR.

(a) Appeals against a RAO (third appellant only)
[85] In terms of section 210B(6) of the 1995 Act, it is stated in clear terms that there is no appeal available in respect of the making of, or a refusal to make, a RAO. The rationale behind that is straightforward. The RAO may, or may not, result in an OLR. It is simply a necessary procedural step on the way towards determining whether the risk criteria are met. It involves the court obtaining certain statutorily prescribed information but, that apart, the process is little different from any adjournment in order to achieve clarity on a particular matter affecting sentence. Even if the statute had not expressly excluded an appeal, it is highly doubtful whether a party could competently have challenged the making (or refusal) of a RAO. The only conceivable method of doing so would have been by Bill of Advocation. This would not be competent other than in quite exceptional circumstances (Muir v Hart (1912) 6 Adam 601; Durant v Lockhart 1985 SCCR 72), which are not present in the third appellant's case.

[86] Since the statute specifically excludes an appeal, it is not competent for the third appellant to seek to challenge the RAO. The arguments supporting that challenge are rejected for that reason. This ground was introduced by way of a successful application to amend the Note of Appeal and was supported by an Opinion of Counsel, which made no reference to section 210B(6). The court can only assume that leave to amend the Note was granted in ignorance of that provision.

[87] In any event, the court is quite unable to sustain the argument that the judge erred in making a RAO. The ground of appeal, which the sentencing judge was belatedly asked to comment upon, starts with a contention that he had failed to state whether the RAO was being made in respect of, on the one hand, (in short) a sexual or violent offence (210B(1)(a)) or, on the other, an offence indicating a propensity to commit a sexual or violent offence (210B(1)(b)). The answer to this was, not surprisingly, regarded by the judge "as so obvious ... that it needed no express statement". It is indeed obvious that the offence was regarded as one of violence.

[88] The only other contention in the ground is that, in relying on the offence itself, the judge ignored a number of mitigatory factors; in particular the gap in offending and the summary nature of the record. The judge responded to this by stating that he determined that the nature of the offences, which involved a series of violent assaults upon women over a long period of time, "might" be such that the criteria were met and that he did not consider the gap or the lack of prior custodial sentences as indicating that the criteria could not be met. In so phrasing his report, it cannot reasonably be contended that the judge ignored the gap or the nature of the record. On the contrary, he makes it clear that he did consider these elements prior to making his decision. Although the submission ultimately made focussed upon the weight put upon these factors by the judge, that submission is not contained in the ground of appeal and the judge has thus not had the opportunity to comment upon that aspect of the case. However, there is nothing in what the judge reports to suggest any failure to attach appropriate weight to a particular factor.

[89] The contention, that the judge proceeded on the basis only of the offences and analysed the other matters solely in the context of whether they undermined the inference of possible risk stemming from the offences, also does not feature in the ground of appeal. It seems to be an argument founded upon the order of words in the judge's description of his reasoning in his supplementary report. There appears to be little merit in this line, even were it to have been focused in a ground of appeal and commented upon. The judge was not addressing weight or his approach in his report. He was responding only to on the appellant's contention that he ignored certain specific factors. He has rejected this contention and there is no basis for attacking his account.

[90] Finally, in relation to the need for a CJSWR prior to making a RAO, this is also not a ground of appeal and the judge's comments upon it are not available. However, as distinct from the position with an extended sentence, there is no requirement to obtain a CJSWR prior to making a RAO or, indeed, an OLR (see infra). That is not at all surprising given the level of investigation and analysis which will be involved in the preparation and presentation of a RAR, as distinct from a CJSWR. A risk assessor can be expected to include almost all the information and advice which would be expected in a CJSWR. In the absence of a statutory requirement to obtain a CJSWR prior to making a RAO or indeed to consider whether an extended sentence might be an alternative to making a RAO (cf an OLR infra), there is no substance in this argument.

(b) The interpretation of section 210E
[91] There is no contention that the language of section 210F(1), which requires the court to make an OLR if "on a balance of probabilities, the risk criteria are met", is ambiguous. It is thus not possible to read into this provision any gloss to the effect that an OLR should not be made unless the person can be classified as a "high risk offender" or that the risk he poses must be categorised as "high" in the mind of the judge or the risk assessor. It is, in any event, clear from the progress of the legislation, that the absence of such qualifying words was noticed by the Parliament but they were specifically not incorporated into the amending legislation which created the OLR as a sentence.

[92] In terms of section 210E, the risk criteria, read shortly, are that the "nature ... or the circumstances of ... the offence ... either in themselves or as part of a pattern of behaviour are such ... that there is a likelihood" that the offender "if at liberty, will seriously endanger the ... the public ...". It is for the court to determine whether there is a likelihood that an offender will seriously endanger the public. That is a question of fact, the answer to which will depend upon a consideration of the particular offence and any pattern of individual behaviour. Judicial experience will play an important part in this determination, but the court must have specific regard to the terms of the RAR and "any other information" (s 210F(1)) before arriving at a conclusion.

[93] There are two areas of potential ambiguity in the section 210E criteria. The first is the meaning of "likelihood". The second is the interpretation of "if at liberty". In relation to both, it is necessary to ascertain the purpose of the provision and thus to have regard to pre-legislation material in order to construe the meaning of the statutory language in its proper context, bearing in mind, however, the recognised constraints on that approach (Tortolano v Ogilvie Construction 2013 SC 313, LJC (Carloway) at paras [47]-[48] under reference to Pepper v Hart [1993] AC 593).

