R v Shah [2025] EWCA Crim 1459: Sexual Fantasy Messaging with Sex Workers Admissible as Bad Character to Prove Mindset and Sexual Intent under s.101(1)(d)

R v Shah [2025] EWCA Crim 1459: Sexual Fantasy Messaging with Sex Workers Admissible as Bad Character to Prove Mindset and Sexual Intent under s.101(1)(d)

Court: England and Wales Court of Appeal (Criminal Division)

Date: 13 November 2025

Neutral citation: [2025] EWCA Crim 1459

Introduction

This appeal arose from the appellant’s conviction at Leicester Crown Court for rape and assault occasioning actual bodily harm, stemming from an incident in the early hours after the complainant left a nightclub. The core evidential controversy on appeal concerned the trial judge’s admission of extensive text communications between the appellant and numerous sex workers in the weeks leading up to the incident. Those messages contained requests and fantasies about rough oral sex and included explicit references to “rape” and other extreme acts.

The single ground of appeal granted leave challenged the admission of this material as bad character evidence. The Court of Appeal dismissed the appeal, upholding the admission of the messages through the gateway in section 101(1)(d) of the Criminal Justice Act 2003 (CJA 2003) as relevant to important matters in issue—most notably the appellant’s sexual intent and mindset on the night, and to rebut his account that he was merely driving around for “me time.”

The judgment clarifies two related points of principle: first, that sexual communications with sex workers—although ostensibly concerning consensual paid activity and sometimes expressed as “role-play” fantasies—can still be admissible as bad character because they are probative of the defendant’s state of mind (including sexual interest and the kind of sexual activity sought) at the relevant time; and second, that redaction of “extreme” language is not necessarily required where it would distort the overall tenor of the communications and therefore the jury’s ability to assess the defendant’s mindset “in the round,” subject always to the trial judge’s fairness balancing.

Summary of the Judgment

The Court rejected the contention that the text messages were too prejudicial or insufficiently relevant to be admitted. It identified two important matters in issue to which the messages were plainly relevant:

  • Purpose on the night: Whether the appellant was looking for sexual activity (as the prosecution alleged, supported by CCTV of kerb-crawling and picking up lone women) or merely taking “me time.” The texts demonstrated systematic engagement with sex workers and the logistics of meeting them, including in his car.
  • Mindset and sexual interest: The communications were probative of what kind of sexual activity the appellant liked—particularly rough, condomless oral sex with strangers—aligning with the prosecution’s narrative of how the offence was carried out and undermining the appellant’s coincidence-based account.

Although the trial judge at points used the language of “propensity to rape,” the Court of Appeal emphasised that the principal rationale for admission was not to show a general propensity to rape, but to prove state of mind, sexual interest, and the likelihood that the appellant acted in accordance with the fantasies he had been expressing. The Court accepted there were aspects of the judge’s jury directions that could have been improved but concluded that no unfairness resulted and, in any event, the case against the appellant was strong enough that the convictions were safe even if the evidential ruling had been wrong.

Legal Framework and Issues

The bad character regime under the CJA 2003 provides:

  • Section 98(a): Excludes from “bad character” evidence that “has to do with” the alleged facts of the offence charged.
  • Section 101(1)(d): Admits a defendant’s bad character evidence if it is relevant to an important matter in issue between the defendant and the prosecution.
  • Section 103: Gives illustrations of “important matter in issue,” including propensity to commit offences of the kind charged or propensity for untruthfulness. It is not exhaustive; state of mind, motive, intent, and sexual interest can also qualify as important matters in issue under s.101(1)(d).
  • Section 101(3): Even where admissible under a gateway, evidence must be excluded if its admission would have such an adverse effect on the fairness of the proceedings that it ought not to be admitted. The trial judge conducts this balancing exercise.

In Shah, the prosecution initially considered whether the messages were so connected to the alleged facts as not to be bad character under s.98(a). Ultimately, the judge and the parties proceeded on the basis that the messages were bad character, and the Court of Appeal assessed their admissibility under s.101(1)(d), with the fairness safeguards engaged.

Analysis

Precedents Cited and Context

The judgment does not cite specific previous authorities. Nonetheless, its reasoning coheres with established principles in leading bad character jurisprudence under the CJA 2003:

  • It aligns with the broad construction of s.101(1)(d) recognised in the Court of Appeal’s jurisprudence: “important matter in issue” is not confined to propensity; it encompasses motive, intent, state of mind and, in sexual cases, evidence bearing on sexual interest and the kind of activity sought. This approach has long been understood as consistent with the statutory text, which deliberately frames s.103 as illustrative rather than exhaustive.
  • The Court’s emphasis on the balancing exercise mirrors the settled requirement under s.101(3) that even probative bad character evidence must be excluded where its prejudicial impact would render the trial unfair. The Court accepted the evidence was extreme in places but found that its probative force—particularly on state of mind and the coincidence issue—outweighed the prejudice.
  • The cautionary observations about jury directions echo the familiar guidance that judges should carefully identify the legitimate use(s) of bad character and avoid inviting forbidden reasoning (for example, “propensity to rape” where propensity has not been proved). The Court noted room for improvement but found no miscarriage of justice.

