Bayley v The Queen [2016] VSCA 160

CRIMINAL LAW - Appeal - Conviction - Application for leave to appeal the convictions in respect of the third complainant - Whether trial judge erred in admitting telephone records - Whether relevant - Whether verdicts unsafe or unsatisfactory - Leave to appeal granted - Appeal dismissed - Evidence Act 2008 (Vic) s 55.

CRIMINAL LAW - Appeal - Sentence - Application for leave to appeal sentence - Whether non-parole period manifestly excessive - Application falls away as first conviction appeal is allowed - Sentences imposed on other convictions not disturbed - Total effective sentence of 12 years and nine months undisturbed - Non-parole period of 40 years fixed afresh.

1 On 19 June 2013, the applicant was sentenced to life imprisonment for the rape and murder of Gillian Meagher (known generally as 'Jill Meagher'). Nettle JA, sitting in the Trial Division, fixed a non-parole period of 35 years.[1]

2 Whilst serving that sentence, the applicant faced three trials in the County Court for rape and other offences perpetrated against a further three female victims, 'GH', 'BC' and 'HS'.[2]

3 In June 2014, the trial judge dealt with a series of preliminary legal issues relevant to the trials. The trial with respect to the offences against GH - the 'first trial' - was heard in July 2014. Next, the trial relating to the offences against BC - the 'second trial' - took place in March 2015. Finally, the trial involving HS - the 'third trial' - was also conducted in March 2015.

4 On 14 July 2014, the jury empanelled in the first trial convicted the applicant of all the offences on the indictment. On 12 March 2015, the jury in the second trial convicted the applicant of the two offences contained in the relevant indictment. On 26 March 2015, the jury in the third trial, by majority, convicted the applicant on all offences on the indictment.

5 On 28 May 2015, following a plea hearing, the judge sentenced the applicant

to 18 years' imprisonment,[3] and fixed a new non-parole[4] period of 43 years.

6 The sentences imposed are reflected in the following table:

First trial: victim 'GH' [Indictment C1309028.4]

Second trial: victim 'BC' [Indictment C1309028.2]

Third trial: victim 'HS' [Indictment C1309028.3]

• Disposal order
• Pursuant to s 6F of the Sentencing Act 1991, sentenced as a serious sexual offender

7 The applicant sought leave to appeal against the convictions returned in the first and third trials (but not the second). He also sought leave to appeal against sentence meaning, effectively, against the new non-parole period of 43 years.

8 So far as the first trial is concerned, the sole ground of appeal against conviction is that the verdicts 'are unreasonable or cannot be supported having regard to the evidence'.

9 With respect to the third trial, the two grounds relied upon are:

1. The verdicts of the jury are unreasonable or cannot be supported having regard to the evidence.
2. The learned trial judge erred by permitting the Prosecution to lead evidence of:

(a) Location codes and location descriptions from the Applicant's telephone records.

(b) CCTV footage depicting a dark sedan outside the Elephant & Wheelbarrow.

10 As will become clear, we are of the view that the convictions sustained in the first trial[10] cannot be permitted to stand. The application for leave to appeal against conviction must be granted; the appeal allowed; the convictions and sentences set aside; and in lieu thereof a judgment and verdict of acquittal entered.

11 With respect to the third trial,[11] leave to appeal against conviction should be granted, but the appeal should be dismissed.

12 Self-evidently, the quashing of the convictions and sentences in the first trial will necessitate the applicant being re-sentenced. For the reasons that we will later set out, we will fix a new non-parole period on all sentences of 40 years' imprisonment. Leave to appeal against sentence will otherwise be refused.

The conviction application on the first trial

13 In the first trial, involving the complainant GH, the identity of her attacker was the principal issue.

14 Prior to empanelment of the jury, counsel for the applicant, relying on s 137 of the Evidence Act 2008 (the 'Evidence Act'), sought exclusion of identification evidence. The evidence included a purported identification made by GH from 'Facebook', and a later photo board identification that she also made.

15 In a careful and detailed ruling, the judge refused to exclude the identification evidence, holding, first, that 'GH's Facebook identification [was] reliable and [was] evidence relevant to the issue in this trial, being the identity of her attacker'; secondly, that the evidence '[had] significant probative value, and the probative value of that evidence was not outweighed by any unfair prejudice' to the applicant; thirdly, that any unfair prejudice 'could be ameliorated by judicial direction and appropriate warnings'; fourthly, that the 'identification by [GH] relevant to the photo board [was] capable of significant probative value'; fifthly, that 'the potential for unfair prejudice did not outweigh the probative value of that evidence'; and, sixthly, the 'photo board identification had probative value in its own right, as did the Facebook identification'.

16 As indicated, the sole ground of appeal against conviction in relation to the first trial is that the verdicts are unreasonable or cannot be supported having regard to the evidence. There is no ground of appeal that challenges the judge's decision to permit the identification evidence to be led.

17 During the course of argument, the Court raised this matter with counsel for the applicant. The relevant exchange proceeded as follows:

WEINBERG JA: You sought, did you not, to have this [identification] evidence excluded.

[COUNSEL]: We did.

WEINBERG JA: And there was a ruling adverse to you but you do not complain about that ruling?

[COUNSEL]: No, we don't. In my respectful submission, your Honour, when one takes into account in particular the judgments of the High Court in Festa [(2001) 208 CLR 593], once this evidence was - it was open essentially to the trial judge to rule the evidence admissible. I cannot complain about that ruling as a matter of law.

We are then in a different position, which is whether the infirmities in that evidence were such as on proper intellectual rigorous analysis such as to mean that it was incapable of reaching the requisite standard. My submission is those are two distinct questions.

Perhaps taking up something associated with your Honour Justice Weinberg's comment. We are dealing here of course with a situation where we accept, as I have just indicated, that I cannot point to a legal error in the admission of the evidence of identification, of Facebook identification.

Equally we make no complaint, nor could we in light of the way the preparation of directions proceeded at the trial consistent with the Jury Directions Act as to the directions that were given by the learned trial judge. So our submission rests on the proposition that, notwithstanding those matters, that when analysed as a matter of logic, when analysed as a matter of rational consideration, the infirmities of that evidence were such that they are not capable of meeting the standard that was required. And that of course will then take us, if the Court accepts that submission - if it doesn't accept that submission then of course the application is over - if the Court accepts that submission then the question will become whether Mini Minor evidence was capable of effectively being added to that identification evidence, not to prove the identification evidence but by consideration of the combined effect of both strands of evidence, a conclusion of identity and in our submission it can't.

18 Therefore, it seems that based upon counsel's understanding of the High Court's decision in Festa v The Queen,[12] it was decided to deal with the identification evidence point under the rubric of what is generally termed the 'unsafe or unsatisfactory' ground, rather than on the basis of its wrongful admission.

19 Since the exclusion of the identification evidence that was challenged at trial would have left the Crown with no evidence capable of implicating the applicant in the commission of the offences, it is perhaps understandable that the matter was approached in this way. Nonetheless, and despite counsel's disclaimer, the argument before this Court developed as an attack upon the decision to permit the identification evidence to be led, as well as a challenge to the convictions themselves based upon their being unsafe or unsatisfactory.

20 Perhaps, as a matter of strict form, there should have been a separate ground of appeal challenging directly the judge's decision to admit the identification evidence. In the somewhat unusual circumstances of this case, however, we consider that the matter should be approached as one of substance rather than form. We therefore proceed upon the basis that the main thrust of the applicant's case is that the judge erred in permitting the identification evidence to be led.

21 For reasons that we will shortly explain, we are of the view that the judge should have excluded the identification evidence.[13] As we have said, without that impugned identification evidence there is simply no evidence upon which a jury could possibly find the applicant guilty. Accordingly, the convictions in the first trial cannot stand.

22 The offences relating to GH were alleged to have occurred sometime between 31 October and 1 December 2000.

23 GH gave evidence that 'around the time' of her 18th birthday, she was addicted to heroin and occasionally engaged in 'street prostitution'. Although her boyfriend, 'CN', was a drug dealer and would supply her regularly with heroin, one particular day after her birthday he held out on giving her drugs so as to watch her suffer.