[94] The starting point is no doubt the description of the order itself as one of "lifelong restriction". Such an order is a sentence of custody for an indefinite period (s 210F(2)). However, at the same time as the OLR was introduced, it was provided (by amendment to the 1993 Act, s 2) that the offender would be eligible for parole upon the expiry of a "punishment part", being a period fixed by the judge to reflect the requirements for retribution and deterrence (1993 Act ss 2(2)). Parole would be granted if the Parole Board considered that it was "no longer necessary for the protection of the public that the prisoner should be confined" (1993 Act, s 2(5)(b)). What was in contemplation therefore is a person who, on the face of things, requires monitoring, of one sort or another, for the rest of his/her life. It is envisaged that he/she will spend a period in custody, but that he/she will be eligible for, and may be granted, parole after the expiry of the punishment part. The order should not be seen as necessarily resulting in lifelong custody (cf a life sentence) and it can form no part of any test that the court considers that the offender must be locked away permanently in order to avoid serious endangerment of the public. At the other end of the spectrum, it cannot be sufficient for the making of an OLR that, at the point of sentencing, the court considers that, if the person were to be immediately set at liberty, he would seriously endanger the public. Were that to be the test, the incidence of OLRs would be likely to increase significantly.

[95] In Liddell v HM Advocate 2013 SCCR 541, "likelihood", in the context of this specific statutory provision, was, following the (Shorter) Oxford English definition, held to mean "something that is likely; a probability" (Lady Paton at para [52]). The court requires to follow that decision, which is binding upon it. There is support for this meaning in other Scottish contexts (see eg Fotheringham v Dunfermline DC 1991 SLT 610, Lord Marnoch at 611; North Uist Fisheries v Secretary of State for Scotland 1993 SLT 333, Lord Cullen, delivering the Opinion of the Court, at 336). What the judge must therefore determine before making an OLR is that it is likely (in the sense of more likely than not) that, were the offender to have been sentenced otherwise than by the imposition of an OLR, he would seriously endanger the public once at liberty.

[96] Where a person's liberty is at stake, the court will adopt the higher of possible alternative tests for the restriction of that liberty. The court has not ignored some strong dicta to the effect that a lesser test may have been intended by the use of the term "likelihood". For example, the court has carefully considered Lord Mance's dictum (with which the other members of the United Kingdom Supreme Court agreed) in R (on the application of Sturham) v The Parole Board [2013] 2 AC 254 (at para 17) that the word "likely" is not to be read in a mathematical sense of "more probable than not". Rather it is "a word of open meaning" and, in the context of the similar legislation under consideration, means "good grounds for believing that the offender might remain a serious danger to the public". The court has also considered Lord Reed's definition of "no likelihood" as "no risk, that is a real risk which ought not to be ignored" in Ludriecus v Thomson 2008 SCCR 996 (para [22]).

[97] In R v Piri [1987] 1 NZLR 66, Cooke P, chairing the Court of Appeal of New Zealand, remarked (at 78) that:

"the words 'likely' and 'probable' ... are variable in meaning. Each can mean, in an appropriate context, more probable than not, but that is certainly not the only available meaning. This may be simply illustrated by noting that the Concise Oxford Dictionary includes in its definitions of likely 'Such as may well happen ... probable'".

Both Cooke P (at 82) and McMullin J (at 83) referred to Boughey v R (1986) 161 CLR 10, in which the majority of the High Court of Australia (Manson, Wilson and Deanne JJ) held that, in a particular statutory context, "likely" was not to be taken to mean "more likely than not" but "to convey the notion of a substantial or real chance" (see also In re O (Minors) (Care: Preliminary Hearing) [2003] 2 WLR 1057, Lord Nicholls at 1081).

[98] A third view is that it is best not to attempt to place any gloss on statutory wording which has an apparently clear meaning (Ludriecus v Thomson (supra), Lord Carloway at para [25]); Three Rivers DC v Bank of England (no 4) [2003] 1 WLR 210, Chadwick LJ at paras 21-22 and especially 33). However, given that the statutory context is liberty of the person, the meaning of likelihood must be that of probability; the court requiring to be satisfied, before making an OLR, not just that there is a substantial or real chance of serious endangerment but that such endangerment is more likely than not to happen or, put another way, that it will occur. In this context, the fourth appellant appeared to be arguing for the application of a lesser test for an OLR than is appropriate.

[99] The time for assessing the likelihood of serious endangerment is, and can only be, at the time of sentencing. However, that assessment is looking forward to the point at which the offender would, but for the OLR, be "at liberty". It is important to note the historical context of the legislation. It was introduced at a time when, apart from life sentences, custodial disposals were all determinate but where, in the case of short term violent prisoners, a SRO was available and, in the case of long term prisoners and sexual offenders convicted on indictment, both parole and later the extended sentence were available. What the OLR legislation was looking at, when it was introduced, was the point at which an offender would be "at liberty" under the pre-existing regime. Thus the judge has to contemplate whether there is likely to be serious endangerment at the point in the future when, but for the imposition of an OLR, the offender might be predicted to be at liberty.

