O'Dowd v R

The court has to consider the safety of the conviction of Kevin O’Dowd of a number of offences against a woman to whom we shall refer as SS committed between 15 and 18 September 2004. He was tried at the Central Criminal Court before HH Judge Hawkins QC and a jury. The trial began on 6 December 2006 and the jury returned its verdicts on 22 June 2007, so the total length was about six and a half months. For a trial involving just one defendant and the relatively simple issues that the jury had to decide to have lasted for this length of time with the consequent vast cost to the public is not only disproportionate but a serious blot on the administration of justice. There is no single reason for the trial having lasted this long. Many of the delays could be justified individually but viewed collectively it is entirely unacceptable for the case to have taken anything like this length of time.

A major reason for the length of the trial was the introduction of bad character evidence admitted pursuant to the Criminal Justice Act 2003 (hereafter “the CJA 2003”) concerning three other allegations of rape, two of which related to events that occurred 22 and 17 years before the indicted charges. The first of these allegations, by RL, resulted in an acquittal, the second, by JD, in a conviction, and the third, by LB, was stayed on the ground of abuse of process. If ever there is a case to illustrate the dangers of satellite litigation through the introduction of bad character evidence this is it.

The overriding objective of the Criminal Procedure Rules 2005, S1 2005 No. 384 (the Criminal Procedure Rules”) is that all criminal cases should be dealt with “efficiently and expeditiously.” This we regard as of core importance in the interests of justice. Those interests include the interests of the prosecution, the defence and, not least, the jury. We shall refer later in this judgment to the Protocol for the Control and Management of Heavy Fraud and Criminal cases 22 March 2005, although we would observe that this case was not in essence a particularly complex criminal case.

The Crown alleged the appellant had locked SS in her flat, frightened her so that she did not dare try to escape, threatened to kill her with a knife, made her consume largactil tablets, sexually assaulted her by making her masturbate him when she did not consent, and raped her vaginally and orally. The appellant was convicted by majority verdicts of 10 to 2 of falsely imprisoning, threatening to kill, twice raping, sexually assaulting, and poisoning SS. He was sentenced to life imprisonment on counts 1, 4 and 7 with a minimum term of 9 years before his case could be considered by the Parole Board, to 7 years concurrent on each of count 3 and count 6, and to 4 years concurrent on count 5.

Limited leave to appeal against conviction was granted by the Full Court. The court stated that, after giving due allowance for the reasons for the time taken, “the case took a wholly exceptional period of time”. Much of the hearing before us concerned the first ground. Mr Kovalesky QC, on behalf of the appellant, submitted that the trial judge was wrong to admit the bad character evidence of the three other allegations of rape because they were all disputed and to admit evidence of them was likely seriously to complicate and lengthen the trial and unfairly to divert its focus from the events charged on the indictment. These difficulties have been considered by this court on a number of occasions: see Hanson [2005] 1 WLR 3169 and Edwards and others [2006] 2 Cr. App. R. 4, and, since the judge’s ruling in the present case, McKenzie [2008] EWCA Crim. 758 and DM [2008] EWCA Crim. 1544. None of those cases concerned a trial of anything like this length.

The second ground upon which leave to appeal was granted concerned the judge’s directions as to the use the jury could make of the bad character evidence. This ground has two limbs. The first concerns the direction that the bad character allegations were capable of mutually supporting the truth of the other allegations. The second is that the judge failed to direct the jury adequately or at all as to the weaknesses the defence alleged in relation to each bad character allegation. Mr Kovalesky did not pursue the first limb. He was right not to do so. It was unarguable in the light of the decisions of this court in Wallace [2007] 2 Crim App. R. 30; DM [2008] EWCA Crim 1544, and Freeman and Crawford [2008] EWCA Crim. 1863.

The decision to admit the evidence of bad character was made by the judge on 15 February 2006, after a two-day pre-trial hearing. On 5 December 2006, the day before the trial was listed to start, he rejected an application to exclude the bad character evidence under section 78 of the Police and Criminal Evidence Act. On 20 February 2007 at the conclusion of the prosecution’s evidence relating directly to SS’s complaints, and before the bad character evidence was called, the defence submitted there was no case to answer and that the bad character evidence should not be admitted. The judge rejected these submissions.

Following the ruling permitting the admission of the bad character evidence it was estimated that the trial would last four months and would end in the middle of April 2007. In the event it started on 6 December 2006 and lasted six and a half months. The defendant completed giving evidence and the defence case was closed on 23 May 2007. The judge started his summing up on 7 June. The jury retired on Thursday 14 June and returned verdicts on Friday 22 June. We have set out the chronology in an Appendix to this judgment. Before turning to the facts of the case and the rulings concerning the bad character evidence, we refer to the factors other than the bad character evidence that contributed to the length of the trial.

There were anticipated breaks for holidays including Christmas and New Year, and Easter. The condition of the appellant’s health meant that there had to be frequent short breaks during the trial to enable him to take medication. There were also other breaks to enable his condition to be assessed by Matron at the Central Criminal Court and for him to receive treatment. The trial was adjourned between 10 and 23 January 2007 (during SS’s cross-examination) because the appellant went into hospital for colon surgery. The Crown’s skeleton argument states that because the defendant had apparently instructed his legal team not to concede any point the Crown was required to prove matters which in many trials would have been uncontentious or admitted. The unpredictability of breaks because of the appellant's health and his general approach undoubtedly made the judge's task a difficult one.

After the judge’s ruling on 20 February 2007 (see [7] above) the appellant dispensed with the services of his counsel, Mr James Turner QC and Mr Hodivala. He was given some days to consider his position and decided to represent himself. It appears from the court record that the appellant acted on his own behalf between 23 February and 21 March 2007. Following an application by the Crown that the appellant not be allowed to cross-examine the three women whose allegations of rape were admitted as bad character evidence (see rule 31.1(2) of the Criminal Procedure Rules) the judge instructed Mr Bruce Houlder QC to cross examine them.

Mr Houlder and the applicant were provided with papers and given time to read and prepare. The court did not sit between 24 February and 6 March to enable Mr Houlder to prepare his cross-examination of JD. The evidence relating to JD’s allegations was heard on three days and completed on 9 March. The court did not sit between 10 and 14 March to enable Mr Houlder to prepare his cross-examination of RL. The evidence relating to RL’s allegations was heard on 6 days between 15 and 20 March.

During this time the jury sent a note expressing concern at the appellant’s ability to represent himself, and he chose to reinstruct counsel. Mr Hodivala was able to return to the case but professional commitments precluded Mr James Turner QC from doing so and the defendant did not wish to instruct Mr Houlder. A second leading counsel Mr Kovalevsky QC was instructed. There was a further adjournment and delay between 21 March and 17 April while Mr Kovalevsky was given time to read and prepare. The trial thus resumed at about the time the judge had told the jury it would end.

