The criminal justice system can be daunting, but you don’t need to go through it alone. Our Criminal lawyers are here to guide you every step of the way.
Contact Our Firm
Joseph A. Neuberger, Michael P. Bury, and Diana Davison, Neuberger & Partners LLP, Sexual Assault Defence Lawyers
In R. v. Hornsby, 2026 ONCA 604, the Court of Appeal for Ontario returned to a familiar but stubbornly difficult question: at what point does a trial judge’s assessment of competing testimony stop being ordinary fact-finding and become an unfair trial? The answer matters, because the line separates a credibility finding that appellate courts will defer to from a “demonstrably flawed assessment methodology” that requires a new trial. Writing for a unanimous panel, Roberts J.A. concluded that the trial judge had crossed that line, quashed all four convictions, and ordered a new trial.
01 • THE CASE IN BRIEF
The appellant, Everton Hornsby, was convicted after a judge-alone trial of one count of assault, one count of assault with a weapon (a laundry basket), and two counts of sexual assault. Each count involved his spouse at the time, and the incidents spanned roughly December 2018 to June 2019. He received a global custodial sentence of 42 months.
The couple had met in 2014, married in 2017, and separated in stages through 2018 and 2019. The relationship deteriorated after the complainant admitted to infidelity, and the two continued to share a home until a June 2019 confrontation ended with a police call and the appellant’s arrest for mischief. The complainant reported the underlying incidents days later.
As is common in domestic and sexual assault trials, the case came down to the competing testimony of two people. The complainant and the appellant gave sharply opposed accounts of essentially every incident, and there was little independent evidence to corroborate either account. The trial judge believed the complainant, disbelieved the appellant, and convicted on all counts. The entire appeal turned on how she arrived at that decision.
02 • THE DOCTRINE: A HIGH THRESHOLD, AND FOR GOOD REASON
The appellant’s central argument was “uneven scrutiny” — the contention that the trial judge assessed the defence evidence more harshly than the Crown evidence, rendering the trial unfair.
It is worth being candid about how hard this argument is to win. The Court of Appeal openly acknowledged that uneven scrutiny is a “notoriously difficult argument to prove,” precisely because credibility findings are deferred to the trial judge and appellate courts rarely interfere with those findings. Appellate courts are, in the Court’s words, justifiably skeptical of what may be veiled attempts to relitigate credibility one level up. An appellant cannot simply point to a credibility finding he dislikes; he must identify something in the reasons or the record showing that the trial judge actually applied different standards to the two witnesses.
Framed a different way, the appellant must show that the conviction rested on a flawed reasoning process, not merely on a difference of opinion as to how the evidence should have been viewed. Uneven scrutiny, the Court noted, is often less a freestanding error than a label for a cluster of related problems — insufficient reasons, misapprehension of evidence, a reversed burden of proof — that together reveal a skewed analysis.
To ground the analysis, Roberts J.A. catalogued the recurring fact patterns in which the argument on appeal succeeded. Among them: seizing on the “leanest opportunities” to reject the accused’s evidence while forgiving similar or worse problems in the complainant’s; treating the accused’s detailed testimony as suspiciously “contrived” while crediting the complainant’s equally detailed account as a mark of reliability; rejecting the accused’s evidence for a reason that applies just as forcefully to the complainant; grounding an adverse credibility finding in stereotype rather than evidence; requiring the accused to explain himself and thereby reversing the burden of proof; and fixating on minor inconsistencies in the defence account while leaving central inconsistencies in the complainant’s account unexamined.
Crucially, the Court reaffirmed the stakes. Even while the evidence at trial could have supported a conviction, a trial conducted with different standards for prosecution and defence evidence is not a fair trial — and an unfair trial is a miscarriage of justice.
03 • WHERE THE TRIAL JUDGE WENT WRONG
The Court found that this case fit squarely within those recognised patterns. Several threads ran through the reasons.
The trial judge treated the complainant’s willingness to concede unhelpful facts as enhancing her credibility. Yet when the appellant made comparable admissions, he received no corresponding credit. The same feature of testimony was a virtue for one witness and a nullity for the other, and the reasons never explained the difference.