[100] Given the terms of section 210A(1)(b), if a judge considered that, at the end of any determinate sentence then under contemplation, the offender (duly released on licence) posed a threat of serious harm to the public, he/she would probably impose an extended sentence. What the OLR legislation was (and is) considering is whether, when an offender is at liberty, he/she will seriously endanger the public. This contemplates the period after he/she would be released from prison, although possibly the subject of parole or extended sentence supervision. It also contemplates, however, and has firmly in mind, the period when the person ceases to be the subject of parole or extended sentence licence.

[101] This approach does not require a precise calculation of when the offender is likely to have been released (but for an OLR) or an accurate prediction of his state at that time. Rather, it envisages the court assessing the existing risk posed at the time of sentencing and an assessment of whether any custodial or post release regime, short of an OLR, will have any material impact on that risk. If the judge, in light of, especially but not exclusively, the RAR, considers that no material reduction of the risk will occur, he/she would be entitled to find that any likelihood of serious endangerment at the time of sentencing will be, for practical purposes, that which will exist on release from custody.

[102] Putting matters in a slightly different manner, the judge contemplating making an OLR is looking at the likelihood of serious endangerment when the offender is at liberty, but taking into account what he/she might be anticipated to have achieved by way of rehabilitation whilst in custody and the predicted effects of post release supervision. However, the judge has to bear in mind that, under standard (not OLR) sentencing regimes: first, rehabilitation programmes, including courses, cannot be forced upon an offender; and, secondly, any period of post release supervision will inevitably expire. If serious endangerment is regarded as likely at any point in what, but for the imposition of an OLR, would have been the offender's post release future, an OLR must be made.

(c) The extended sentence comparison
[103] Sentencing is "a delicate art based on competence and expertise" (Gemmell v HM Advocate 2012 JC 223, LJG (Gill) at para [59]) rather than an exact science. A judge should not have to go through a formal check list of procedures before arriving at the appropriate and proportionate sentence. Having ordered a RAR, the judge will have considered that one sentencing option is an OLR. Once the RAR is available, the judge will have the risk assessor's view on the category of risk posed by the offender in terms of the RMA definitions of high, medium or low. He must take the content of the RAR into account and consideration of it must form an important part in his decision making process. However, the sole question which the judge requires to ask himself/herself is whether the risk criteria have been met; that is to say whether, again putting matters shortly, "there is a likelihood that [the offender], if at liberty, will seriously endanger" the public. In answering that question, there is no alternative question to be posed and answered, such as whether a determinate or extended sentence might be appropriate. There is one question; the answer to which is either positive or negative. It is only if it is negative that the judge will go on to consider other sentencing options and, in particular, whether any tests for the imposition of these options have been met.

[104] Nevertheless, although the reasoning process required before making an OLR need not involve any express contemplation of an alternative determinate or extended sentence, that exercise, as already observed, does involve a consideration of the offender's likely conduct once at liberty from what would, but for the OLR, have been a determinate or extended sentence. Clearly, if the judge reached the view that there was no likelihood of serious endangerment after release from incarceration in the context of either a determinate or an extended sentence, he would not impose an OLR. That, however, involves a slightly different approach from one which requires a judge to move up a ladder of escalating punishments before fixing upon a particular rung as being at the appropriate level. Such an approach is too mechanistic in the context of a sentencing decision which, although often involving a complex matrix of factual and legal material, will be instantaneous, if not quite instinctive (cf Gemmell v HM Advocate (supra) LJG (Gill) at para [59]), once the material is ingathered and understood. Thus, the judge will, from his knowledge and experience, have in mind such general elements as the effects of maturity on criminal conduct, the prospects of reform or rehabilitation as a result of programmes in the prison setting and the effectiveness of post-release supervision under parole licence. The judge will also have the specific predictions and recommendations of the risk assessor in relation to the particular offender. However, the decision on whether the statutory test has been made remains that of the judge and the judge alone.

(d) The relevance of risk categories
[105] The utility of the risk assessor's categorisation of an offender's risk as high, medium or low is as a tool for the sentencing judge. The risk assessor must, in accordance with the RMA guidelines, advise the court on whether the risk posed by an offender is "high, medium or low" (s 210C(3)). There is no statutory guidance on what the court is to make of the categorisation. There is no indication that only those categorised as a "high risk" should be made the subject of an OLR or that those thought by the risk assessor to be at "low risk" should not be made so subject. This lack of instruction accords with the general principle that sentencing is for the judge, and not the risk assessor. In that context, the judge may disagree with the assessor's categorisation, having regard to his/her own knowledge and experience or because he/she does not consider that the risk assessor's conclusion is supported by his analysis of the facts.

[106] Nevertheless, the structure of the legislation, and its reference to the need for a RAR to be compiled in accordance with RMA guidelines, makes it clear that the judge ought to pay particular attention to the views of the expert risk assessor and hence to the categories of risk as defined by the RMA before reaching a decision. The risk assessor's view has a statutory function. If an offender is categorised as at high risk then he has, in terms of the definition, "an enduring propensity" seriously to endanger the public and the prospects of this changing are at least "significantly limited". This, if accepted by the judge, is a strong indicator pointing towards the necessity of an OLR. It is not, however, binding upon the judge.