The defendant was admitted to hospital on various dates between 16 and 22 May, while he was being cross-examined, because of concerns about his health. Mrs. May, who led for the Crown at the trial, informed us that, after the defence case was closed on 23 May, there was a two week break in the trial to accommodate a pre-booked holiday by a juror.

We return to the allegations that led to the charges concerning the index offences. SS was a registered drug addict. She said that she met the appellant on 8 September 2004 at the home of a man called Owen Brown where crack cocaine was being smoked. She agreed to rent him a room in her flat. She said she did so because she wanted a shoulder to cry on and needed a friend, and that she allowed him to cuddle her, but told him she was not interested in men because she liked women. On 15 September she allowed him to perform oral sex on her but pushed him away when he tried to have sexual intercourse with her. She said she smoked crack supplied by the appellant and, against a background of continued use of crack, remained in his company. She said she felt that he expected her to have sex with him and was threatened. She said he said he wanted to have sex with her, accused her of “sucking cocks for rocks”, and kept on about the fact that she slept with her dealer who was black, and could not understand why she would not sleep with him.

SS said the appellant continued to ask her for sex and on 16 September she told him she would have sex with him if he could get her some crack. While he was out getting the drugs, she changed her mind and, when he returned, told him she did not want to have sex with him. She said he became angry because she was blowing hot and cold and told her that, but for the children, he would have killed his wife when he found she had cheated on him, and that he burned prostitutes and made them jump out of the window. He got a large knife from a kitchen drawer and said he would make her death look like suicide, held the knife to her throat, threatened her, and told her to swallow the largactil tablets. She said he used shoe laces to tie her to the bed and, after he noticed that she had not taken the tablets, made her take them. At some stage he untied her so she could use the bathroom but then told her to jump from the window. She said she was frightened he would kill her and submitted to sexual intercourse with him from behind in order to get away. She screamed while it was occurring.

SS said that the following morning, Friday 17 September, she was expected to visit her sister but the appellant said he would keep her hostage until Monday and demanded that she “fuck him”, but she refused. She said he then made her masturbate him and perform oral sex on him and then left the flat briefly. When he returned he told SS that she could go to her sister but when she told him she did not want to see him again he locked the door, picked up the knife, and said she was staying with him all day. She said he continued to threaten her with a knife, and then pulled her out of the flat and said he was going to kill members of her family. Eventually he threw the knife away saying he did not want to hurt her. They went to where her sister lived by bus and he returned her keys.

SS said she told her sister about the events and that on Saturday 18 September the appellant telephoned asking where she was and made threats to kill her family. She complained to the police on Monday 20 September.

The appellant claimed that SS’s account was completely unreliable. He maintained that he was a client of an escort agency for which SS worked as a prostitute and that he had had a number of paid encounters with her. He told police officers that as far as he knew SS was working for an escort agency and produced an escort agency’s card from his wallet. He said he knew that SS was a drug addict, was concerned about this and wanted to help her. He thought that over the months a romantic attachment had grown between them and said he believed all the sexual relations between them were consensual. He did not admit the particular offences alleged between 15 and 17 September but said that if they did occur, there was consent or he believed there was consent. His case was that SS had been put up to making the allegation by a woman named SH with whom she had formerly been in a lesbian relationship.

When the appellant was arrested he had a handwritten letter in his possession which purported to be signed by SS giving him authority to remain in her flat. He claimed he wrote the body of the letter and she signed it. He said he thought he should have a letter of this sort to avoid any aggravation with neighbours who did not know him. A handwriting expert considered that there was strong evidence that SS did not write the letter but was not able to give an opinion about who wrote the signature. The owner of the escort agency gave evidence. She said the agency knew the appellant as a customer but was sure that SS had not worked for them and said that the agency avoided drug users.

SS denied that she was a prostitute or had been in a consensual relationship with the appellant. She admitted that she had consensual sex with her drug dealer. She was cross-examined over 4 days during which time her credibility and her character were challenged. She agreed she had not told police what had happened when she first contacted them. She said she had not told the police about the allegation of rape because she had not found it easy to speak about rape to the police and she felt she had brought it upon herself. She was cross-examined about this and other failures and about inconsistencies between her statement and her evidence. For example, she was cross-examined about the fact that her statement said that on the Wednesday night she let the appellant perform oral sex on her followed by his starting to have sexual intercourse with her until she pushed him away, but in her evidence she had insisted it was only on the Thursday that the appellant had sex with her and that it was rape.

There was medical evidence that SS had serious problems probably linked to drug and alcohol abuse and a history of sexual abuse. Analysis of her blood and urine detected alcohol and a number of drugs including largactil and citalopram. Professor Forrest stated that taking the largactil and then falling asleep after sexual intercourse was consistent with her account, but the findings of largactil in her system was also consistent with her taking it at other times. Accordingly, he could not exclude the possibility that the largactil had been taken at an earlier date. He said that in combination largactil and crack cocaine might result in aggressive and unpredictable behaviour.

The defence relied on a number of weaknesses in the prosecution’s case about SS’s allegations and in particular in her evidence. These were set out in a schedule. They included the unreliability of her evidence as to when she first met the appellant because she said that when she first met him she was “cracked out of her nut”. When giving evidence she said she met him at Owen Brown’s home, but also said that she had not met the appellant before she went to his flat. The defence also relied on inconsistency between her evidence that they only had sex once and the contents of her statement, and on changes in her evidence as to the date on which she was imprisoned. As well as 15 September (the commencement of the indictment period), she referred to 14 and 10 September, but a CCTV image dated 14 September shows her standing in a queue at the Post Office with the appellant.

The court also heard evidence from SS’s parents, her two sisters, N and H, and from SH, Owen Brown, and the police officers who attended her flat in response to a request to assist in getting a person out of it. Her family gave evidence of telephone calls from the appellant asking for SS, claiming to be her partner and saying she was pregnant and was having his baby. SS’s sister H, who lived at the family home, said she opened a letter addressed to SS and found a photograph of herself inside with a message from the appellant on the back to SS to call him or else he would “come down your mum’s” which she took to be a threat although the appellant denied this. Her mother said his manner was angry and she took what was said as a threat and her father said he found the appellant’s tone threatening.

SS’s mother said that, after her she spoke to the appellant on the telephone, SS, in distress, told her that she had been held hostage but did not mention any sexual assault. SS’s father said that after SS complained about the appellant and said he had been involved with a firearms offence, he called the police. SH was a reluctant witness. She said SS told her about the events including that the appellant hit her, knocked her over, gave her different tablets and raped her. She also said she had not seen a man in SS’s house during August and September, and had never seen the appellant before although the appellant claimed he had met her.