24 GH was 'sick', and decided to obtain drugs to make herself 'feel better again'. That afternoon, she went to a health service for prostitutes, the 'Prostitutes Collective', located in Inkerman Street, St Kilda. She there obtained a pamphlet, entitled 'Ugly Mugs', which set out details 'of all the bad things that happened to the other girls on the streets'. GH then went to Grey Street, St Kilda.

25 GH said that whilst she was walking along the street, a male in a two door red Mini Minor car pulled up beside her. The driver said something like, 'Do you want to make some money?' GH replied, 'Yeah'. She then got into the car.

26 Asked to describe the driver, GH said that he had long, spikey hair; his hair colour was 'ranga' (that is, red); and he had 'orange, hairy arms'. He looked like a normal 'every day guy'. GH said that the male was wearing blue jeans. When asked the colour of his shirt, she said that she was 'pretty sure it was a blue shirt' (although she added that she was 'a bit iffy on the colour').

27 About five minutes after they drove off, GH was reading the Ugly Mugs pamphlet and remarked, 'There's so many bad men out there', and, 'Look at all the attacks'. The driver then punched her to the side of the face and said, 'You know what, I am one of them [sic] bad men'.

28 The male then drove his car to a narrow alleyway which backed onto houses. He parked in such a way that although the driver's door could be opened, the passenger door could not. The male called GH a 'junkie slut' and inserted his fingers into her vagina. He then obtained a condom from the glove box - about which GH said, 'the glove box was full of condoms, which was weird' - and then dragged her into the back seat of the car.

29 In cross-examination, GH agreed that the front and back seats were 'tan' coloured, as were 'the doors', 'the door trim' and 'the door panelling'.

30 Whilst GH was 'pleading [for her] life', the male had 'violent' penile-vaginal sex with her. He then had 'really bad anal sex' with her, such that GH 'was horrified and in agony'.

31 When another car drove into the alleyway, GH started 'smashing on the window'. This prompted the male to put his fingers down her throat, telling her that he would kill her if she did not stop screaming and hitting the window.

32 After about half an hour, the male took GH's purse and obtained her name and address from her driver's licence. The male then drove off with GH, heading away from the city. GH eventually managed to escape from the car and was picked up by a female driver who wanted to take her to a police station but she refused because she thought that police would not believe her.

33 GH was taken by the female driver to Balaclava Station. She later met CN at a hotel in the city. She told him that she had been raped. CN did not believe her, called her a 'slut' and punched her in the face.

34 GH gave evidence that 'some days' after the attack, she went back to the Prostitutes Collective, where 'a lady' advised her to report the matter to police. The lady also suggested she 'write it down on paper and put it on the noticeboard', which GH did. She described the red Mini Minor, and the male as having 'short, blonde spikey hair' and 'ranga arms'. She said that he was 'evil eyed during the attack'.

35 In the course of her evidence in chief, GH outlined what had eventually prompted her to contact police, nearly 12 years after she had been attacked. She said:

Well, I was home one day and I remember Jill Meagher, that case, on Facebook, they had a missing persons page for her and I was flicking on it one day and all of a sudden I've seen Adrian Bayley's face and I knew, a hundred percent, that's my guy.

36 GH subsequently identified a photograph, Exhibit E, as the photograph that she had seen displayed on Facebook.

37 We pause to note that other evidence established that Jill Meagher disappeared in the early hours of Saturday, 22 September 2012. As it transpired, she had been raped and murdered by the applicant in Brunswick shortly after 1.38 am. The applicant disposed of her body in Gisborne South later that morning. He was arrested five days later, at 2.08 pm on 27 September 2012.

38 GH said that, having seen the applicant's photo on Facebook, she called Crime Stoppers. Under cross-examination, she agreed that at the time she made that call, she was aware from media coverage 'of the discovery of [Jill] Meagher's body in a shallow grave out some distance from Melbourne'. She said that she also knew 'from Facebook' that 'Adrian Bayley had been arrested for the rape and murder of [Jill] Meagher'.

39 After contacting Crime Stoppers, GH subsequently had a telephone conversation with a member of the Homicide Squad. She was later contacted by the informant, Detective Senior Constable Christine Stafford, and made a statement to Senior Constable Stafford on 29 October 2012.

40 On 1 February 2013, Senior Constable Stafford told GH that the applicant had been charged with the offences against her. On 11 February 2013, Senior Constable Stafford asked GH to go to the Lilydale police station in order to view some photographs.

41 At the Lilydale police station, Detective Senior Constable Susan Weinbergs asked GH to view a photo board containing photographic images within a manila folder, Exhibit D. Senior Constable Weinberg gave evidence that when she showed Exhibit D - the photo board containing 12 images - to GH on 12 February 2013, GH opened the folder and identified photograph 4 (a photograph of the applicant) by saying, 'It definitely looks like No. 4'.

42 In cross-examination, GH agreed that she had seen the applicant's image in the media on a number of occasions, from the time of her Facebook identification up until the time of her identification from the photo board. Thus, at the time she was shown the photo board, GH knew not only that the applicant had been charged with the rape and murder of Jill Meagher, but also that he had been charged with the offences against GH herself.

43 Other evidence indicated that the Facebook photograph of the applicant had been taken in 2011 by his former partner. The applicant is shown in that photograph to be wearing a blue, short-sleeved T shirt, with 'Michigan State Athletic Squad 79' emblazoned on the chest. What appears to be a full circumference tattoo is visible in the area of his left bicep.

44 'TN', who was in a de facto relationship with the applicant between 1996 and 2002, gave evidence that the red Mini Minor that the applicant owned in 2000 did not have a glove box, but was instead equipped with a parcel tray. Moreover, she said the car's interior was black. (The significance of this evidence was, of course, twofold. First, GH had described as 'weird' the fact that her attacker's car had a 'glove box' which was 'full of condoms'. Secondly, GH agreed in cross-examination that the front and back seats were 'tan' coloured, as were 'the doors', 'the door trim' and 'the door panelling').

45 TN also gave evidence that in about August 2000, she and the applicant had purchased another car, a Mitsubishi Pajero. (A Road Corporation certificate, Exhibit 7, tendered during the evidence of the informant recorded that the applicant had acquired the Pajero on 14 August 2000). At that time, the Mini Minor was 'off the road' and 'on blocks'. Among the work being done, the Mini Minor was spray-painted white. Although it was not clear when the Mini Minor was again capable of being driven - TN gave evidence that the Mini was back on the road before Christmas 2000 and she was driving it by April 2001 - the evidence established that, from August 2000, it had been stripped back to bare metal, and had been painted 'pearl white'. We pause to note that, if that evidence was accurate, the applicant's Mini Minor could not have been the car driven by GH's attacker. It will be remembered that the offences were said to have occurred between 31 October and 1 December 2000, and GH was adamant that the car was 'red'.

46 For the sake of completeness we note that, in the course of the first trial, Associate Professor Richard Kemp gave evidence in support of the defence case. He was an 'experimental forensic psychologist', lecturing in the Master of Forensic Psychology program in the School of Psychology at the University of New South Wales.

47 Given the conclusions that we have reached regarding the identification evidence led in that first trial, it is unnecessary to set out Professor Kemp's evidence in detail. It is enough to observe that his evidence was directed to the frailties of identification evidence generally, including the capacity of memory to be altered by 'post-event information'.

48 The judge's decision as to whether to exclude GH's Facebook and photo board identification evidence was governed by s 137 of the Evidence Act.[14] That section provides that in a criminal proceeding 'the court must refuse to admit evidence adduced by the prosecutor if its probative value is outweighed by the danger of unfair prejudice to the accused'.

49 At the time that the judge ruled GH's identification evidence to be admissible, the approach to be taken to s 137 of the Evidence Act in this State was as laid down in Dupas v The Queen.[15]

50 In Dupas it was held that a trial judge undertaking the balancing task required by s 137 is obliged to assume that the jury will accept the evidence to be truthful, but is not required to make an assumption that the jury will accept it as reliable.[16] This represented a departure from the approach adopted in New South Wales in R v Shamouil.[17] In that case, it was held that a trial judge determining the probative value of evidence for the purposes of s 137 of the Evidence Act 1995 (NSW) should assume that the jury will accept the evidence. The trial judge should not, however, have regard to the credibility or reliability of the evidence.

51 Recently, in IMM v The Queen,[18] the High Court disapproved of that specific aspect of Dupas, and instead endorsed the approach set out in Shamouil.