[107] With a medium risk offender, the offender has a "propensity" seriously to endanger the public with characteristics which are problematic, persistent or pervasive, but which may be amenable to change. In such a case, the judge may well have to take a view on the realism of the prospects of such a change in order to decide whether the criteria are met. It is of some significance, however, that the starting point is the existence of the "propensity"; being an inclination or tendency to behave in the way described in the criteria. The categorisation assumes the necessity of either significant change or effective management in order to dampen the propensity. It may not be surprising, therefore, if a judge decided that a mature and persistent offender in the medium risk category met the statutory criteria. Such a decision could not be successfully challenged merely on the basis of the risk assessor's categorisation. That was the ratio of LBM v HM Advocate 2012 SLT 147 under the previous RMA definitions. However, the court in LBM was not suggesting that a medium risk always or even normally merited an OLR, merely that it may do so depending upon the judge's own assessment. Where the offender is a young man or one whose actions on the particular occasion did not appear to be prompted by his underlying personality traits but by the ingestion of drink or drugs, the prospect of change over time as a result of maturity or rehabilitation measures would render it unlikely that a judge could reasonably consider that the statutory criteria were met.

[108] With a low risk offender, there is only a capacity seriously to endanger the public. The offender has no "long term or persistent motivation or propensity to do so". If the judge accepts that categorisation, such an offender would not meet the statutory criteria (LBM v HM Advocate (supra), Lord Carloway, delivering the Opinion of the Court, at para [11]).

(d) Individual Appellants
[109] In relation to the individual cases, the ultimate question for the court is whether the sentences were either excessive or inappropriate. This involves a consideration of whether the sentencing judge was correct to hold that the risk criteria were met. In each case, the age of the offender, his record and the nature of the offence under consideration are, along with the RAR classification and analysis, of particular importance in that exercise.

[110] The appellant was 29 at the date of the offences. He was then at an age when he would have been expected to have reached mature adulthood and be beyond the violent impulsiveness present in some male youth. He had no history of employment or steady relationship. He had a significant criminal record for violence extending back over a decade and largely devoid of any significant interval. He had only very recently been released from a sentence of almost four years, when he had engaged in acts of violence. That sentence was mainly for concern in the supply of heroin, but there were four assaults, whilst on bail, included in the conviction at that time. There is a background of previous convictions for assault and robbery in both 2007 and 2008. The 2008 conviction must have been for a significant offence, since it attracted a sentence of over two years. It appears to have involved a knife. There were assaults in 2003 and 2004. The present offences again involved the use of weapons.

[111] There was good reason to conclude that the CJSWR was accurate in its conclusion that there was a "high risk of imminent reconviction" and "a high risk of serious harm" upon release unless the appellant developed "alternative strategies for managing conflict". It stated that the appellant required a "very high level of monitoring and supervision in the community". The RAR was, as already narrated, hardly more optimistic in its description of the appellant as "currently ... unmanageable" in the community. Re-offending was likely on release even if the RAR contained some positive content in the event of "significant input from others, and significant effort" from the appellant.

[112] This appellant is, therefore, a significant recidivist who has, on many occasions, been involved in episodes of assault and knife carrying. He re-offended repeatedly almost immediately upon release from custody; attacking four different people, including a woman, for very little, if any, reason. The sentencing judge described him as a "violent and dangerous man who meets the risk criteria". Given what was reported as his "pro-criminal" attitude and his stated intention to behave in a similar manner if faced with similar situations, it is not possible to fault the judge's conclusion. But for the imposition of an OLR, a custodial term of 4 years and 6 months would have been imposed. It is likely the appellant would have struggled to be paroled, but he might have anticipated release, subject to restrictive conditions under an extended sentence, after about 3 years, ie in about July 2015. He would then be 32 years of age.

[113] From the terms of both the CJSWR and the RAR, the sentencing judge was entitled to the view that there would be no material change in the risk posed by this appellant between the date of sentencing and his anticipated release date (but for the OLR). In all these circumstances, the risk criteria were met. Not only would the appellant be a danger to the public upon release but also it is amply demonstrated that there is a likelihood that he will seriously endanger the public then, even if subject to restrictive conditions. For all of these reasons this appeal must be refused.

[114] The circumstances of the offence, the appellant's personal life and his record have already been noted. It is of particular importance to notice that the appellant was 28 at the time of the offence. This is a similar age to the first appellant. Unlike that appellant, however, he had lived in family with his partner and children and was in employment. The offence was a single one, albeit one of attempted murder. The injuries were severe, but there was no aggravation by way of lasting sequelae. The offence has to be seen in the context of the directly analogous previous conviction for assault to the danger of life.

[115] The sentencing judge confined his comments on why he imposed an OLR to the content of his "sentencing remarks". It is, perhaps for that reason, not entirely clear why he reached the view that the appellant would, if at liberty, be likely seriously to endanger the public. There is thus force in the criticism that there is barely any adequate explanation as to why the judge departed from the clear terms of the RAR, whose author was clear that an OLR was not appropriate since the appellant did not fall into any exceptional category. The latter is not, however, the test. The issue is whether the statutory criteria were met.