Owen Brown said that SS first met the appellant when they smoked crack at his (Brown’s) flat. Later the same day or within a couple of days she said she liked the appellant and wanted to go to bed with both the appellant and Brown, adding that she was a lesbian. Brown said that a week later she passed him a note for the appellant to the effect that the appellant was a nice man and the appellant may have misread the note as a sexual advance. He said that after that the appellant and SS visited his address on two further occasions. On the first of these he understood that the appellant was lodging with SS. On the second visit he was holding her by the arm and dragged her in. He thought that on the second visit when the appellant was holding her he was being possessive but not holding her prisoner.

One of the police officers who attended SS’s flat in response to a request to lend assistance in getting somebody out of it said SS said she had not asked the appellant to leave because she was scared and he had forced her to have sex with him. He said her account included threats to kill, use of a knife, making her swallow pills, and crying when he penetrated her from behind. She said that she had given in to the appellant’s demands on the Friday because she just wanted to get out of the house but that he then prevented her from leaving. The officer stated that SS was shaking and distressed while she gave her account.

Before turning to the evidence of the allegations by the three other women; RL, JD, and LB we set out the relevant parts of sections 101 and 103 of the CJA 2003.

“‘Defendant’s Bad Character’

101.

–(1) In criminal proceedings evidence of the defendant’s bad character is admissible if, but only if -

…

(c)

it is important explanatory evidence,

(d)

it is relevant to an important matter in issue between the defendant and the prosecution …

(3)

The court must not admit evidence under subsection (1)(d)… if, on an application by the defendant to exclude it, it appears to the court that the admission of the evidence would have such an adverse effect on the fairness of the proceedings that the court ought not to admit it.

(4)

On an application to exclude evidence under subsection (3) the court must have regard, in particular, to the length of time between the matters to which that evidence relates and the matters which form the subject of the offence charged.

‘Important explanatory evidence’

102.

–For the purposes of section 101(1)(c) evidence is important explanatory evidence if –

(a)

without it, the court or jury would find it impossible or difficult properly to understand other evidence in the case, and

(b)

its value for understanding the case as a whole is substantial.

‘Matter in issue between the defendant and the prosecution’

103.

–(1) For the purposes of section 101(1)(d) the matters in issue between the defendant and the prosecution include-

(a)

the question whether the defendant has a propensity to commit offences of the kind with which he is charged, except where his having such a propensity makes it no more likely that he is guilty of the offence;

(b)

the question whether the defendant has a propensity to be untruthful, except where it is not suggested that the defendant’s case is untruthful in any respect.

(2)

Where subsection (1)(a) applies, a defendant’s propensity to commit offences of the kind with which he is charged may (without prejudice to any other way of doing so) be established by evidence that he has been convicted of –

(a)

an offence of the same description as the one with which he is charged, or

(b)

an offence of the same category as the one with which he is charged.

(3)

Subsection (2) does not apply in the case of a particular defendant if the court is satisfied, by reason of the length of time since the conviction or for any other reason, that it would be unjust for it to apply in his case.”

Section 101 requires a two-stage procedure. The first stage is to determine whether the bad character evidence meets the “gateway” criteria for the admission of evidence of a defendant’s bad character contained in section 101(1). Mr Kovalesky accepted that the evidence of the allegations made by RL, JD and LB met the criteria in section 101(1)(c) or (d) and was accordingly prima facie admissible: see Weir [2006] 1 WLR 1885 at [35]-[36]. This appeal is concerned with the second stage of the statutory procedure, which is contained in section 101(3). That requires the court to exclude bad character evidence that is prima facie admissible under section 101 if it would be unjust to admit it.

Section 101(3) only applies to bad character evidence relevant to an important matter in issue between the defendant and the prosecution (section 101(1)(d)) and where the defendant has made an attack on another person’s character (section 101(1)(g)). It is in mandatory terms. Its wording is more emphatic than that of section 78 of the Police and Criminal Evidence Act 1984 which uses the word “may”. However, if, when considering the application of section 78, the court decides the admission of the evidence would have such an adverse effect on the fairness of the proceedings, it cannot, as Auld LJ stated (Chalkley [1998] QB 848, at 874), logically exercise a discretion to admit. Accordingly, this difference in the wording of section 101(3) and section 78 may in itself not be significant: see Tirvanenanu [2007] 1 WLR 3049.

The bad character provisions in the CJA 2003 are based on recommendations made by the Law Commission: Evidence of Bad Character in Criminal Proceedings, Law Com. No 273, (2001). They do not, however, contain the main safeguard proposed by the Commission. The Commission’s proposal was that, subject to limited exceptions, the prosecution should only be permitted to adduce bad character evidence with the leave of the court. As we have observed, the Act provides that bad character evidence that meets one of the gateway criteria in section 101(1) is admissible. Professor John Spencer, in his Evidence of Bad Character (2006) at [1.15] has said that by this and other changes the Law Commission’s proposals “had been significantly ‘bent’ to make evidence of the defendant’s bad character more readily admissible”. The only safeguards contained in the 2003 Act are those in section 101(3) and section 107. Section 107 only applies after the close of the case for the prosecution and requires the judge to stop the case where the evidence is contaminated in a way that would make a conviction unsafe.

The importance of section 101(3), to which we return later in this judgment, is that where bad character evidence qualifies under the criteria and is prima facie admissible, it is the only provision in the 2003 Act itself available to a court when considering whether in fact to admit such evidence. This court indicated in Highton [2005] 1 WLR 3472 that that section 78 of the Police and Criminal Evidence Act 1984 also applies in this context, although it did not consider the provision should have been applied to exclude the evidence in that case . In Weir [2006] 1 WLR 1885 (the appeal of Somanathan), it was said that there was “no reason to doubt” that section 78 should be considered where section 101(1)(f) is relied on (and section 101(3) does not apply), but again the court did not consider the provision should have been applied in that case to exclude the evidence. Although we note that in Davis [2008] EWCA Crim 1156, it was said that the role of section 78 is “possibly controversial” under the CJA 2003, we agree with the indications in Highton and Weir that section 78 should be considered, as the judge did in this case.

We have referred to the defence position. At the pre-trial hearing the defence submitted that admitting evidence of the bad character allegations by the three women would make a simple case complicated and would expand it out of all proportion and would be unjust. It would expand the case because the appellant denied the allegations, two of which had not resulted in a conviction, and the Crown would have to prove them. It would be unjust to admit the allegations because the defence would be handicapped in dealing with them. This was because of the age of the first two allegations, and because of the absence of some key statements and transcripts. In relation to the allegations by JD for which the appellant had been convicted it was argued the defence would be handicapped because the appellant had represented himself and absented himself from part of the trial.