52 In IMM, French CJ, Kiefel, Bell and Keane JJ jointly held that, when assessing the probative value of evidence for the purposes of s 137 (and several other provisions of the Act), a trial judge must proceed on the assumption that the jury will accept the evidence. No question can arise at that stage as to matters of credibility or reliability. Their Honours observed:

Once it is understood that an assumption as to the jury's acceptance of the evidence must be made, it follows that no question as to credibility of the evidence, or the witness giving it, can arise. For the same reason, no question as to the reliability of the evidence can arise. If the jury are to be taken to accept the evidence, they will be taken to accept it completely in proof of the facts stated. There can be no disaggregation of the two - reliability and credibility - as Dupas v The Queen may imply. They are both subsumed in the jury's acceptance of the evidence. [19]

It must also be understood that the basis upon which a trial judge proceeds, that the jury will accept the evidence taken at its highest, does not distort a finding as to the real probative value of the evidence. The circumstances surrounding the evidence may indicate that its highest level is not very high at all. The example given by J D Heydon QC[20] was of an identification made very briefly in foggy conditions and in bad light by a witness who did not know the person identified. As he points out, on one approach it is possible to say that taken at its highest it is as high as any other identification, and then look for particular weaknesses in the evidence (which would include reliability). On another approach, it is an identification, but a weak one because it is simply unconvincing. The former is the approach undertaken by the Victorian Court of Appeal; the latter by the New South Wales Court of Criminal Appeal. The point presently to be made is that it is the latter approach which the statute requires. This is the assessment undertaken by the trial judge of the probative value of the evidence. [21]

54 As may be gleaned from these passages, French CJ, Kiefel, Bell and Keane JJ were attracted to the example given by former Justice Heydon.[22] He had observed that the 'disputation between and within the intermediate appellate courts of New South Wales and Victoria is detailed', but the 'detail may obscure the possible fact that the gap is narrow'.[23]

55 Adopting the approach described by Heydon, and seemingly endorsed by the majority in IMM, GH's purported identification from Facebook was, in our view, not merely weak, but 'simply unconvincing'. Moreover, given the circumstances of the Facebook identification and the publicity surrounding the applicant's known involvement in the Jill Meagher case, the later photo board identification was virtually of no probative value whatever.

56 Any probative value that this identification evidence might theoretically have had was significantly outweighed by the risk to the applicant of unfair prejudice. With respect, we consider that the judge was wrong to have reached a different conclusion.

57 Before this Court, counsel for the applicant submitted that the problems inherent in suspect identification cases are well known to the law. It is recognised that wrongful convictions have resulted because even honest witnesses have confidently given mistaken identification evidence. The deficiencies in flawed identification evidence, it was submitted, could not be overcome by other circumstantial evidence, as to which in this case there was, in any event, none.

58 In particular, it was submitted that GH's evidence of identification had the following 'problematic features':

• first, GH initially purported to identify the applicant from a single image on Facebook;
• secondly, the Facebook photo was viewed some 12 years after the attack upon GH;
• thirdly, at the time that GH viewed the Facebook photo, she knew that the applicant had been charged with the rape and murder of Jill Meagher;
• fourthly, the photograph of the applicant on the Facebook page was taken about 11 years after the alleged offending;
• fifthly, the evidence of the photo board identification was exceedingly weak, since there was an obvious risk that GH selected the applicant's photo as a result of the 'displacement effect' or 'unconscious transference', she having been influenced by the Facebook image; and
• sixthly, GH's photo board identification was, in reality, evidence only of similarity, since she said, it '[d]efinitely looks like No. 4'.

59 It has been said that s 137 of the Evidence Act largely mirrors the Christie[24] discretion at common law.[25] Whether or not s 137 and that common law discretion are wholly coextensive, the kinds of considerations flowing from Christie have long been applied to identification evidence.

60 Hence, in Alexander v The Queen[26] Gibbs CJ observed:

The authorities support the conclusion that I have reached, which is that, as a matter of law, evidence of an identification made out of court by the use of photographs produced by the police is admissible. However, a trial judge has a discretion to exclude any evidence if the strict rules of admissibility operate unfairly against the accused. It would be right to exercise that discretion in any case in which the judge was of the opinion that the evidence had little weight but was likely to be gravely prejudicial to the accused.[27]

61 In our opinion, the probative value of GH's purported identification from Facebook was scant. That identification was made by GH from a single photograph, some 12 years after the attack upon her. The photograph itself had been taken 11 or so years after the offending. Further, the circumstances were highly suggestive in that, at the time that she made her purported identification, GH knew that the individual depicted in the photograph had been charged with both rape and murder.

62 The law has long set its face against identification evidence obtained from a single photograph. Thus in R v Simmonds,[28] in circumstances where a single photograph of a suspect had been presented to an identifying witness, it was said that the 'danger of misidentification in such a situation must be regarded as extremely high', and that the identification evidence was 'so inherently unreliable that it should have been withdrawn from the jury'. In such a situation, the risk of the 'displacement effect' was 'very high'.

63 In Alexander,[29] Stephen J described the 'displacement effect', to which we have made reference, in the following terms:

Having been shown a photograph, the memory of it may be more clearly retained than the memory of the original sighting of the offender and may, accordingly, displace that original memory. Any subsequent face-to-face identification, in court or in an identification parade, may, on the identifying witness's part, in truth involve a matching of the man so identified with the remembered photograph, which has displaced in his memory his recollection of the original sighting.[30]

64 Facebook, and other social media and networking sites have, in recent years, become a ubiquitous aspect of daily life. Potential witnesses are free to peruse photographs on Facebook (and similar sites) - without police oversight or other supervision - before attempting a formal identification process with law enforcement authorities. Accordingly, the risk of the displacement effect will, depending upon the circumstances, be exacerbated.

65 Certainly, none of the safeguards which ordinarily attend photographic identification performed under the supervision of police will be present.

66 In Strauss v The Police,[31] Peek J adverted to some of the difficulties attending identification evidence derived from Facebook. In such cases, the probative value of the identification evidence will, of course, as the authorities demonstrate, have to be considered according to the particular circumstances of each case.[32]

67 So far as the approach to evidence obtained by viewing photographs on the internet is concerned, a ruling delivered by Garling J, as the trial judge in R v Smith (No 3),[33] is instructive. In that case, a trial for murder, the prosecutor sought leave to lead identification evidence in the course of re-examination of a prosecution witness. It was not in issue in the trial that the accused had shot and killed the deceased, the principal issue being whether he acted in self-defence.

68 Shortly before the shooting, and in its near vicinity, the witness had seen a man wearing a blue 'hoodie' and jeans leaping a fence. Were it to be accepted that that man was the accused, the claim of self-defence would have been weakened.

69 The prosecution wanted to lead evidence that, in the course of the afternoon following the shooting, the witness, whilst at work and using his computer, saw an article on the MSN News site on the internet which contained a report of the shooting, together with a photograph of the accused. The photograph had been circulated by police as being that of a person whom they were seeking with respect to the shooting.

70 It was anticipated that the witness would give evidence to the effect that, when he saw that photograph on the internet, he recognised the accused as the person he had seen in the blue hoodie and jeans.

71 Garling J said generally of identification evidence:

Identification evidence of the kind and nature which is intended to be led, is notorious for the problems and difficulties which can arise. Such identification evidence can be unreliable. Caution is always to be exercised in admitting this evidence, and once admitted, juries must be given directions about using the evidence unhesitatingly: see s 116 of the Evidence Act 1995.

When a person sees, or is shown, a single photograph rather than a photo array from which one photograph is chosen, it is obvious that an identification arising from that single photograph may be unreliable. As well, where the photograph is associated with the description that the person is wanted by the police, albeit for the provision of assistance, those difficulties are magnified.

As Gibbs CJ said in Alexander v R [(1981) [1981] HCA 17; 145 CLR 395, 400], in relation to a single photograph identification process:

'... it would be unfair and improper to show to a witness, ... a single photograph of a person who is said to be the suspect, and it would be unsafe to act on evidence of identification given in those circumstances.'

Attention must also be paid to the judgment of Stephen J in Alexander at 409, where his Honour describes the phenomenon known as the 'displacement effect' which is relevant here. ...