Note on statutory references: Paragraph 34 of the judgment states that “Section 103 requires exclusion.” In strict statutory terms, the exclusionary safeguard is in s.101(3). Section 103 sets out examples of “important matter in issue” (propensity, etc.) and does not itself contain the exclusion test. The Court’s reasoning, however, clearly applied the correct fairness balancing principle.

Legal Reasoning: Why the Messages Were Admissible

The Court’s reasoning can be distilled into the following steps:

  1. Relevance to disputed purpose on the night: The appellant claimed he was out for “me time,” not seeking sex. The messages showed extensive, recent contact with sex workers, planning logistics that included using his car, and a strong interest in specific sexual acts. Together with CCTV and other evidence, this was capable of rebutting his account and supporting the prosecution’s case that he was kerb-crawling for sexual opportunities.
  2. Mindset and sexual interest as “important matters in issue”: The messages illuminated the appellant’s sexual desires—rough, condomless oral sex with strangers—corresponding with the prosecution’s account of what happened in the car. The Court treated this as state-of-mind evidence admissible under s.101(1)(d), not to prove a propensity to rape as such, but to show the kind of sex he was actively seeking to experience. This closed the distance between fantasy and alleged conduct in a way probative of whether he committed the act charged.
  3. “Whole picture” approach to presentation: The Court accepted that showing the messages “in the round”—including their more extreme aspects—gave the jury the necessary context to evaluate the appellant’s mindset, and that selective redaction risked distorting that context. Fairness was preserved by careful judicial directions, the overall strength of the case, and the fact that the jury acquitted on the strangulation count, suggesting they were not improperly swayed.
  4. Jury directions—imperfections but no unfairness: Although the trial judge could have framed the directions more precisely (for example, making clear that the messages were not to be relied upon as proof of a general propensity to rape, and cautioning against moral prejudice), the Court held that any imperfections did not render the trial unfair or the convictions unsafe.
  5. Safety of the convictions in any event: Even if there had been an error in admitting the messages, the Court would have upheld the convictions in light of the compelling cumulative evidence: CCTV and ANPR, the complainant’s presentation and injuries, DNA findings consistent with oral sex and vomiting, the recovery of the complainant’s phone in the appellant’s bedroom and her clothes in his boot, and the appellant’s inherently improbable account.

Why “Propensity to Rape” Was Not the Right Label Here

A central theme is the distinction between using bad character to prove propensity to commit offences of the kind charged (s.103(1)(a)) and using it to prove other important matters in issue, such as intent, motive, or sexual interest under s.101(1)(d). The Court stressed that the messaging did not demonstrate that the appellant had previously committed rape or that he had a general propensity to rape. Instead, the messages showed:

  • a persistent sexual interest in rough oral sex without a condom;
  • a pattern of cruising and arranging encounters, using his car; and
  • expressions of fantasy using the term “rape” in a sexualised context.

Those features, when placed alongside the independent evidence of what occurred, were probative of the likelihood that he acted in accordance with his stated desires on the night in question.

Impact and Future Significance

The judgment has several practical consequences for sexual offence trials:

  • Broader use of sexual communications as bad character: Prosecutors can rely on pre-incident sexual communications—particularly with sex workers—as evidence of mindset and sexual intent under s.101(1)(d), even if the messages refer to consensual role-play or appear to concern lawful transactions. The key is their probative connection to disputed issues at trial.
  • Redaction is not automatic: Courts may prefer to present the communications “in the round” where redaction would sanitize or misrepresent the tenor of the messages and thereby mislead the jury about the defendant’s state of mind. This does not mean anything goes; judges must still balance probative value against prejudice under s.101(3).
  • Sharper jury directions: Judges should carefully explain the legitimate relevance of such messaging (e.g., to state of mind or sexual interest), make clear that it cannot be used to convict on the basis of bad character alone, and caution against being influenced by “extreme or deviant” elements that do not directly bear on the issues the jury must decide.
  • Defence strategy: Defence teams will need to challenge the probative pathway—by demonstrating that the messages are temporally remote, dissimilar in kind, mere fantasy, or that their prejudicial content substantially outweighs their probative value. Proposing targeted and principled redactions, coupled with limiting directions, remains crucial.
  • Consistency with scientific and circumstantial evidence: Messaging evidence will carry greatest weight where it dovetails with “real world” proof—CCTV/ANPR, forensic findings, witness accounts—thereby converting what might otherwise be dismissed as mere fantasy into cogent circumstantial evidence of conduct.