The judge was referred to and considered the decisions of this court in Hanson [2005] 1 WLR 3169, Edwards [2006] 2 Cr. App. R. 4, and to the decisions of the House of Lords in O’Brien [2005] 2 AC 534; UKHL 26, and Z [2000] 2 AC 483. The judge stated his task was to assess the potential significance of the evidence in the context of the case as a whole and assuming it to be true. In doing so he said, in the light of the decision in O’Brien, it was necessary for the evidence to have “enhanced evidential value”, and that the lengthening of the trial by the additional evidence must not render the trial process unfair because what is required is a trial process fair to all parties. The judge had regard to the decisions in Edwards and Hanson and in particular the three questions set out in Hanson:

“1 Does the history of conviction(s) establish propensity to commit offences of the kind charged? 2 Does that propensity make it more likely that the defendant committed the offences charged? 3 Is it unjust to rely on the material and, in any event, will the proceedings be unfair if the material is admitted?”: [2005] 2 Cr. App. R. 21 at [7].

We turn to the evidence of the allegations. The earliest allegation was by RL, who had lived with and entered into a bigamous form of marriage with the appellant, but they parted. She claimed that on 22 September 1982 he had smashed his way into her home, threatened her with a knife, threatened to kill her, and raped her. In his defence the appellant claimed he had gone to RL’s house to see the children and, in order to get him out of her life, RL made up the allegations with the encouragement of a Detective Inspector Wallace with whom she had had an affair. At the trial the appellant produced a photocopy of a photograph of RL with Sir Edward Heath and alleged there were photographs depicting her in compromising sexual circumstances in front of their child. RL denied that she had an affair with DI Wallace or that there were compromising photographs in front of their child.

On 14 March 1984 the appellant was acquitted of the charges of threatening to kill RL and of raping her. Subsequently the photocopy of the photograph produced at the trial was found to be a forgery, and the appellant was tried for perjury, making false statements and perverting the course of justice. He agreed that the photographs he had produced were forged but maintained that there were original photographs of RL, DI Wallace, Sir Edward Heath, and his children in pornographic poses. He maintained that those photographs were taken from him by the police. In October 1985 he was convicted of the forgery and perverting the course of justice.

In the present case the Crown submitted RL’s allegations were admissible under sections 101(1)(c) and (d) of the CJA 2003 as important explanatory evidence, and relevant to an important matter in issue between the defendant and the prosecution. The Crown submitted that RL’s evidence showed a propensity by the appellant to commit offences of the nature and with the same characteristics as those alleged by SS. It argued that RL’s allegations had the following common elements with SS’s allegations; an association that turned sour, threats to kill, use of a knife, reversion to a more normal state of mind after submission by the woman, and an attempt to discredit the complainant by calling into question her moral standards.

The judge did not consider the evidence was admissible under s 101(1)(c), but decided that it was admissible under s 101(1)(d). The principle in R v Z [2000] 2 AC 483 that similar fact evidence of an allegation which had led to an acquittal was in principle admissible had survived the CJA 2003 and RL’s allegation was one of reprehensible conduct by the appellant. The judge found that, because of the similar features he had identified, the evidence had the necessary enhanced probative value in relation to the allegations by SS. He took into account how long ago RL’s allegation’s were made, the fact that some statements were missing, there was only a partial transcript of RL’s evidence, and the need to ensure the trial of the charges based of SS’s allegations did not degenerate unreasonably into trial of satellite issues. He concluded that RL’s evidence showed a propensity to commit offences of the type charged, and made it more likely that the appellant committed the index offence, and that it was not unjust or unfair for it to be relied on.

As far as the convictions for perjury and perverting the course of justice are concerned the Crown argued the conviction was admissible to show a propensity for untruthfulness (sections 101(1)(d)) and 103(1)(b)) in that the appellant falsely sought to blacken RL’s character in defending the charge of rape by presenting her as a woman of loose morals and gross untruthfulness. The judge referred to the letter the Crown alleged the appellant had forged indicating SS had given him authority to stay in her home. Although the convictions were 20 years earlier the judge found the convictions by their nature showed a propensity for the appellant to be untruthful, and by their nature made it more likely he committed the offences charged.

The evidence concerning RL’s allegations came from her and three other witnesses, and took six days of court time. In his evidence the appellant maintained the account he had given at the two trials.

When summing up, the judge drew attention to the aspects of RL’s evidence which the defence relied on to show she had lied. These included the absence of any reference to rape in her initial complaint, which the defence said was a significant inconsistency, RL saying that her daughter had been present at the time but that the daughter was keen to go to Margate with the appellant the next day, and her admission to lying on oath at the appellant’s committal hearing. The judge told the jury that it was not very clear what RL was admitting to lying about, that it was all a long time ago, and that they should bear that in mind in assessing the appellant’s ability to defend himself on these matters. He said “that is a ground for considering [RL’s] evidence with caution. Take a special care over her evidence in those circumstances”.

The second tranche of bad character evidence related to the allegations made by JD. She said she had a friendly non-sexual relationship with the appellant who gave her cannabis which she could not afford to buy, and that he frequently made sexual advances to her which she indicated were not welcome. She said that in December 1987 she went out with him and her three-year old son to find the boy’s father. They did not do so and, after visiting various public houses in which she drank some two pints of lager and smoked some cannabis with the appellant, they returned to the flat.

JD said that in the flat the appellant made sexual advances and she made it clear they were not welcome. When he tried to kiss her and she pushed him away he punched her in the face and then kept her in the flat against her will for 12 hours. She said he threatened her and she submitted to sexual intercourse through fear. She escaped with her son when he fell asleep. The appellant said that he paid for sex with JD, and that she was on the game. He said she made the allegations up after they had a row because he realised she was ripping off charities. On 9 November 1988 the appellant was convicted of raping JD and sentenced to 6 years imprisonment. In March 1989 he was also sentenced to a consecutive sentence of 8 years imprisonment for grievous bodily harm and a firearms offence, making a total of 14 years. It appears from the ruling on the bad character applications that the judge was told the appellant was released on 9 October 1997.

The Crown relied on section 101(1)(d) and 103(1)(a) of the CJA 2003, submitting that JD’s evidence showed the appellant has a propensity to commit offences of the same nature and with the same characteristics. Those characteristics were; befriending a victim, paying for her cannabis, and making sexual overtures which were repulsed, and then telling the victim she could not leave, threatening to kill her, and using violence against her which led to her being scared into submitting to sexual intercourse. The judge found that the similarities in the allegations were significant and cumulative and that the enhanced probative value of the evidence well exceeded the prejudicial effect of introducing it. The judge took into account the time since the conviction, but also the fact that during that time the appellant had spent a lengthy period in custody.