[116] Had he not imposed an OLR, the judge would have sentenced the appellant to 6 years and 8 months imprisonment and, presumably, might have imposed an extended period of supervision beyond that. With the appellant's record, it is reasonable to assume that he may not have been granted parole immediately upon reaching one half of his prison term. He may have been given parole before he had served two thirds of that term (ie about 4 years 5 months). Assuming that he would serve approaching 4 years in custody, he would be aged about 34 when at liberty and then subject to a period of licence and supervision for an additional period of several years. The question is whether there is a likelihood that, after the time of his hypothetical release in or about July 2016, when the appellant would be about 32, or thereafter when any supervision ceased, he would seriously endanger the public.

[117] It is of importance to note that, in the case of this appellant, the offence was prompted in part by the level of alcohol consumed rather than being, for example, brought about by any permanent psychological or other condition. As the risk assessor said, this appellant's profile is not an exceptional one. He does have problems, but these are of a type which can be addressed both in prison and under the types of condition which might be imposed if the appellant were released from the custodial element of an extended sentence. The risk posed by the appellant is potentially manageable, even if there can be no guarantee that any custodial or post release regime will succeed in eliminating all risk. Nevertheless, in these circumstances, it cannot be said that the statutory criteria for an OLR have been met, although those for an extended sentence undoubtedly have. The OLR should be quashed and an extended sentence of 10 years (with a custodial element of 6 years, 8 months) imposed in substitution.

[118] This appellant is the only one of the four offenders who has been categorised as "high risk". It is not difficult to see why he was so assessed. As already noted, the risk assessor described him as suffering from a "dissocial personality order" with "psychopathic, narcissistic and paranoid personality traits". Other factors are "young age at first violence, history of previous violence, lack of insight, negative attitudes, con-compliance with remediation attempts, plans lack feasibility, substance misuse, supervision failure, employment and relationship problems and ... a high risk of exposure to destabilisers which would make him recidivate in a violent manner".

[119] Of significance was the fact that:

"1.6 It is not possible to identify early warning signs of the likelihood of his violence as his decision to act violently does not, for example require him to be destabilised or disinhibited. Violence is functional for [the appellant] in that he uses it to have his needs met."

The risk assessor reached the view that:

"1.7 ... treatment and risk management strategies cannot be implemented at the present time which would be effective in reducing this risk of future violence in the short or medium term. Any prospect of change is likely to require long term intervention."

The reasons for this were the appellant's "negative attitudes towards addressing his offending as well as his personality disorder". The risk assessor referred to the likely scenario if the appellant were set at liberty now but she also considered "future scenarios" and expressed the opinion that the appellant:

"1.9 ... presents with a chronic and enduring long-term risk of serious violence to a range of victims including adult female partners (current and former), and male peers. He also poses a risk to family members and children. The way in which his violence would manifest is likely to be diverse."

The sentencing judge took into account these findings and concluded that none of the recommended supervision, monitoring and victim-safety measures could confidently be predicted to be effective.

[120] The submission that the RAR addressed only current risk cannot be sustained. It is clear, even from the above summary, that the risk assessor considered both present and future risk. Equally, the contention that there was no indication that future management programmes would not be effective is not borne out by the RAR, which states that the anticipation is that they will fail. There is no indication that the risk assessor failed to notice the lack of convictions recorded from 1996 to 2001. As the judge himself observes, this gap has to be seen in the context of "frequent and serious violent offending ... against a number of victims" since then. The judge has stated that he did not regard the risk assessor's conclusions to be dependent to any extent upon allegations of sexual violence.

[121] For all of these reasons, the judge was entirely justified in concluding that the statutory criteria were met in this appellant's case. He has given clear adequate reasons for so finding. It is evident from the RAR that, were the appellant to be released from a finite sentence, with or without an extended period of supervision with conditions, he would remain a serious danger to the public. In that regard, the narrative in relation to the charges put these offences into a quite exceptional category of repeated domestic violence over a prolonged period and involving several different women. This appeal must therefore be refused.


[122] This appellant has the distinction of having three High Court convictions for extreme violence. The first was in 2005, when he was imprisoned for 5 years for an assault to severe injury and the danger of life. This involved stabbing a man a number of times on the trunk and abdomen. The second, which may have occurred during a period of licence, was in 2007 for attempted murder, when an extended sentence of 10 years (6 years custodial) was pronounced. This involved wilful fireraising at a flat occupied by a couple and their teenaged son. The third is the present conviction for attempted murder, which was committed whilst on licence for the earlier analogous crime. The appellant had been released in December 2011. These convictions alone put the appellant into an unusual category, quite apart from his assault and robbery offence in 1991, which attracted 30 months, and two convictions for carrying offensive weapons in 2003 and 2004. The appellant was fortunate not to be facing a mandatory life sentence for murder.

[123] It is true that the risk assessor categorised the appellant as only at "medium risk". He thought that the appellant suffered from an antisocial personality disorder, which is common in the prison population, but was not psychopathic. The latter factor was particularly important. The risk assessor considered that the appellant's risk would be "high" if he failed to address his problems of substance abuse, poor emotional and behavioural control, unstructured lifestyle and unstable relationships. However, because he had the capacity to change, "measures short of lifelong restriction may be sufficient to minimise the risk of serious harm".