It is the existence of this effect upon which the accused relies here, even though [the witness] may not accept its existence.[34]

In this case, because [the witness] had seen a single photograph which identified a person, shortly after the shooting, whom the police were seeking to interview, there is a real risk that [the witness] has, entirely unconsciously, substituted that image in the photograph for the image in his mind of the face of the person whom he saw in the blue hoodie. After all, he did not see anyone else who could have been the shooter, and [the witness] also saw the person leaping the fence at the scene.

Accordingly, it is easy to see that, quite unintentionally, [the witness] may have associated the male in the blue hoodie as being the shooter, and that he has then associated the photograph circulated by the police as being that person.

Because the association caused by the displacement effect is unintentional, it cannot be expected that [the witness], if cross-examined about it, would necessarily understand what has occurred. Accordingly, any cross-examination of [the witness] by counsel for the accused is more likely than not to result in an affirmation of the evidence of identification, rather than anything else.[35]

The identification of the male figure is a piece of evidence of relevance. Accordingly, assuming that it was an appropriate piece of evidence, and in admissible form, the Crown would be entitled to lead it.

However, having regard to the nature of the evidence, its relatively low probative value, and the very high risk of unfair prejudice to the accused, in light of the matters which I have discussed above, I have not been persuaded that I should exercise my discretion to permit the Crown to lead it.

On the contrary, I am persuaded that by the application of s 137 of the Evidence Act, I must reject the evidence because the danger of unfair prejudice to the accused from the admission of such evidence, far outweighs any probative value which the evidence may have for the prosecution case. It follows that leave to adduce the evidence must be refused because the evidence is inadmissible.[36]

74 Smith is one example, among many, where identification evidence has been rejected pursuant to s 137, because of the inherent defects in the nature of the identification carried out. In that case, a witness had viewed a single photograph on the internet in a setting that strongly suggested that the person depicted was wanted by police. As Garling J made clear, because the association caused by the displacement effect is unintentional, cross-examination is more likely than not to result in an affirmation, rather than a rejection, of the identification.

75 Garling J's ruling in Smith represents a recent application of well-established doctrine. In general terms, identification evidence of a questionable nature should not be admitted because there is a real risk that, irrespective of the strength of any warning that might be given to the jury as to its possible infirmity, it will be accorded more weight than it actually merits.

76 There are many examples, at trial level, of weak identification evidence having been excluded rather than left to the jury, with strong warnings, for their consideration.[37] There are also examples at appellate level of such evidence having been said to be inadmissible.[38]

77 As long ago as Davies and Cody v The King,[39] the High Court held that where evidence of the identity of the accused person is given by a witness whose previous knowledge has not made that witness familiar with the appearance of the accused, and where the witness has been shown the accused alone as a suspect and has on that occasion first identified him or her, a Court of Criminal Appeal should quash the conviction as unsafe unless his or her identity is further proved by other evidence, direct or circumstantial.

78 Turning to examples of more recent authority, in R v Blick[40] a witness had been shown a group of photographs in which only one was of a man with a goatee beard. That beard was a pivotal identifying factor in the witness's mind. The trial judge had, somewhat surprisingly, permitted that identification evidence to be led. The New South Wales Court of Criminal Appeal quashed the conviction, holding that that the photographic evidence had little probative value, but 'substantial' prejudicial effect.

79 A similar conclusion was reached by that Court in R v Fisher.[41] In that case, the appellant had been convicted of assault with intent to rob. Two witnesses had identified the appellant at an identification parade in which the appellant was the only person to match the description given by witnesses of the offender. The trial judge had ruled that evidence admissible. Sheller JA (with whom Sully and Hulme JJ agreed) considered that the evidence should have been excluded pursuant to s 137. Its admission resulting in a substantial miscarriage of justice, the conviction was quashed and a new trial ordered.

80 In R v Marshall,[42] the appellant had been convicted on a charge of armed robbery. He had elected to be tried by judge alone. On appeal against conviction, the issue was whether the evidence of the victim, who identified the appellant as the offender while he was seen to be in police custody, should have been excluded. There was also discussion of the 'displacement effect'.

81 The appeal was allowed. Spigelman CJ (with whom Heydon JA and James J agreed) noted that, without the victim's identification evidence, there was no case against the appellant. The issue raised on the voir dire had been whether or not that evidence should be excluded under s 137 of the Evidence Act 1995 (NSW).

82 Chief Justice Spigelman observed that the prejudice often associated with identification evidence was that, though mistaken, it was frequently given great force and assurance by the person making the identification. It appeared that the trial judge had been impressed by the confidence with which the victim made their identification. However, the danger of unfair prejudice in a case such as this was particularly strong.

83 Moreover, as the Chief Justice pointed out, the identification was flawed because of the 'displacement effect'. The victim had seen several photographs of the accused on the morning that he made his identification. Interestingly, when shown a photo board which included the appellant's photograph, the victim could not identify the offender. However, shortly afterwards, upon seeing the appellant in a single suspect situation at the police station, he indicated that he was able to do so.

84 The combination of a single suspect identification and the displacement effect, coupled with the absence of any additional evidence implicating the appellant, meant that the trial had miscarried. The identification evidence was of limited probative value, and of a high order of prejudice. It should have been excluded.[43]

85 To similar effect was R v Clark,[44] a decision of the South Australian Court of Criminal Appeal. In that case, the sole issue at trial was one of identification of four offenders. The victims of the particular offence identified the appellants from a book of 250 photographs that had been prepared by police. They also identified them in the dock at trial.

86 It was held, allowing the appeal by one of the alleged offenders, that the verdict against him was unsafe and unsatisfactory. That was because the identification evidence against that appellant was so 'sketchy' as to render the verdict unsafe.

87 The Court in that case accepted that a dock identification, generally, had little or no probative value. However, such an identification might be permitted in circumstances where the witness had earlier identified the accused in the course of a physical line up or, failing that, a properly conducted photographic identification process.

88 The one appellant whose conviction was overturned in Clark was able to call in aid a series of significant discrepancies between the description given by the witnesses of the particular offender alleged by the Crown to be him, and his actual physical characteristics. These discrepancies were so potent as to enable Cox J (with whom Perry and Lander JJ relevantly agreed) to conclude that this particular conviction could not stand. Implicitly at least, it seems that his Honour considered that the judge should have excluded the identification evidence in the exercise of his discretion.

89 There are a number of other cases, at appellate level, that cast doubt upon the admissibility of questionable identification evidence. See, for example, R v Burchielli[45] (single suspect identification); Corke v The Queen[46] (single suspect identification); R v Scifleet[47] (in-court single suspect identification); and Grbic v Pitkethly[48] (in-court single suspect identification).

90 Returning to the present case, the circumstances of GH's evidence of identification from the Facebook photograph are similar in some respects to those that arose in Smith.[49] Of course, the fact that such evidence was excluded in that case is in no way determinative of whether it was rightly admitted in this case. Smith does, however, illustrate that there is nothing whatsoever novel about the rejection, in the exercise of discretion, of suspect identification evidence gleaned from the internet.

91 It is true that in R v Dupas (No 3),[50] Weinberg JA remarked that there are instances in which identification evidence that could be said to be questionable had nonetheless been admitted.[51] His Honour regarded Domican v The Queen[52] as providing a 'paradigm example'.

92 In the context of alluding to the dangers of admitting weak evidence of identification, his Honour made reference to R v Jamal.[53] That was a case involving dock identification which as we have said is, in the absence of prior out of court identification, ordinarily devoid of any probative value. His Honour observed that in some cases where dock identification had been permitted, it had been held that the giving of appropriate directions might be an adequate safeguard against the possible misuse by the jury of such evidence.[54]

93 There are indeed instances of dock identifications having been made without any previous out of court identifications, or other such questionable identifications being made, and the evidence nonetheless having been held to be admissible. It must be said that, in modern times, such cases are the exception.[55] Weinberg JA in Dupas (No 3) did not suggest that evidence of that kind should be admitted as a matter of course. Whether or not a dock identification or evidence of a similarly questionable nature should be admitted must, ultimately, depend upon the particular circumstances of each case.

94 GH's evidence regarding the Facebook photograph was in some respects no better than a dock identification. Indeed, it could reasonably be viewed as worse. She was faced with a single photograph in circumstances that were themselves highly suggestive of guilt of the most serious crimes imaginable. Evidence of that kind has often been excluded and is commonly regarded as insufficient, on its own, to support a conviction.