The jury heard evidence about JD’s allegations and the conviction from her, four other live witnesses, and two witnesses whose statements were read. It took three days of court time. In his evidence to the jury in this case the appellant substantially repeated his account at the earlier trial. He denied he had supplied JD with cannabis or threatened her and said she was on the game and he paid her for sex.

When summing up the evidence about JD’s allegations, after summarising the appellant’s account, the judge referred to the aspects of JD’s evidence which the defence submitted showed she had lied and the inconsistencies in her evidence. He also referred to the evidence of the doctor who examined her that he had found no marks on her body, and that JD had consumed more alcohol than was wise.

The third allegation was by LB. She, like SS, was a drug addict. In her case the addiction was to cannabis. She alleged that in 2000 the appellant raped, falsely imprisoned and indecently assaulted her. She said the appellant struck up a friendship with her and her boyfriend, supplied them with cannabis and visited their flat frequently. There was a row when they requested him to stop calling on them so much after he showed LB photographs of a person who he said he had shot in the kneecaps, and she did not want him to visit. LB said that about a week later he apologised and, after spending the day in her flat, sat on her bed and said he wanted to have sex with her and knew she felt the same. She said he behaved aggressively. Although she did not consent she submitted to intercourse through fear. She did not shout out because she was scared. She also said that he made her sit in his car and drove away with her. While they were in the car he threatened her, saying he was going to kill her and dump her body. She said he also told her that if she was pregnant he did not care if the boyfriend was the father because he would treat the baby as his, and only drove her back to her flat when she agreed to tell her boyfriend she was leaving him for the appellant. At the flat he pushed her into the bedroom, made further threats and tried to have sex with her. LB said her boyfriend banged on the door and the appellant took a knife saying he would stab the boyfriend. She said she persuaded the appellant to call the police whose arrival defused the situation. She fled to Scotland with her boyfriend a few days later and did not report the matter for a year.

As a result of LB’s allegations the appellant was charged and tried at Harrow Crown Court in September 2003. However, the judge in that case stayed the proceedings on the ground of abuse of process. She did so because, while being cross-examined by defence counsel, LB said, “you are trying to put a rapist back on the street for a second time” and thus confirmed an earlier submission made by the defence that her evidence had been contaminated because she had been told of the appellant’s conviction for rape.

The Crown also relied on section 101(1)(d) and section 103(1)(a) of the CJA 2003 in respect of LB’s evidence. The similarities in her allegations and those of SS were said to be; the appellant striking up a friendship with a drug addict victim, supplying the victim with cannabis, recounting stories of violence committed against others, making sexual advances to the victim in her flat, and, after rejection of those advances, forcing her into sexual intercourse through fear.

The defence submitted LB’s allegations had been contaminated and should not be admitted in the present case. They relied on section 107 of the CJA 2003 which requires a case in which bad character evidence has been given to be stopped if the evidence has been contaminated and a verdict would be unsafe. The judge rejected this submission. He said he had power to keep the admission of evidence under control and did not consider the matter had so undermined LB’s credibility that she should not be regarded as reliable in relation to her allegations against the appellant. He also considered that he could deal with this issue by an appropriate direction about contamination. Again the judge found the similarities showed enhanced probative value and that the evidence should be considered by the jury. It showed a propensity by the appellant to commit offences of the kind charged and it was not unfair to admit it.

The evidence concerning LB’s allegations came from her and three other witnesses and took seven days of court time. In his evidence the appellant claimed LB was lying. He said he had a consensual sexual relationship with her for about two months and had paid her for sex.

The judge drew the attention of the jury to the discrepancies between LB’s account and doctors’ notes, and her acknowledgement that she had lied to doctors. He directed the jury that they should take special care about LB’s evidence in the light of a comparison of her account that her boyfriend was not violent and certain 999 calls which she had made. He also drew their attention to LB’s failure to complain about being sexually assaulted and falsely imprisoned in any of the large number of telephone calls she made at the material time, or to the police when they called at the appellant’s flat, and to the different reasons she had given for failing to report matters for a year.

Mr Kovalesky submitted that the judge erred in his approach to the balancing exercise required by section 101(3). Although some of what subsequently happened during the trial had not been foreseeable, the difficulties of allowing evidence of the three allegations of rape, each of which amounted to a complicated case, were foreseen. As a result he submitted the convictions are unsafe.

At the heart of the appellant’s case is the submission that, although the bad character material may have been admissible, it seriously diverted the focus of the trial and amounted to the sort of satellite litigation which this court in Hanson and in Edwards (which were before the judge when making his ruling) stated should not occur. In Hanson [2005] 1 WLR 3169 at [12] this court stated that “where past events are disputed the judge must take care not to permit the trial unreasonably to be diverted into an investigation of matters not charged on the indictment”. In Edwards and others [2006] 2 Cr. App. R. 4, at [1(vii)], in the context of an allegation rather than a conviction, the court stated; “this is an area in which it is important to guard against satellite litigation”. Similar injunctions against this danger have been given in the recent decisions of McKenzie [2008] EWCA Crim. 758 at [22]-[24] and DM [2008] EWCA Crim. 1544 at [22].

There were a number of difficulties with the bad character evidence the Crown wished to adduce in this case. One was the age of the incidents involving RL and JD. In the case of JD the conviction was for conduct some 17 years before the allegations made by SS. The allegations by RL were even older. They alleged conduct by the appellant 5 years before those by JD; that is 22 years before those by SS. Secondly there was only a single conviction. In the light of what this court stated in Hanson at [9] it is normally difficult to show propensity from a single conviction in itself. It is only where the conviction shows a tendency to unusual behaviour or where its circumstances demonstrate probative force in relation to the offence charged that a single conviction may show propensity.

The third and, in our view, the most serious difficulty with the Crown’s application to adduce evidence of all three allegations and to call all three complainants is that so much of what the Crown wished to adduce was disputed. It was known at the time the Crown’s application to adduce the bad character evidence was first considered that all the facts of the allegations by the three women were contested. Accordingly, particularly in the case of the allegations that did not result in a conviction, proof of the previous alleged misconduct would require the trial of a three collateral or satellite issues as part of the trial of the applicant for the offences with which he was charged.