[124] The sentencing judge was entitled to hold that the criteria had been met and that the appellant, once at liberty would seriously endanger the public. He would, but for the OLR, have been given an extended sentence with a custodial element of in excess of 9 years consecutive to a section 16 return order. It might be anticipated that the appellant, who was 40 at the time of the offence, would spend at least 8 years in prison and be released when he was approaching 50. Nevertheless, he has twice been released from significant custodial sentences for acts of violence. On the second occasion he was the subject of extended sentence supervision. On both occasions he committed further acts of serious violence. In these circumstances, the "likelihood" of repetition was made out. An OLR was mandatory and this appeal must be refused.


[2014] HCJAC 19

Lord Justice Clerk

Lord Drummond Young

Lord Clarke

Appeal Nos: XC224/13, XC258/13, XC292/13 and XC385/13


in the Appeals against Sentence by







Act (1): JDM Macara QC (sol adv), NG Allan, Solicitor Advocate; BCKM

Act (2): J Keenan, Solicitor Advocate; Capital Defence (for Keegan Smith, Livingston)

Act (3): SM McCall, Gilchrist; John Pryde & Co

Act (4): A Ogg, Solicitor Advocate; Drummond Miller (for JC Hughes, Glasgow)

Alt: Fairley QC AD; the Crown Agent

28 February 2014

[125] I am in agreement with your Lordship in the chair as to both the interpretation of the statutory provisions that govern OLRs and their application to the four cases that are presently under consideration. Because of the importance of the issues that are involved, I add some observations of my own.

[126] In the first place, it is important in my opinion to emphasize that the statutory provisions governing OLRs leave important elements of judicial discretion intact. This is so despite the fact that the legislation at first sight appears significantly more prescriptive than most other legislation governing sentencing. Central to the legislation is the definition of the risk criteria in section 210E, namely that

"... the nature of, or the circumstances of the commission of, the offence of which the convicted person has been found guilty either in themselves or as part of a pattern of behaviour are such as to demonstrate that there is a likelihood that he, if at liberty, will seriously endanger the lives, or physical or psychological well-being, of members of the public at large".

Initially, if the judge considers that the risk criteria may be met, he is obliged by section 210B(2) to make a RAO; I agree with your Lordship in the chair that there is no discretion in this matter. Moreover, at this point the threshold is not especially high; that is apparent from the word "may". The intention here is clearly that an expert risk assessor should be brought into the sentencing process, to provide advice as to the level and type of risk that the offender presents. Once the RAR is available, the judge must consider whether an OLR should be made. In this regard, section 210F stipulates that, if the court is satisfied that "on a balance of probabilities the risk criteria are met, [the judge] shall make an order for lifelong restriction...". Thus the making of an OLR is mandatory if the judge is satisfied that the risk criteria are met. In considering this question, it is the statutory test that must be addressed, and other considerations, such as whether the offender can be considered at "high risk" of reoffending, are not relevant. Nevertheless, the evaluation of whether the risk criteria are in fact met is an exercise that involves important elements of discretion on the part of the judge.

[127] In exercising this discretion, the judge must in my opinion take careful account of the information contained in the RAR. The function of the risk assessor is to provide expert advice as to the risk that the offender is likely to pose at a future date, an obviously specialized task that is dependent on a wide range of different factors. The risk assessor's categorization of the risk presented by the offender as "high", "medium" or "low" is not decisive, but in my view it is a significant part of the overall assessment of the offender, and should therefore be taken into account by the sentencing judge. That applies in particular to offenders who are categorized as being at "high" risk; what that means is an assessment that the individual's behaviour indicates "an enduring propensity to seriously endanger the lives, or physical or psychological well-being, of the public at large". That clearly goes some way to meeting the requirements of the risk criteria in section 210E. Furthermore, "high" risk indicates that the individual has "problematic, persistent, pervasive characteristics that are relevant to risk" which are either not likely to be amenable to change or alternatively the potential for change with time and treatment is significantly limited. The enduring nature of the propensity is important, because a major factor in imposing an OLR will always be the risk to the public presented at a future date, when the offender may be released from custody. Nevertheless, even when an offender is categorized as being at high risk, the judge must in my opinion still consider the remainder of the RAR and the nature of the offences for which a sentence is being imposed, to ensure that the conclusion can properly be drawn that the risk criteria are satisfied. Mere acceptance of the risk assessor's categorization as high risk will not be enough by itself.

[128] Consequently the judge must evaluate the whole of the information contained in the RAR, together with the information that he has about the instant offences, and in the light of those reach a judgment as to whether the risk criteria are met. That means in particular that he must determine whether there is a "likelihood" that the offender will, if at liberty, "seriously endanger the lives, or physical or psychological well-being, of members of the public". In this respect, I agree with your Lordship in the chair that the word "likelihood" means probable, in the sense of more likely than not; that is a stringent construction, which is clearly justified by the fact that the liberty of the individual is involved.

[129] It is also important in my opinion to have regard to the demanding terms of the risk criteria. The particular risk that is referred to is that of seriously endangering the lives, or the physical or psychological well-being, of members of the public. In addressing this risk, three elements are in my opinion important: first, the likelihood, in the sense of probable occurrence, of offences that seriously endanger the public; secondly, the probable seriousness of those offences; and thirdly the time at which the existence of the risk criteria must be considered. As to the third of these matters, for the reasons stated by your Lordship in the chair it is clear that the material time is not the time of sentencing but the time when the offender will become eligible for parole on the expiry of the punishment part of his sentence; that is the time when he may be "at liberty", and it is then that the relevant risk emerges. It is obviously essential that RARs should address the risk at that time. In this way, factors such as the beneficial effect of programmes undertaken in prison and the general effect of increasing age will be taken into account in determining whether there is a likelihood that the offender will seriously endanger the public in the manner suggested in section 210E. It is also important that, in assessing the risk at that time, the availability of post-release supervision should be taken into consideration. The sentencing judge must clearly have regard to the comments in the RAR on all of these matters.