95 In R v Preston,[56] David J remarked:

There may be cases where an in-dock identification is appropriate, but only where there has been some form of positive identification on a previous occasion and the in-dock identification merely relates that previous identification to the case at hand. That is not the case here. There was no previous positive identification, merely evidence that amounted to a piece of circumstantial evidence. Therefore, the only positive identification led as part of the prosecution case against the appellant was that in-dock identification. The evidence is clearly valueless and, if objected to at the time, should have been excluded.[57]

96 Moreover, GH's subsequent photo board identification of the applicant was, if anything, even less worthy of credence than her earlier Facebook identification. By that stage, the risk of the displacement effect had to have been acute. By the time GH came to view the photo board, she had seen the applicant's image in the media repeatedly, and she knew that the applicant, as well as having been charged in relation to the rape and murder of Jill Meagher, had also been charged with her rape.

97 Given the circumstances under which GH made her two identifications, there was a serious risk that the jury, no matter how well directed, would fail to appreciate just how devoid of substance this evidence truly was. The danger of unfair prejudice was substantial. For these reasons, the evidence should, in our opinion, have been excluded.

98 Even if, contrary to the view that we have expressed above, GH's evidence was properly received,[58] we would hold that the verdicts in the first trial are unsafe and unsatisfactory.

99 Having made our own independent assessment of the evidence as a whole (including what seems to us to have been credible evidence given by TN as to the applicant's Mini Minor), we are firmly of the view that a properly instructed jury, applying themselves appropriately to their task, must have had a reasonable doubt as to the applicant's guilt. GH's identification evidence was so weak[59] that no other conclusion was reasonably open.

100 It follows that the convictions in the first trial must be set aside as unsafe and unsatisfactory. A judgment and verdict of acquittal must be entered with respect to each charge on the indictment in that trial.

The conviction application on the third trial

101 In April 2012, HS was living at a 'backpacker place' in Balaclava with 17 others.

102 On Saturday, 14 July 2012, at about 11.00 pm, HS went with two 'flatmates' to the 'Elephant & Wheelbarrow' Hotel in Fitzroy Street, St Kilda, to meet some other 'housemates'. Before leaving for the hotel she had 'a couple of wines', and while at the hotel she had a 'couple of beers'. She left the hotel alone at 'something like' 2.30 am. When she left the hotel she was 'a bit tipsy' or 'drunkish', since she 'had a little bit to drink'.

103 HS gave evidence that she walked down Fitzroy Street to the Esplanade where it turns into Carlisle Street. When she was about 20 metres from St Kilda Road, a car 'pulled out' and stopped. The front passenger door opened and a man sitting in the car started talking to her. He pointed to a car in front which was 'standing there'. HS said they 'were both small cars' and 'didn't have like a big boot at the back, just like a stump', 'like a Volkswagen Polo'. In cross-examination she agreed that in her statement to police she had said that the car that she got into 'was round at the back, it did not have a boot'. She thought that the car that pulled out was red, but said that both cars were dark.

104 HS gave evidence that the man said, 'Okay, well that car in front of you is following you so I can give you a ride safe home'. Eventually, she sat in what she had described as the 'red car'. By this time it was about 3.00 am. The interior of the car 'was really new, it smelt new'. HS said that the 'seats and everything' were black in colour, and in the 'middle' there was (she thought) an analogue clock. In cross-examination, HS agreed that when she made her statement on 17 July 2012, she had told police that there was a digital clock in the car, closer to the driver than to her, with 'a blue screen with dark numbers'. She also agreed that she had told police that she did not think that the dashboard had an analogue clock.

105 The man then drove the car off Carlisle Street into Westbury Street, and into 'like a parking place away from the road'. While HS sat there, 'all of a sudden' the man 'jumped at' her. He 'jumped' right at her face and started kissing her, and he put his tongue in her mouth. HS was 'flabbergasted' and 'in distress'. It 'all happened very fast'. The man 'just came right at [her]'. HS screamed at the man to let her go and 'struggled against him'. She tried to open the door with her left hand, but the man grabbed her hands and 'really clutched' her into her seat. The man then hit her to the left side of her face.

106 HS testified that the man then put his hand over her mouth because she was 'yelling and screaming'. He also grabbed her by the throat to hold her back, and told her that she could not get out and no one would hear her screams. The man took her passport from her and put it where she could not grab it easily. He lifted her top, unbuttoned her jeans and pulled them partially down. HS was trying to push him off. The man then took her hand, placed it on his penis and made her masturbate him for a 'few seconds'. He then told her to take the right leg out of her jeans.

107 Once HS got her leg out of her jeans he pulled her underpants down. The man was 'aggressive'. He penetrated her vagina with his penis for 'five minutes or so'. So as to avoid further aggressive behaviour, HS pretended willingly to take part. She suggested that they go back to her home, telling him 'the car was not a good place to do this'. The man withdrew his penis. He told her she could not put her jeans back on, and draped the loose leg of the jeans over her leg.

108 They then drove to where HS was staying. The man allowed HS to put her jeans and underpants back on. When they got out of the car, the man grabbed HS's left hand with his right and walked her towards the house. HS then opened the front door and ran inside 'straight to the bathroom'. Some of her flatmates were in the living room. The man had come inside a 'few steps'. That was the last she saw of him.

109 Another occupant, 'SC', came into the bathroom where HS was crying. HS told her that a man had 'forced himself' on her and she thought she was going to die. She told other flatmates the same thing.

110 The police were called. HS told them that the man had 'force fucked' her. She gave evidence that she described the man as 'an Australian man, around 35 years old', of '[m]edium build, not too high ... about 170 to 172 [centimetres]'. HS said 'he was Caucasian, I think, like white skin, blonde hair, clean shaven'. She also said the man had 'gel' in his hair, which was 'spiked'. In cross-examination, she agreed that she had also described the man to police as having 'blonde short hair, about two centimetres in length', and told the police that 'his head was sort of squared shaped'.

111 Constable Elizabeth Tonkin obtained a description from HS of her attacker, which included that he was wearing a dark grey or black jacket. It should also be noted that SC, who had briefly seen the man, described him as wearing a 'grey blazer'.

112 On 17 July 2012, two days after HS was attacked, she told a police officer, Leading Senior Constable Andrew Grutzner, who compiled a 'FACE'[60] image of HS's assailant (Exhibit B), that her attacker 'had acne scars on his cheeks, not too much of it'. HS could not remember any tattoos or scars. The man was wearing something like blue jeans and a shirt. HS said that the image compiled was '65 per cent like the person who had assaulted [her]'.

113 Some six months later, on 27 February 2013, police showed HS eight photographs of car dashboards (Exhibit A). HS was unable to pick any photo which matched her memory of the inside of the car.

114 Also on 27 February 2013, Detective Senior Constable Christine Stafford showed HS a photo board (Exhibit C) containing 12 photographs of males. HS selected photograph '5', which was a photograph of the applicant taken on 16 August 2011. In cross-examination, HS agreed that she told police at the time, 'That looks like him ... He looks most like the guy ... 75 to 80 per cent sure'.

115 Some of the other evidence established that, on 14 July 2012, the applicant went to the football with his girlfriend, 'RA', together with a friend, 'DF', and others. They left the football at the end of the match and tried to get back into a bar at which they had been drinking earlier. From there they went to the applicant's car - a dark aqua-blue Holden Statesman - for about ten minutes, before walking back to the Post Office Hotel in Swan Street, Richmond. They remained there until the hotel was shut.

116 DF recalled the applicant as having worn a dark coloured leather jacket at the start of the night.

117 The applicant and RA were arguing. The applicant drove RA home to Coburg in the early hours of the morning. She told the applicant that he was not welcome in their bed, and the applicant went outside. The next morning, when RA awoke, the applicant was asleep on the floor. He said that he had slept in his car.

118 Exhibits Q and R, to which we will return, showed calls between the applicant's and RA's mobile telephones between 2.42.01 am and 2.57.07 am on 15 July 2012. An analyst from Vodafone, Petine Tuhukava, produced Exhibit Q (which was a record of the applicant's telephone calls between 14 July and 15 July 2012) and Exhibit R (a record of RA's telephone calls for the same period).