This was not a case like McKenzie [2008] EWCA Crim 758, in which the allegations of prior misconduct had never been investigated by the police and the incidents had not been drawn to the defendant’s attention at the time they were said to have happened. What was said in that case about the difficulties is, however, relevant. Delivering the judgment of this court, Toulson LJ referred to three potential difficulties that need to be considered in such cases. The first is the need to consider whether admission of such evidence would result in the trial “becoming unnecessarily an undesirably complex even if not unfair”: [2008] EWCA Crim. 758 at [22]. The second is the danger of a trial of collateral issues not only adding to the length and cost of the trial but “complicating the issues which the jury has to decide and taking the focus away from the most important issue or issues”(ibid at [23]). The third is what the court described as the dilemma that “…if allegations of previous misconduct are few in number, they may well fail to show propensity even if they are true, but the greater the plethora of collateral allegations, the greater the risk of the trial losing its proper focus”: (ibid at [22-24]).

In this case the Crown, perhaps mindful of the difficulties in prosecuting rape cases, considered it needed the evidence of all three bad character complainants. It did so because it considered the appellant was masterly at manipulating vulnerable women such as SS, a drug addict, and defending himself by reviling them in a way which, but for the evidence of similar allegations by others and similar defences, might well cast doubt on her allegations. Mrs May did not accept the suggestion that the bad character evidence was adduced to bolster a weak case. She submitted that in this case the evidence was vital to a fair consideration of SS’s allegations. She submitted that the judge clearly understood the test in Hanson to which we have referred and directed himself correctly in admitting the evidence. Indeed, she said the judge applied a more stringent test than was required because he required each element of the bad character evidence to possess “enhanced probative value, a requirement greater than in fact necessary for admissibility under the CJA 2003: see Weir [2006] 1 WLR 1885, at [36].

It is undoubtedly the case that the feel of a trial judge for the case in hand is very important in this context and that an appellate court should hesitate before interfering with the trial judge’s conclusion on a matter of judgment: see Renda [2006] 1 Cr App R 380 at [57]; Edwards and others [2006] 2 Cr. App. R. 4, at [1(viii)], [27] and [51], and DM [2008] EWCA Crim 1544. The Crown submitted that in this case the judge was in a particularly good position to weigh the strength of the prosecution case and the other issues concerning the admission of the bad character evidence before that evidence was admitted. This is because, apart from the extensive pre-trial hearing at which the matter was first considered, he reviewed the position on 20 February 2007 when considering the renewed application after all the evidence directly linked to SS’s allegations had been adduced.

The judge approached the question of admitting each of the three categories of bad character evidence with considerable care. There were undoubted similarities between SS’s allegations and the three other allegations. There were particular similarities to the allegations made by JD and LB. Both were vulnerable young women who alleged the appellant befriended them, supplied them with their drugs, when they rejected his sexual advances because unpleasant, threatened to kill them and raped them. In both cases the appellant had sought to discredit them in the same way he sought to discredit SS; by saying that he had paid them both for sexual favours to which they consented.

We have referred to the importance of section 101(3) of the CJA 2003 at the stage the court first considers an application to admit bad character evidence: see [29] – [31] above. It is incumbent upon a judge considering the application to try to project forward to see the problems which might later arise in the trial as a result of the disputed bad character evidence being admitted before ruling on the application. This is because, once the evidence is admitted, unless it has been contaminated (and section 107 applies) the question is of its weight. The judge will then only have limited remedies open to deal with problems that arise thereafter. After a decision to admit the evidence the prosecution will, absent a direction not to do so, open the case on the basis that the bad character evidence will be adduced. If, during the trial but before the evidence is adduced, it becomes clear that admitting it will be unfair or will so complicate the trial that the jury cannot properly consider the matters before them, since they have been open to the jury, there may be difficulties in revisiting the matter and it may be necessary to discharge the jury. After the evidence has been adduced the difficulties will increase.

The judge’s concern to give the appellant every opportunity to put his case was evident both in conducting the balancing exercise under section 101(3) when making his ruling in February 2007, and during the course of the trial. While recognising the difficult task that the judge had in this case, we have concluded that he fell into error in his consideration of the section 101(3) matters. This is because he did not or did not adequately consider the cumulative effect of the introduction of three separate contested issues into the trial on its overall length and on the jury, or how the evidence might be timetabled or truncated. A judge has wide case management powers under the Criminal Procedure Rules. We have referred to the overriding objective, as set out in Part 1 Rule 1.1(2)(e), which requires all criminal cases to be dealt with “efficiently and expeditiously”. All participants in criminal proceedings must prepare and conduct the case in accordance with the overriding objective; see Part 1, Rule 1.2 (1)(a) and the judge’s powers under the rules include power to give directions on the court’s own initiative (rule 3.5(2)(b)), and to identify a timetable (rule 3.10 (h)).

The importance of controlling the length of trials and of their being conducted in a way that enables juries to retain and assess the evidence which they have heard was emphasised by Lord Woolf, then Lord Chief Justice, when handing down the Protocol for the Control and Management of Heavy Fraud and Complex Criminal Cases on 22 March 2005. We have observed that this case was in essence not even a particularly complex criminal case, but what is stated in the Protocol is of relevance. The Protocol states that if the jury cannot retain and assess the evidence which they have heard “the trial is not fair either to the prosecution or the defence”. It also states that “there is a consensus that no trial should be permitted to exceed a given period, save in exceptional circumstances; some favour three months, others an outer limit of six months”. Lord Woolf favoured the three-month period and, when handing down the Protocol, stated that: “it is implicit in the Protocol that trials of six months are just not capable of satisfactory disposal”. The ruling that evidence of the three bad character allegations could be adduced led to an estimate that the trial would last for four months. That in itself should have rung warning bells to the judge when undertaking the balancing process under section 101(3), especially since this was a trial of a single defendant concerning a single complainant, albeit on an indictment containing seven counts. If as we understand to be the case when the decision to admit the bad character evidence was made it was estimated that the trial would last for four months, this indicates to us that those who subscribed to this view were in the wrong mindset. The judge should have told the parties it was not going to last for four months and taken steps accordingly, the most obvious being to limit or exclude the bad character evidence; other steps include setting a firm timetable.

In Hanson (at [12]) the court stated it is necessary to examine each individual conviction rather than the name of the offence or the defendant’s record as a whole. In principle the same is true of allegations. When considering the impact of the bad character evidence on the fairness of the trial where there are several convictions or allegations, it is also necessary to consider their cumulative effect on the trial. The judge considered the position of the defendant in relation to dealing with each individual conviction or allegation, but does not appear to have considered the cumulative effect of the three disputed allegations on his position. But it is not only the position of the defendant that must be considered. It is particularly important to assess the cumulative effect on the jury of receiving the evidence where it is disputed.