Complex Concepts Simplified

  • Bad character evidence: Evidence about a defendant’s past conduct or disposition that is not part of the alleged facts of the current offence. Admissible through statutory “gateways,” including where it is relevant to an important issue in the case (s.101(1)(d)).
  • Section 98(a) CJA 2003: Carves out material that “has to do with” the alleged offence; such material is not treated as “bad character” at all. The prosecution initially floated this but ultimately treated the messages as bad character and relied on s.101(1)(d).
  • Section 101(1)(d): Allows bad character evidence if it helps prove an important matter in issue, which can include motive, intent, state of mind, or sexual interest—beyond mere propensity.
  • Section 103: Provides examples of “important matter in issue,” most notably propensity to commit offences of the kind charged or to be untruthful. It is illustrative, not exhaustive.
  • Section 101(3): Even if evidence is admissible through a gateway, it must be excluded if admitting it would have such an adverse effect on the fairness of the proceedings that it ought not to be admitted. This is the key fairness safeguard.
  • Propensity vs. Mindset: Propensity means a tendency to behave in a particular way (e.g., to commit sexual offences). Mindset/state of mind refers to what the defendant was thinking or intending (e.g., seeking rough oral sex), which can be probative without proving a general tendency to offend.
  • ABE interview: “Achieving Best Evidence” interview is a structured police interview method for vulnerable witnesses, designed to obtain high-quality accounts for use in criminal proceedings.
  • ANPR: Automatic Number Plate Recognition—technology used to track vehicle movements across camera networks.

Observations on Jury Directions

The Court acknowledged scope to improve the directions:

  • The judge should have avoided implying that the messages could establish a “tendency to commit rape” where that was not the prosecution’s case.
  • He might have explained more exactly how the messages were relevant (mindset/state of mind) rather than framing them in propensity terms.
  • A specific caution against allowing “extreme or deviant” language to inflame prejudice would have been desirable, alongside the standard warning not to convict wholly or mainly because of bad character.

These are useful reminders for trial judges managing sensitive sexual messaging material. In Shah, however, they did not undermine the safety of the verdicts.

Practical Guidance for Practitioners

  • For the prosecution:
    • Articulate a clear probative pathway under s.101(1)(d): show how the messages go to state of mind, motive, intent, or to rebut a specific account (e.g., “me time”).
    • Explain why redaction would materially distort the “tenor” of communications and impede the jury’s understanding of mindset.
    • Propose tailored directions that identify legitimate uses of the evidence and warn against forbidden reasoning.
  • For the defence:
    • Challenge the connection between fantasy and conduct; emphasise consensual nature of sex work communications and the difference from the alleged offence.
    • Press the s.101(3) fairness balance; seek redaction where particular passages add moral prejudice without incremental probative value.
    • Request precise directions that confine the use of the messages to demonstrated issues (e.g., mindset) and forbid inferences of general criminal disposition.
  • For judges:
    • Identify with precision the issue(s) to which the messages are relevant.
    • Record the s.101(3) balancing, acknowledging prejudicial features and explaining why probative worth prevails.
    • Direct the jury in terms that avoid “propensity creep” and warn explicitly against being swayed by the shock value of extreme content.

Conclusion

R v Shah confirms that sexual messaging with sex workers—however framed as consensual or couched as fantasy—may be admitted as bad character under s.101(1)(d) to prove state of mind and sexual intent, and to rebut anodyne alternative explanations of the defendant’s conduct on the night. The Court’s acceptance that the jury could see the messages “in the round” (including extreme references) is significant: redaction is not obligatory where it would misrepresent the overall tenor of the evidence. At the same time, judges must maintain rigorous s.101(3) scrutiny and provide carefully calibrated directions to prevent unfair prejudice.

The decision will influence future sexual offence trials where pre-incident digital communications are available. It underscores that the probative force of such material is greatest when it dovetails with the objective evidence—CCTV/ANPR, forensics, eyewitness accounts—and when the prosecution’s use is tightly tethered to clear issues in dispute. Shah thus meaningfully develops the law’s practical treatment of sexual communications as bad character, emphasizing mindset rather than propensity, and provides concrete guidance on presentation and judicial directions to preserve fairness.

Appeal outcome: Dismissed. Evidence properly admitted; convictions safe in any event.

Case Details

Year: 2025
Court: England and Wales Court of Appeal (Criminal Division)

Comments