119 Ms Tuhukava gave evidence that Exhibit Q contained a location code for the cell tower that took each telephone call. Each cell tower has its own individual code. A mobile phone on the Vodafone network constantly monitors the strongest cell sites within range of the telephone. Cell sites are situated on cell towers. Multiple cell sites can be on the one tower. Each cell site can have one, two or three antennae. In the city, cell sites are designed in a network and cell sites overlap so far as their range of coverage is concerned.

120 In an urban environment you would expect a phone to be within range of more than one tower. A call connects to the strongest signal available to it at the time. The only site or tower that gets listed in the records is the cell site to which the phone first connects, the strongest signal at that time. Typically, the range of a cell site in the Vodafone network can be between two and 35 kilometres. In suburban or urban areas the average range of a cell site is between two and 15 kilometres.

121 Doing the maths, if one assumes a range of 15 kilometres, that equals a coverage area of 706 square kilometres. Ms Tuhukava could not say from looking at the location description records what the range of any of the relevant antennae or cell sites was as at July 2012.

122 Detective Julio Salerno, the informant, collected CCTV footage, including from outside the Elephant & Wheelbarrow Hotel. Having studied the footage with a view to locating any red cars that might have been in the vicinity at the relevant time, he found two vehicles which appeared to be red in colour, one travelling east along Carlisle Street at 3.40 am and the second at 3.54 am. Plainly, given the time at which these vehicles were seen, neither could have been the vehicle in which HS had been attacked.

123 Footage from the Elephant & Wheelbarrow Hotel also showed, at the lights closest to the hotel, at 2.57 am, what Detective Salerno believed to be a blue Holden Statesman. The footage showing the presence of that car was pointed out to the jury.

Were the telephone records and CCTV footage relevant?

124 It is convenient first to turn to ground two, which asserts that the judge erred in admitting the telephone records and the CCTV footage depicting a dark sedan outside the Elephant & Wheelbarrow Hotel.

125 Counsel for the applicant at trial unsuccessfully sought exclusion of the evidence contained in the telephone records (which ultimately became Exhibits Q and R) and of the CCTV footage (including that taken outside the Elephant & Wheelbarrow Hotel).

126 Before this Court, counsel initially sought to press both limbs of ground two. However, in the course of oral argument he made a partial concession:

But in respect to the CCTV footage we don't press that ground of appeal at all in relation to the CCTV footage. I cannot sensibly mount an argument that that is not relevant, in terms of s 55 [of the Evidence Act].

127 In an endeavour to support what remained of ground two, counsel for the applicant relied on s 55(1) of the Evidence Act. That section provides:

The evidence that is relevant in a proceeding is evidence that, if it were accepted, could rationally affect (directly or indirectly) the assessment of the probability of the existence of a fact in issue in the proceeding.

128 As will become clear when we come to deal with ground one, we consider that the telephone records were clearly relevant to a fact in issue, that fact being the applicant's opportunity to commit the crimes with which he was charged. At the very least, the records were capable of supporting an inference that the applicant was not at home between 2.42 am and 2.57 am, the attack on HS having occurred at approximately 3.00 am.

129 Moreover, although that part of ground two which related to the CCTV footage was abandoned, we should observe that, when taken as part of the overall circumstantial case, the CCTV footage was capable of supporting an inference that the dark car shown outside the Elephant & Wheelbarrow Hotel was that belonging to the applicant.

130 In determining whether evidence might rationally affect the assessment of the probability of the existence of a fact in issue - directly or indirectly - it is wrong to focus upon whether a piece of evidence, considered in isolation, may prove anything. One must also look to the other evidence in the case.

131 As Heydon J observed in Evans v The Queen:[61]

The relevance of evidence does not depend on its capacity by itself to prove the prosecution case on a particular issue, or to raise a reasonable doubt in favour of the defence on that issue. The effect on assessing probability which is to be looked for is the effect of the contested evidence taken with other evidence either admitted by the time the controversial evidence is tendered, or to be called.[62]

132 Viewed in isolation, the evidence of the car shown in the CCTV footage may not have been particularly powerful. However, when taken with the other evidence, it was capable of bearing on the identified fact in issue. Hence, counsel's concession as to relevance was properly made.

133 For the sake of completeness, we note that the trial judge gave the following directions about the impugned evidence:

I must give you three directions in relation to the CCTV footage and the mobile phone records. (1) You can only use evidence of the CCTV footage of the car outside the Elephant & Wheelbarrow to support the conclusion that Mr Bayley was in the vicinity of the scene of the offence if you are satisfied that the car in the CCTV footage is Mr Bayley's car. It is not enough for you to be satisfied it is similar to Mr Bayley's car or even of the same make and model. (2) You cannot use the evidence of mobile phone records to 'track' Mr Bayley's movements. The evidence given by Ms Tuhukava only permits the conclusion that Mr Bayley's phone was within the range of the cell tower noted in the records. (3) You cannot use the conclusion that the Crown asks you to draw, namely that Mr Bayley was in the vicinity of the scene of the offence, such that he had the opportunity to commit it, unless you are satisfied of that conclusion beyond reasonable doubt.

134 In our view, and in the reasons that follow, it will be made clear why the evidence of the telephone records was clearly relevant. What is left of ground two cannot be upheld.

135 The approach to be taken by an appellate court to a claim that a verdict is unsafe or unsatisfactory is governed by well-established principles. The court must ask whether it was open to the jury to be satisfied beyond reasonable doubt of the applicant's guilt.[63]

136 In determining whether it was open to the jury to be satisfied of guilt, the court must make its own independent assessment of the evidence.[64] Indeed, it is wrong for the court to concern itself with 'whether, as a question of law, there was evidence to support the verdicts, rather than making its own independent assessment of the evidence'.[65] But for the applicant to establish that a verdict is unsafe or unsatisfactory, it will not be sufficient merely to show that there was material which might have led the jury to entertain a reasonable doubt about guilt.[66] The relevant enquiry is whether the jury must, as distinct from might, have entertained a reasonable doubt as to guilt.[67]

137 Counsel for the applicant submitted that this Court, having made its own independent assessment of the evidence, ought to entertain a reasonable doubt as to the applicant's guilt. As will become clear, however, having made our own independent assessment of the evidence, we entertain no such doubt. In our opinion, it was properly open to the jury to return the verdicts that they did.

138 In support of the contention that the verdicts are unsafe or unsatisfactory, the applicant's counsel submitted that, taken at its highest, the photo board evidence was merely evidence of similarity, and not evidence of positive identification. It was submitted that, on its own therefore, as a matter of law, it was incapable of sustaining a guilty verdict.[68]

139 Moreover, it was submitted that the evidence of opportunity was 'very weak'. It relied on telephone records and CCTV footage of a dark sedan (arguably a Holden Statesman) outside the Elephant & Wheelbarrow Hotel, seemingly recorded around the time that HS had left the hotel. It was submitted that there were no features of the car depicted in the footage that clearly identified it as being the applicant's car.

140 Expert evidence established that the telephone records went no further than showing that the applicant was within 15 kilometres of St Kilda when the offender approached HS.

141 Further, the applicant's counsel submitted that the evidence positively suggested that the applicant was not the offender. HS said that she was attacked by the driver of a small red car which she said was like a Volkswagon Polo. She said the car was round at the back; did not have a boot; was 'kind of new'; had all black material inside; and had a digital clock which had blue backlit lighting and dark numbers.

142 On the other hand, the applicant drove a blue Holden Statesman with a large boot. In 2012, the vehicle was 15 years old. It had a grey interior, and wood-like panelling. It also had an analogue clock. Counsel pointed out that the evidence showed that the police had conducted a media conference on 1 August 2012, during which it was asserted that HS's attacker had been driving a red car. Thus, so it was submitted, the evidence demonstrated that the applicant's car was not that driven by the offender.

143 Counsel for the applicant also submitted that HS had described her attacker as having a square shaped head, whereas he argued that the applicant had a triangular face. HS had also described her attacker as having some acne scarring, yet there was no evidence that the applicant had ever had any such scarring.

144 Additionally, SC described the offender as wearing a 'grey blazer'. The evidence was that the applicant was wearing a dark coloured leather jacket on the night HS was attacked. Police located a black leather jacket when they searched the applicant's home, but they did not find a grey blazer.