Although, as was said in McKenzie, the greater the number of collateral allegations the greater the risk of the trial losing its proper focus (see [56] above), it is not the number of allegations in itself which is problematic. For instance, in Z [2000] 2 AC 483, there were four previous complainants and the defendant had been acquitted on three occasions. As Mrs May observed, there are many trials involving charges of sexual offences by a number of complainants where the jury is able to cope. The important factor is not the number of allegations but their nature and complexity, and the time it will take to put them before the jury where they are contested.

In this case there were significant complications with the allegations. Only one was supported by a conviction, which this court in McKenzie (at [23]) referred to as the launch pad for establishing propensity. Without that launch pad, a trial of the collateral or satellite issues is necessary with the dangers to which we have referred. Because the evidence was disputed, significant factual issues would have to be explored in relation to all three allegations, each of which needed witnesses. As Moses LJ stated in DM at [22] the jury would need “to consider with as much detail and concentration all the facts” in relation to each of the three allegations as they would in relation to the offences with which the appellant was charged, before relying on it in relation to the index offence. This is because the jury would have to be sure those allegations were true before relying on them in relation to the index offence. Our summary of the bad character evidence shows that the issues that would have to be explored in relation to each of the allegations and each of the bad character complainants were not straightforward: see [32] – [33], [38], [43], [45] and [49].

Additionally, because the appellant had been acquitted of the rape of RL and the proceedings arising from LB’s allegations had to be stayed, the judge had to consider and direct the jury as to the meaning of, in RL’s case, the implications of the acquittal, and in LB’s case of the stay. The position has similarities to that in DM where the Scottish proceedings resulted in a verdict of not proven and if bad character evidence were admitted the judge presiding over in DM’s trial would have to direct the jury as to what that meant. In that case Moses LJ stated this, together with the need for the detailed examination and scrutiny of the evidence to which we have referred, combined “to provide a paradigm of the satellite trial which a trial judge ought to avoid lest the focus of the jury should be diverted and deflected”.

In this case the Crown’s response to the difficulties the defence said it would face if the bad character evidence were adduced was to offer to make admissions and to call witnesses other than RL, JD, and LB, whom they wished to call. For example, in relation to the oldest set of allegations, those by RL, the defence pointed to the absence of a contemporaneous note of her original complaint or transcripts of her evidence. The Crown maintained that the difficulties said to arise could be met by a combination of admissions on some matters and the availability of transcripts prepared for the perjury trial, including transcripts of the appellant’s evidence at the earlier trial, and of the original prosecution papers. In relation to JD’s allegations, while there was no transcript of her evidence and no statement by the person to whom she first made the complaint, there was a statement by Pauline Briscow, to whom she went very soon afterwards and whom the Crown offered to call and who was called. The Crown also offered to call and called Dr Craig who had examined JD after she made her allegation, and a forensic scientist. The judge accepted the submission that the fairness of the proceedings would be ensured by the admissions, the additional evidence and the available transcripts. The effect of admitting the evidence was, however, significantly to increase the number of issues the jury had to consider.

Mrs May submitted that, although the Crown wished the three bad character complainants to give evidence, it was not that which complicated the trial. She submitted the complications that arose from this evidence were the result of the position of the defence in relation to it, including the request that other witnesses give evidence. The remaining complications and delays did not result from the bad character evidence but from the appellant’s various health problems and his dismissing his legal representatives in February.

The appellant’s health problems, his instructions that no point was to be conceded, and the later disruption when he sacked his counsel and subsequently sought new representation undoubtedly caused significant and unforeseeable delay and understandable difficulty for the judge. The appellant was responsible for the delays caused by sacking his counsel and some three weeks later deciding he did after all want legal representation. He cannot, however, be criticised for the delays caused by his significant health problems during the trial. We also consider that the complications that arose from the bad character evidence cannot be ascribed to the defence in the way suggested by Mrs May.

Although the Crown wished to reduce the evidence to be put before the jury in respect of each allegation, it wanted all three bad character complainants to give evidence. Since the evidence was challenged it was inevitable that more witnesses would be required and that, in particular there would be substantial cross-examination of the three bad character complainants. Delivering the judgment of this court in L [2007] EWCA Crim 1912, Latham LJ (V-P) stated at [12]:

“… the proper position in cases where the prosecution, in order to put forward evidence of bad character in these circumstances, effectively has to ask the court to evaluate the evidence at a previous trial; is that, whilst the prosecution will inevitably seek to reduce to a necessary minimum the amount of material which goes before the second jury, nonetheless it can only do so if it is prepared to be in a position to put before the jury all the evidence which was available at the previous trial if that is the only proper way to ensure fairness to the defendant”.

The duty that lies on all parties to prepare a criminal case in accordance with the overriding objective includes consideration of the nature and extent of any application to admit bad character evidence. Here it was clear from the outset that the allegations by the three women were disputed and that they would be subjected to detailed examination and scrutiny.

What should a judge do when faced with an application to admit bad character evidence which is disputed but which the Crown submits is, in the light of the position to be taken by the defence, vital to a fair consideration by the jury of a complainant’s allegations? In the light of the particular warnings about the admission of such evidence given in Hanson and Edwards, warnings reiterated in McKenzie and DM, a number of matters should be considered. These include the exercise of the judge’s wide case management powers under the Criminal Procedure Rules to impose a timetable on both the Crown and the defence, and, after ascertaining whether evidence can be truncated without unfairness, to give directions to this effect. In this case consideration should have been given to directing the Crown to pick the best of the three allegations and only to consider whether to admit that evidence. Although, in Hanson it was said that it is normally difficult to show propensity from a single conviction in itself, the court did not rule it out in an appropriate case. Consideration should also have been given to the use of section 74(3) of the Police and Criminal Evidence Act 1984 in respect of the appellant’s convictions for raping JD and for forgery and perverting the course of justice. It was common ground before us that these matters were not considered at the pre-trial hearing that considered the admission of the bad character evidence.

The need for the witnesses and the extent of the issues concerning the bad character matters that were contested and would have to be scrutinised in detail were, as we have noted, known at the time of the pre-trial hearing and the judge’s subsequent ruling. The approach of the defence was known then. In the light of the nature of the similarities of the allegations made by the three women, we accept that there was a good case for some of the evidence to be put before the jury. But what happened was that all three allegations were put forward and there was no consideration of timetabling or whether the evidence could be truncated in some way. The judge was concerned to give the appellant every opportunity to put his case.

Mrs May submitted that the bad character evidence lengthened the trial by only some two weeks. In fact, it was lengthened directly by 16 days; that is by three weeks and one day of court sitting time. The Crown also relied on the fact that there were no indications that the jury was not focussed whether by notes from them or in any other way. Indeed, a juror who had serious problems with his feet towards the end of the trial was keen to remain part of the jury and to return, and did so. Mrs May also relied on the fact that all the questions asked by the jury were about the allegations concerning SS and not the allegations by the three other women. This is, however, not altogether surprising. The jury were told by the judge, as they had to be, that the bad character material was only of help to the prosecution if it helped them to come to conclusions so far as SS was concerned.