On one incident there was some agreement as to what had happened but dispute about how and why it happened. The appellant said the complainant taunted him while she smoked marijuana and that he masturbated at a distance when challenged to do so; the complainant said he returned during an argument and ejaculated on her without consent. The trial judge’s evident disapproval of the act itself — which she described in strong terms — appears to have closed her mind to whether the complainant’s own evidence might partly corroborate the appellant’s version. His account was branded “concocted” and “bizarre” without explanation, even though the trial judge had herself called the complainant’s version the most bizarre of the allegations.
The trial judge repeatedly treated the appellant’s references to the complainant’s marijuana use as attempts to “besmirch” her character. But the complainant admitted using marijuana on other occasions, and — importantly — the Crown never asked the judge to view these references as character assassination, nor did she raise the concern with counsel. More troubling, the marijuana evidence was potentially relevant to the appellant’s account of why the complainant was on the porch at 2:00 a.m. Rather than weigh it as evidence, the trial judge used it to discount his credibility.
A recurring problem was that the trial judge rejected the appellant’s evidence wholesale — except where it happened to support the complainant. Evidence disbelieved when it helped the defence was quietly accepted when it helped the Crown. That selective use of a single witness’s testimony is a hallmark of skewed scrutiny.
On an incident involving a laundry-basket assault, the complainant’s claim that her finger had been broken surfaced only at trial, was absent from her police statement, and lacked any photographic or medical documentation — even as a photograph of a scraped arm was treated as decisive corroboration. The trial judge never grappled with whether this late-emerging allegation reflected inconsistency or exaggeration. On an alleged bedroom sexual assault, the complainant’s account of how the appellant came to be in her room shifted materially on cross-examination, yet the reasons did not analyse what that shift meant for her reliability. Worse, the trial judge imported her flawed findings from the porch incident to resolve the consent question on the bedroom incident, reasoning from one contested finding to another rather than assessing each count on its own evidence.
Across the counts, the appellant’s evidence — especially his testimony that he had not been angry — was dismissed as “Nonsense!” rather than analysed. On the final assault count, the Court found essentially no meaningful engagement with the defence evidence at all; a photograph was again treated as sufficient evidence, and the appellant’s competing account went unaddressed.
04 • THE STRUCTURAL ERROR: A FIXED LENS
Beneath the count-by-count problems lay a larger methodological one. The trial judge was entitled to reach general conclusions about each witness’s credibility. What she could not do was let those general conclusions replace the discrete analysis each count required, or harden into a fixed lens through which every subsequent incident was judged.
That is what happened. Having assessed the complainant as unfailingly credible and the appellant as unrelentingly not, the trial judge carried those global verdicts into each count, rejecting the appellant’s evidence whenever it conflicted with the complainant’s simply because it conflicted. The Crown must still prove each count beyond a reasonable doubt on its own evidence; a general credibility impression cannot do that work. Because the general assessments were themselves infected by uneven scrutiny, the infection spread to every count that relied on them.
05 • WHY IT MATTERS
Hornsby does not lower the bar for uneven scrutiny arguments — the Court took pains to reaffirm how demanding that bar remains, and how much deference trial credibility findings deserve. What the decision offers instead is a clear illustration of the difference between a credibility finding an appellate court will not touch and a reasoning process it cannot let stand.
The practical lessons are several. Symmetry is the touchstone: a factor that boosts one witness’s credibility (candour, detail, admissions against interest) cannot silently be denied its mirror effect for the other. Moral disapproval of conduct — however warranted — is not a substitute for analysis of whether disputed events occurred as alleged. Judges should be cautious about attributing improper motives (character assassination, for instance) that neither party advanced and that were never tested with counsel. Genuine inconsistencies in the Crown’s case, including allegations that surface only at trial, demand the same critical attention as inconsistencies in the defence’s. And general credibility findings, however firmly held, cannot become a lens that pre-decides every count.
None of this is a finding that the appellant is innocent, or that a properly conducted trial could not convict. The Court was explicit that evidence capable of supporting a conviction is not enough if it was assessed unfairly. The remedy for an unfair process is a fair one: the convictions were quashed, and the matter returns for a new trial.
For trial judges, counsel, and anyone interested in how criminal courts handle he-said-she-said cases, Hornsby is a useful reminder that the fairness of a verdict lives not only in the conclusion but in the even-handedness of the path taken to reach it.
READ THE FULL DECISION HERE: R. v. Hornsby, 2026 ONCA 604
This post is a summary and commentary on a public appellate decision and is not legal advice. Anyone seeking advice about a specific matter should consult a criminal lawyer. Neuberger & Partners LLP, Criminal Lawyers Toronto.