145 In response, counsel for the Crown acknowledged that the prosecution case was largely circumstantial, but submitted that it was nonetheless strong. It was submitted that, although certain items of evidence, viewed in isolation, may have been insufficient to prove guilt, the combined strength of the evidence was such that there was ample justification for the jury to have concluded that the applicant was the offender.

146 Several aspects of the evidence were relied upon in that regard.

147 First, counsel contended that HS's description of her attacker as an Australian man; around 35 years old; of medium build; not 'too high', around 170 to 172 centimetres; Caucasian; with white skin; clean shaven; and with blonde spikey hair, was generally consistent with the applicant's appearance.

148 The jury had available to them Exhibit P, a blown up image of the photograph of the applicant from the photo board that had been taken closer in time to the offences. It was submitted that it was well open to the jury to find that HS's description of the offender was consistent with the appearance of the applicant at the relevant time.

149 It was next submitted that the fact that HS described to police that the offender's head was 'sort of square shaped' did not diminish the strength of the identification evidence. Indeed, it was open to the jury to find that the FACE image that HS compiled with police, which she described as being 65 per cent like the person who assaulted her, depicted a person with more of a triangular shaped face, quite similar to that of the applicant.

150 The Crown further submitted that the description of the offender as having 'acne scars on his cheeks, not too much of it', was of little significance, since the jury might reasonably have concluded either that HS was mistaken about that matter, or that such a feature was not a 'particularly identifying feature' in the context of the identification evidence as a whole.

151 In the same way, it was submitted that the description of the offender as having worn a grey jacket or blazer should not lead to the conclusion that the applicant was not correctly identified as the offender. The only description of the jacket as having been a 'grey blazer' came from SC. Her observations were made when she had just woken up, and she had only a fleeting opportunity to observe the offender.

152 It was submitted that it was open to the jury to prefer HS's description of the offender's clothing given that she had a much longer and better opportunity to make her observations.

153 Ultimately, it was submitted that the jury may have concluded that the difference between a description of a black jacket and a grey jacket was immaterial, since HS had told the police that the offender was wearing a dark grey or black jacket. It was open to the jury to find that the black leather jacket said to have been worn by the applicant earlier in the night, and later found by police, was generally consistent with the description of the jacket given by both HS and SC.

154 Next, it was submitted that, even assuming that HS's identification from the photo board amounted to evidence of 'similarity' rather than evidence of positive identification, the evidence was strongly probative as part of the overall circumstantial case. The jury might have regarded the evidence of similarity as particularly strong because of the manner in which HS described her level of certainty. She told police that she was 75 to 80 per cent certain 'that the image was of him'. It was submitted that this was not merely a vague assertion that the photograph resembled the offender, but rather an expression of a high level of certainty.

155 Counsel next submitted that a further important strand of the prosecution case was the evidence of opportunity. Telephone records indicated that the applicant was not at home at the relevant time, since he was telephoning his girlfriend who (it was agreed) was at home in Coburg. At that time, telephone records reflected the applicant's phone as connecting from towers located at Fitzroy West, Fitzroy and St Kilda North.

156 Further, a car similar to that driven by the applicant could be seen on CCTV footage at the relevant time and in the relevant vicinity. It was open to the jury to examine the CCTV footage and conclude that the vehicle there depicted was the applicant's vehicle.

157 Insofar as HS had described a small red car which did not match the description of the applicant's car, counsel submitted that it was open to the jury to conclude that HS was mistaken about the make and model of the car in which she was abducted. HS had explained during her evidence that she had been drinking and was more focussed on the violence being done to her rather than the details of the vehicle. With respect to the dashboard, it was open to the jury to conclude that HS might have been mistaken about 'peripheral matters' such as the precise colour of the exterior and interior of the vehicle, and layout of the dashboard.

158 In our opinion, the respondent's submissions should be accepted. Having made our own independent assessment of the evidence, we have concluded that it was well open to the jury - on all of the evidence - to convict the applicant. There are at least five aspects of the evidence which, when taken in combination, are important in leading us to that conclusion.

159 First, and significantly, the telephone records show that on 15 July 2012, a total of 15 calls were made from the applicant's mobile telephone to his girlfriend, RA's, telephone, between 2.42.01 am and 2.57.07 am, those calls having been recorded by cell towers at various locations. The time of the calls, the locations of the cell towers[69] and the duration of the calls, are:

160 Significant also is the fact that the records show that, after the call at 2.57.07 am, there was no activity on the applicant's telephone until 4.13.42 am, when a text message was sent from his telephone to RA's telephone. Thus, despite the frantic activity between 2.42.01 am and 2.57.07 am, the applicant did not use his telephone for a period of over an hour, between 2.57.07 am and 4.13.42 am. The rape of HS occurred, of course, within that period.

161 Telephone records also show that between 2.43.33 am and 3.30.58 am, some 19 calls which were not answered (and diverted to voicemail) were made from RA's telephone to the applicant's telephone. There were also eight text messages sent from RA's telephone to the applicant's telephone between 3.15.08 am and 3.41.57 am.

162 The telephone records amply demonstrate that the applicant had the opportunity to commit the relevant offences. Clearly, he was not at home in Coburg, and was, at the very least, in a position to be in the St Kilda area at about 3.00 am.

163 More importantly, however, the records are capable of founding an inference that the applicant was either unable or unwilling to use his telephone for a little over an hour (between 2.57.07 am and 4.13.42 am), during the period when HS was attacked.

164 Therefore, at a time when most people would be expected to be asleep, the applicant was up and about, making calls and texting repeatedly, at all but the very time of the commission of the rape. That could be coincidence,[70] but it is a matter that, at the very least, calls for some explanation.[71] Of course, only the applicant could give such an explanation, and it goes without saying that none was ever proffered.

165 Secondly, a car fitting the general description of the applicant's vehicle, a dark Holden Statesman, is shown to be in close proximity to the Elephant & Wheelbarrow Hotel, very near to the time when the attack occurred. Although, perhaps, of limited probative value when considered alone, the CCTV footage adds something to the overall circumstantial mosaic when viewed with all of the other evidence in the prosecution case.

166 Thirdly, the description given to police within a very short time of the attack largely coincides with the applicant's appearance. Two days later, at 9.00 am on 17 July 2012, HS gave a further description to Leading Senior Constable Andrew Grutzner, of the Criminal Identification Unit of Victoria Police. Again - with, perhaps, the exception of the reference to the acne scarring - the description is generally consistent with the appearance of the applicant.

167 Fourthly, as a matter of our own impression, the FACE image that Leading Senior Constable Grutzner produced bears a telling resemblance to the applicant. That fact, when coupled with the description given by HS, to which we have just referred, provides support for the conclusion that the applicant was the offender who had attacked HS. That is so notwithstanding the fact that HS could only say the image was '65 per cent like' the person who had assaulted her.

168 Fifthly, on 27 February 2013, HS selected the applicant's photograph from a photo board containing 12 photographs. At the risk of repetition, HS told police at the time, the selected photograph 'looks like him' and is 'most like the guy'. Indeed, she was '75 to 80 per cent sure' that the photograph was that of her attacker. And it will be remembered that the relevant photograph of the applicant had been taken on 16 August 2011, less than a year before the attack.

169 In our view, the combination of these five factors makes a convincing case for guilt.

170 As we have mentioned, the applicant's counsel drew attention to a number of factors which, so it was submitted, compelled the conclusion that the jury must have had a reasonable doubt as to the applicant's guilt.

171 Most prominent among those factors was HS's description of a small red car, without a boot, having a black interior and a distinctive clock. This was in circumstances where the applicant drove a large blue car with a substantial boot and an interior different to that described by the victim.

172 In our opinion, however, it was open to the jury to reason that HS, being significantly affected by alcohol, and not immediately realising that the person who offered her a ride posed a threat, was not concerned to take any particular notice of the car into which she got (or of any of its attributes). The jury might well have reasoned that once she was set upon, however, her attention was focussed, so that she was likely to have taken particular notice of the features of the man who was attacking her. Indeed, it would seem that HS had ample opportunity to observe her assailant's face.

173 The applicant's counsel also put store in the fact that HS had described her attacker as having acne scarring. There was no evidence that the applicant had any such scarring. In our view, however, it was open to the jury to conclude that this detail might have been explained by the lighting. HS gave evidence: 'Well I saw his face clearly in the car, but it was still a bit dark, but ah towards my house I could see in the light, I could have a better look at him, yeah'.