[130] As to the first and second of the elements that must be considered in addressing the risk criteria, it is obvious that both the likelihood that the offender will re-offend and the probable seriousness of any such future offending must be taken into account. The critical question is whether those factors indicate that it is more likely than not that the offender will, if at liberty, seriously endanger the lives or physical or psychological well-being of the public. This requires the exercise of judgment by the sentencer. I agree with your Lordship in the chair that it is inappropriate to lay down detailed procedural requirements; what is involved is essentially the exercise of an intuitive judgment, based on experience and a feeling for what is fair and just in the circumstances of the case. In imposing sentence, particularly a very lengthy sentence, proportionality is clearly of great importance. The availability of other forms of sentence, notably the extended sentence, may be relevant in some cases, especially if it appears that any serious risk to the public may decrease in the future. Nevertheless, it must be borne in mind that, at the end of an extended sentence, the offender must be released, even if rehabilitation has been unsuccessful; there is no power to deal with any continuing risk through procedures such as release on licence subject to stringent conditions. If an OLR is imposed, by contrast, release on licence at any time may be made subject to conditions, for supervision and the like.

[131] The four cases that are the subject of these appeals indicate differing circumstances in which an OLR may be appropriate. The appellant Ferguson had a record that was marked by persistent reoffending throughout his adult life, together with the misuse of alcohol and drugs. While the offences were not at the most serious level, they included repeated assaults, including the use of a knife and an assault and robbery, for which he received a sentence of two years' imprisonment in 2008. Various forms of non-custodial sentence had been attempted, including a drug treatment and testing order, without success. His last major sentence, in 2010, was of four years' imprisonment for drug-related offences. Both the Criminal Justice Social Work Report and the RAR were in unfavourable terms. The offences for which the present sentence was imposed occurred very soon after release from custody. They were in themselves relatively serious: three assaults on acquaintances with weapons, on two occasions a knife, and assault and robbery using a knife. These offences took place over a period of only five weeks. Against that background I consider that the sentencing judge was fully entitled to take the view that the appellant met the risk criteria and that, on the basis of the CJSWR and the RAR, that risk would not lessen before the release date that would otherwise have applied.

[132] The appellant Cameron is in a different situation. He had been in regular employment and lived with a partner and children. This contrasted markedly with Ferguson, who had never been employed and appear to live a wholly disorganized life. The offence to which Cameron pled guilty, attempted murder, was clearly very serious and had been preceded by a conviction for assault to the danger of life. Nevertheless, the present offence appeared to have been brought about by excessive consumption of alcohol. That is a problem that can be addressed by appropriate programmes in prison. Thus it is impossible to state that at the expected date of release, in approximately July 2016, there will be a serious danger to the public that meets the risk criteria. In this case an extended sentence is clearly appropriate, however, in the manner indicated by your Lordship in the chair.

[133] The appellant Balfour was categorized as high risk in his RAR. He pled guilty to charges of repeated assaults against five different women over a period of 17 years. It is clear that the assaults were serious and persistent. The appellant was assessed in the RAR as suffering from a dissocial personality disorder together with paranoid, narcissistic and psychopathic personality traits. The risk assessor thought that new relations with women were likely to result in broadly similar physical violence, possibly at a severe level. Furthermore there were no protective factors. In those circumstances I am of opinion that the sentencing judge was clearly entitled to hold that the risk criteria were met; the offences were serious and frequent, and the appellant clearly posed a serious risk to any women with whom he entered relationships, to their families and to others.

[134] The appellant Nolan was different from Ferguson and Balfour in that he only had three serious convictions, but all of these were for extreme violence, and the first two had resulted in sentences of five years' and 10 years' imprisonment respectively, the latter being an extended sentence with six years in custody. The present offence was attempted murder, committed while the appellant was on licence. In this case the seriousness of likely offending was clearly very high, and the risk of future offending was clearly substantial, particularly if problems including substance abuse and self-control were not addressed. Thus in this case the seriousness, as against the frequency, of the appellant's behaviour was especially important. In all the circumstances I agree that the sentencing judge was fully entitled to hold that the risk criteria were met.

[135] In conclusion, I should note one further factor. We were informed that at the date of the hearing of these appeals no offender who had been made subject to an OLR had yet been released. At the present stage this is perhaps not surprising, as such orders are imposed with reference to the probability of offending at a future date when the offender is released from prison. OLRs have only been competent since 2006, and if the risk criteria were indeed met when the sentence was imposed it is quite likely that the risk to the public would still exist in every case after seven or eight years. Nevertheless, the punishment parts have expired in a substantial number of cases. In these circumstances it is perhaps worth emphasizing that prisoners subject to OLRs must have their cases reviewed regularly, to ensure that continued custody is necessary to meet the objectives of the OLR. That is clearly contemplated by the terms and structure of the governing legislation, and it is in my opinion essential to ensure that the OLR does not become an unjustified form of preventive detention.