We have referred to the number of court days taken by the bad character evidence. In total 16 out of 42 days, i.e. some 38% of the days on which evidence was heard were taken up with bad character evidence. The significance of the bad character evidence in the duration of the trial can also be seen from the proportion of the summing up concerned with it. The summing up runs to 434 pages, of which 148 pages, about one third of the summing up, deal with the bad character evidence. This is a greater proportion than the quarter of a much shorter summing up in McKenzie’s case, and about which the court (at [33]) expressed concern.

We have been primarily concerned with the position at the time of the original decision to admit the evidence. After the significant delays caused by the appellant’s ill health and the adjournments that resulted from his dispensing with his original representatives the consequence for the duration of the trial was manifest. As we have said, after the trial started and the Crown had referred to the bad character evidence in opening its case, the judge’s remedies were limited. But, nevertheless, after the other problems in the trial had occurred and the consequences were appreciated, the various methods by which the bad character evidence could be timetabled, truncated or otherwise controlled should have been considered or, if they had been previously considered, reconsidered. They were not. We also understand there was no consideration at that stage of longer or more flexible sitting hours, again out of concern for the defendant. It might, for example, have been possible to sit more flexible hours to accommodate the appellant’s health problems without losing sitting time.

We have noted (at [6]) that only the second limb of the second ground upon which leave to appeal was granted was pursued. Mr Kovalesky submitted that the judge failed to direct the jury adequately or at all as to the weaknesses the defence alleged in relation to each bad character allegation.

The judge directed the jury that they should only take into account those items of bad character evidence of which they were sure were reliable, that if it were possible that any of the evidence was contaminated that went to its weight and, “if there is a real possibility” a witness’s evidence was “contaminated in a significant way” they “should ignore that witness’s evidence entirely”. The remainder of his directions followed the Judicial Studies Board’s Specimen Directions. The judge said that, if the jury were sure a witness was independent and that it was not possible that the witness was lying or mistaken, they should consider how similar the allegations were, and should not take account of the evidence if they thought it was insufficiently similar. Secondly, he said bad character evidence could not be used to bolster a weak case, and it was for them to decide whether the evidence showed a tendency to behave in the way alleged by SS, to be untruthful, or to make false attacks on prosecution witnesses. Thirdly, he directed them that the bad character evidence was only one factor to be taken into account and did not mean the appellant had the tendencies alleged or had made a false attack on a witness.

After giving the jury his directions on bad character the judge briefly outlined the various accounts given by the complainants and the defendant’s answer to them. The judge then summarised what the prosecution said and what the defence said. The prosecution’s case was that the bad character complainants had given remarkably similar accounts of their experiences at the hands of the defendant and that their evidence demonstrated his conduct to vulnerable women who came into his sphere. He said he did not specifically set out at that stage the significant similarities relied on by the prosecution because he had seen that many members of the jury noted them when Mrs May listed them. As to the defence he said:

“[The defence say] that the witnesses are not reliable, plainly they say that the evidence of [SS] is not reliable and they say that the evidence of the three other women is likewise not reliable and counsel invited you, in effect, to look at their background in assessing their reliability. He submitted boldly that they are all liars. He included his client in that description, the defendant, he said, lies to the police, but he has admitted that he lied to the police. … The defence say there are common features between the allegations but are they significant? None are unusual in themselves, that is what the defence are saying. Indeed, they developed matters by saying there are significant differences between the accounts. So those are all matters that you will have to consider.”

Mr Kovalevsky said that the direction that bad character evidence could not be used to bolster a weak case came towards the end of “complicated” bad character directions and did not explain to the jury in what ways the defence suggested the case was a weak one or in what circumstances they should ignore the bad character evidence. Our summary of the judge’s directions shows that the second of these submissions is misconceived. The judge directed the jury as to the probative force and limits of propensity evidence, and as to the circumstances in which they should ignore the bad character evidence in this case.

What of the submission that the judge did not explain to the jury in what ways the defence suggested the case was a weak one? We have referred to what the judge said about the weaknesses in the bad character evidence when summarising the evidence about each of the three allegations and as part of his narrative about those allegations: see [40], [45], and [51]. He dealt with RL and JD’s evidence on 7 June, the same day as his direction on bad character. He dealt with most of LB’s evidence on 12 June, but had also referred to it, mainly in relation to the contamination issue, on 7 June.

As to the weakness of SS’s allegations, the judge’s summary of her cross-examination runs to 57 pages of the transcript. Individual criticisms, for example that she did not initially allege she had been raped, her drug-taking, hallucinations, mental health, previous convictions, and alleged promiscuity, are dealt with in that narrative. This part of the narrative was dealt with on 12 and 13 June, the third and fourth day of the summing up, and five and six days after the judge’s directions on the bad character evidence. All that he said on 7 June, the day he directed the jury about bad character evidence, was the passage we have set out stating the defence suggested the evidence of SS was not reliable, but giving no particulars or examples.

We consider that in the case of bad character evidence, save in the simplest of cases, it is ordinarily desirable not just to comment on the strengths and weaknesses of the evidence as part of a narrative summary of the evidence of a witness, but also to draw the threads together. As well as directing the jury as to the probative force and limits of propensity evidence, we consider that it is ordinarily desirable to draw those threads together by giving guidance as to the strengths and weaknesses of the bad character evidence tailored to the facts of the case. This would put the cumulative impact of bad character evidence; both its strengths and its weaknesses, directly before the jury. In the present case, although the judge did not do this, in the light of all that he did say to them when summing up the case, we do not consider that this in itself affected the safety of the conviction.

Although an appellate court should hesitate before substituting its view for that of the judge in this context, for the reasons we have given, and notwithstanding the care with which the judge approached his difficult task, we consider that this is a proper case to do so. The admission of the totality of the bad character evidence and the consequent need to scrutinise the evidence concerning the three disputed allegations, two of which concerned events 17 and 22 years before the allegations made by SS, made the trial unnecessarily and undesirably complex.

In the end we have to ask ourselves whether the conviction of the appellant is safe. This trial should not have lasted for six and a half months or anything approaching that. In our view the combination of the introduction of bad character evidence that led to the intensive investigation of satellite issues combined with the numerous interruptions to the trial and its overall length made it very difficult for the jury to keep its eye on the ball. Each member of the court is regrettably driven to the conclusion that the verdicts of the jury are not safe and therefore cannot stand. The appeal is accordingly allowed.