174 The jury might well have thought that although the lighting in the car was sufficient to enable HS to observe her attacker's face - which was very close to hers in a confined space - it may have affected her capacity to make accurate observations about fine detail, such as acne scarring.

175 Finally, in our view the evidence about the jacket worn by the offender adds little to either the guilt, or non-guilt, side of the ledger. It can be said, however, that the jacket worn by the applicant (and the jacket later located by police) is not inconsistent with the description of that worn by the offender. SC's reference to the offender having worn a 'grey blazer' does not strike us as particularly compelling one way or the other.

176 For these reasons, we are satisfied that it was properly open to the jury to be satisfied of the applicant's guilt. We ourselves entertain no reasonable doubt as to that guilt. The verdicts are not unsafe or unsatisfactory.

177 The applicant should be granted leave to appeal against conviction in relation to the third trial. However, the appeal itself should be dismissed.

178 Given our conclusion that the convictions in the first trial must be set aside, it is necessary to consider anew the applicant's sentencing.

179 In structuring the total effective sentence for the convictions returned in all three trials, the judge directed that charge 5 on the indictment in the third trial[72] was to be the 'base' sentence. The total cumulation of sentences in the first trial[73] on the base sentence is five years and three months' imprisonment.

180 Although the Court has power to vary the sentences imposed following the second and third trials,[74] we see no reason to disturb either the individual sentences, or the orders for cumulation, made by the trial judge.

181 In the result, the total effective sentence on the remaining charges will be 12 years and nine months' imprisonment. The quashing of the sentences relating to the first trial, however, makes it necessary to consider afresh the fixing of a new non-parole period.

182 There is no need to rehearse the circumstances of the offending against HS which was the subject of the third trial. These have been fully set out in our analysis of the relevant application concerning conviction. It is sufficient to observe that the offending against HS was a protracted violent sexual attack perpetrated against a vulnerable young woman, committed at a time when the applicant was on parole for earlier sexual offences, and on appeal bail for other violent offending.

183 It is necessary to set out, briefly, the circumstances of the offending against BC (in the second trial). In her sentencing remarks, the judge described that offending in the following terms:

In the early hours of 5 April 2012, [BC] was working as a sex worker. You drove up to her in a white Holden Astra motor vehicle. She got into your car for what she understood was to be a sexual service for money. You drove your car with [BC] in it into a dead-end alleyway and parked your car next to a fence. You parked the passenger side of your car close to that fence, described by [BC] as about 30 centimetres between the car and the fence. She was unable to open her door to get out.

[BC] described seeing your face all of a sudden, when you jumped over onto her. She started shaking, stuttering 'No, no, no', and 'What the hell?', and she started screaming. You grabbed hold of her and adjusted the seat that she was on to make it lie down. [BC] yelled out 'Help, help, fire', to try and get attention. You started to get rough and closer to her. She was scared and felt you were strong. She knew she was overpowered and that you had the upper hand. This deprivation of her liberty by trapping her in the car, as she described, was relied upon by the prosecution as false imprisonment (charge 1).

Regarding charge 2 [rape], [BC] said that she could not believe what was happening when you were trying to put your penis into her vagina. She was asking you to use a condom, 'Please use a condom', pleading with you, saying 'No, no, what are you doing?' and 'Help'. You raped her by putting your penis into her vagina. You told her to shut up. She was crying 'Help'. [BC] described being scared, and that it was the 'worst moment of her life'. At one stage you took her handbag and put it under the steering wheel on your side of the floor. You told her she was 'silly' and 'a stupid girl' for not having a spotter, pimp or someone to watch out for her.[75]

184 As we have observed, the offences against both BC and HS occurred when the applicant was on parole and on appeal bail.

185 On 26 April 2002, in relation to charges of rape against five victims, the applicant had been sentenced to a total effective sentence of 11 years' imprisonment with a non-parole period of eight years. He was released on parole on 17 March 2010. Whilst on parole, on 12 August 2011, he engaged in conduct which led to the charge of recklessly causing serious injury. On 27 February 2012, the applicant was sentenced in the Magistrates' Court to three months' imprisonment on that charge. He lodged an appeal to the County Court and was given appeal bail. Then, whilst on appeal bail (and parole), on 5 April 2012, the applicant committed the offences of false imprisonment and rape on BC; and, on 15 July 2012, he committed the offences of false imprisonment, indecent assault, common assault and rape against HS.

186 Prior to the commission of the offences for which he had been granted parole, the applicant had committed other disturbing sexual offences. On three occasions - 8 June 1990, 30 August 1990 and 12 December 1990 - the applicant committed offences for which, on 7 June 1991, on charges of detention for sexual penetration; attempted rape (two charges); rape; threat to kill; and assault with intent to rape, the applicant was sentenced to be imprisoned for five years, with a non-parole period of three years. He was granted parole on 26 February 1993. Between September and October 2000, the applicant committed the rapes against the five victims which were the subject of the sentence imposed on 26 April 2002.

187 The application for leave to appeal against sentence, before this Court, relied on a single ground, which asserted that the 'new non-parole period fixed, 43 years imprisonment, is manifestly excessive'. Given that it will be necessary, in any event, to fix a new non-parole period, the application for leave to appeal against sentence falls away.

188 We note, however, that in contending that the non-parole period fixed by the trial judge was excessive, the applicant's counsel relied on his client's age (he was born on 14 July 1971); the inability to predict the level of risk that the applicant will pose when he is eligible for parole; the manner in which the three trials were conducted; the fact that the applicant will spend his time in custody in protection; and the principle of totality.

189 In our view, the individual sentences imposed by the trial judge and the orders for partial cumulation made were, if anything, moderate.

190 The applicant's offending was utterly abhorrent. It was not, however, atypical so far as he was concerned. Over many years, the applicant has shown a depraved predilection to attack, degrade and humiliate women. As his criminal history amply demonstrates, his base inclinations have not, in the past, been curbed by substantial periods of imprisonment.

191 In our view, the applicant's appalling history leaves little or no room for optimism concerning his prospects for rehabilitation. That is so even after he has served another very lengthy period of incarceration. There can be little doubt that at the present time he remains a serious danger to women. His frequent - and recent - viciousness directed principally at young women emphatically demonstrates that this is so. Whether he will remain a serious danger to women at the expiration of the non-parole period that we are to fix is a matter incapable of sensible prediction.

192 We recognise that the purpose of fixing a non-parole period is to provide for the possible mitigation of punishment through conditional freedom - when appropriate - once the prisoner has served the minimum time that justice requires him to serve having regard to all the circumstances of his crimes.[76]

193 In our view, general and specific deterrence, and the need for denunciation and just punishment, demand that the applicant serve no less than 40 years' imprisonment before being eligible for parole. Of course, whether the applicant is granted conditional freedom at the expiration of that period will be a matter for the Parole Board (or its successor).

194 We will direct that the new non-parole period take effect from 28 May 2015.

195 Further, we make it clear that the other orders made by the County Court are to stand.[77]

196 We wish to acknowledge the assistance given to the Court by senior and junior counsel who appeared pro bono on the applicant's behalf, as did their instructing solicitors. Legal aid was, we were told, declined for the preparation and presentation of the applications in this Court. We were not told why that was so.

197 In our opinion, that decision was regrettable. The applicant's case required the assistance of experienced and competent counsel. He was in no position adequately to represent himself. Any proper appraisal of the available material would have made it abundantly clear that, at worst, the applicant had a strongly arguable case for acquittal in relation to the convictions that he sustained in the first trial. As regards the third trial, his application, though ultimately unsuccessful before this Court, was at least arguable.

[1] R v Bayley [2013] VSC 313. An application for leave to appeal against that sentence was refused: Bayley v The Queen (2013) 43 VR 335.

[2] To ensure that there is no possibility of identification of the victims of sexual offending, this judgment has been anonymised by the adoption of initials in place of the names of persons.

[3] By operation of law, that sentence is, of course, to be served concurrently with the sentence of life imprisonment imposed on 19 June 2013. See Sentencing Act 1991 s 16(1). See also R v Jolly [1982] VicRp 4; [1982] VR 46; R v Taikmaskis (1986) 19 A Crim R 383.

[4] Sentencing Act 1991 s 14(1).