[2014] HCJAC 19

Lord Justice Clerk

Lord Drummond Young

Lord Clarke

Appeal Nos: XC224/13, XC258/13, XC292/13 and XC385/13


in the Appeals against Sentence by







Act (1): JDM Macara QC (sol adv), NG Allan, Solicitor Advocate; BCKM

Act (2): J Keenan, Solicitor Advocate; Capital Defence (for Keegan Smith, Livingston)

Act (3): SM McCall, Gilchrist; John Pryde & Co

Act (4): A Ogg, Solicitor Advocate; Drummond Miller (for JC Hughes, Glasgow)

Alt: Fairley QC AD; the Crown Agent

28 February 2014

[136] I agree with your Lordship in the chair's analysis of the relevant statutory provisions and, in the event, I agree with your Lordship as to what the appropriate disposal is in each of the individual appeals before the court. Standing the importance of the subject-matter of the appeals, I would merely wish to make the following brief observations.

[137] While it seems apparent that Parliament, in employing the language it did in the relevant statutory provisions, provided for the imposition of an OLR in respect of offenders who may not fall within the particularly difficult class of offender, which the MacLean Committee recommendations were concerned with, it is, in my judgment, clear that the imposition of an OLR is only appropriate in relation to those who can be regarded as, in all the circumstances, exceptional offenders. Such orders are not to be imposed simply because of an offender's recidivist history. It has, in particular, to be borne in mind that, in this jurisdiction, the bulk of criminal offending, both minor and serious, is committed by young male persons. Experience also shows that, with, or without, assistance, many of them, in time, as they mature, grow away from their criminal behaviour. That is not, of course, to understate the damage that they do to society in the meantime. The imposition, however, of an OLR, in relation to such persons may have the effect of inhibiting, if not preventing, such a development. That is not, of course, to say that the imposition of an OLR would be inappropriate in relation to a male person of a young age, but it may well be a factor which, in a particular case, the sentencing judge would require to have careful regard to.

[138] There was argument before us that statutory provisions in relation to sentencing have arrived at a point which involves an incremental approach, by which was meant that the availability of an extended sentence for offenders who are considered to be in need of some continued monitoring after release from prison, means that such a disposal should normally have been attempted before the sentencing judge imposes an OLR. It is clearly not the case that statute requires such an approach. There will be clear cases where the sentencing judge will be satisfied that, having regard to all of the material before him, and having regard to the relevant statutory provisions, the appropriate disposal has to be the imposition of an OLR, notwithstanding that the offender in question has never been the subject of an extended sentence in the past. But, nevertheless, I would suggest that the availability of the extended sentence, notwithstanding its limitations, which were referred to on behalf of the Crown by the advocate depute before us, is something that the sentencing judge should not overlook before electing for the imposition of an OLR, particularly where such an approach has never been attempted in relation to a relatively young offender.

[139] Sentencing can be one of the most difficult and sensitive duties that a judge has to carry out. The legislature, in imposing statutory requirements on the sentencing judge, as it is entitled to do, does not lighten the burden. The description of the mental exercise the sentencing judge has to carry out in applying the provisions of section 210B et seq of the 1995 Act, as described by your Lordship in the chair, amply demonstrate this. In that respect, I require to add that while I, of course, accept that the ultimate responsibility for applying the legislation, and making the appropriate disposal, lies with the sentencing judge, the input, in that exercise, from the contents of the RAR, requires to be given very serious consideration by him, and are not to be lightly disregarded. That is because Parliament itself has directed the RAR shall have a very particular significance in the application of the relevant provisions.

[140] In the appeal by the appellant Balfour, it was contended that the sentencing judge had erred in not obtaining an SER prior to making the order in question. While it was accepted that the statutory provisions in question do not refer to this requiring to be done by the sentencing judge, it was submitted that, in the majority of cases, it would be in "good practice" to do so. I do not consider that the sentencing judge's failure to order an SER, in the case of the appellant Balfour, formed any basis of legitimate complaint. It is, no doubt, to be assumed that in the majority of cases the RAR will cover much of the ground that any SER would do. Having said that, however, a sentencing judge, in carrying out the exercise he is required to do in relation to deciding whether an OLR should be imposed or not, should feel, as confident as he can be, that he has all the necessary material before him to reach such a decision and, in certain cases, additional information and guidance that might be found in an SER may prove to be of assistance in that respect.

[141] I have indicated that, in the event, I am in agreement with your Lordship in the chair as to the disposal of each of these appeals. I have, however, to say that I had an initial hesitation as to the appropriateness of the imposition of an OLR in the case of the appellant Ferguson, having regard to the nature of his previous offending, prior disposals and the assessment of risk in the RAR. I have, however, come to the conclusion that, having regard to all the material before the sentencing judge, particularly in relation to the repeated random, highly antisocial behaviour of the appellant and his apparent total inability to operate normally in the community, the sentencing judge was entitled to reach the view that he did.

[142] Lastly, as regards the appellant Balfour, I consider that he clearly falls to be regarded as an exceptional offender, in view of the very extreme and repeated nature of the violence which he visited upon those females with whom he found himself in a relationship.