The Wessels Case: What the Numbers Actually Tell Us About Intent

On October 17, 2025, something happened in a ruck during Bulls vs Connacht. Connacht flanker Josh Murphy immediately struck Bulls hooker Jan-Hendrik Wessels in the head and told the referee that Wessels had grabbed his testicles. Murphy received a 20-minute red card (later rescinded). Wessels received nothing during the match—the referee and TMO reviewed footage roughly 10 times and found no conclusive evidence.

Two weeks later, a disciplinary panel consisting of three Welsh officials found Wessels guilty of violating Law 9.27 (genital grabbing) and banned him for nine weeks. The panel acknowledged there was no "irrefutable visual confirmation" but convicted based primarily on Murphy's testimony and the citing commissioner's subjective interpretation of "unnatural arm movement."

The Bulls are appealing. South African rugby media are furious about the precedent. And everyone has an opinion about what "really" happened—except nobody's actually done the probability properly.

So we did.

Why Bayesian Analysis for a Rugby Disciplinary Case?

This isn't about taking sides. It's about quantifying uncertainty using evidence. A Bayesian framework lets us start with baseline probabilities (how often do these things actually happen in rugby?), incorporate each piece of evidence systematically (Murphy's reaction, the video absence, testimony details, expert assessments), and compute posterior probabilities that tell us what we should believe given what we know.

The key insight: We need to distinguish between three separate hypotheses:

H1: Deliberate Foul Play - Wessels intentionally grabbed Murphy's genitals (violates Law 9.27)
H2: Accidental Contact - Unintentional genital contact in the ruck chaos (does NOT violate Law 9.27)
H3: No Contact - Murphy misperceived or made a false accusation (Wessels innocent)

The disciplinary panel treated this as binary: either guilty or innocent. But that framework misses the critical middle ground—accidental contact. Law 9.27 prohibits deliberate acts: "grab, twist, or squeeze." These verbs require intent. Accidental brush contact during a ruck, while uncomfortable, isn't a Law 9.27 violation.

Let's see what the evidence actually tells us.

The Prior Probabilities: Base Rates Matter

Before examining any evidence from this specific incident, we need baseline probabilities. How common are these scenarios in professional rugby?

Law 9.27 violations are vanishingly rare. Documented genital grabbing cases number roughly 1-2 per year globally at the elite level. That's approximately 0.05-0.1% of all professional matches. The most famous recent case—Joe Marler grabbing Alun Wyn Jones in 2020—involved clear video evidence and Marler's admission that something occurred (though he disputed whether he "grabbed and squeezed"). Marler received 10 weeks.

Accidental genital contact in rucks is more common. Rucks involve bodies pressed together, hands near opponents' lower bodies, and chaotic movement. Former England player Danny Care stated: "I've lost track of how many times players have touched my genitals in rugby." Accidental contact happens; it's rarely reported because everyone understands the context.

False accusations or misperceptions exist but are uncommon. Research on sexual misconduct allegations generally shows false reports in the 2-8% range. Murphy had a conflict of interest (he faced his own hearing for striking Wessels), but that doesn't mean his accusation was false—it means we need to weigh this factor.

Given the ruck context and Murphy's immediate reaction (suggesting something provoked him), our prior probabilities are:

P(H1: Deliberate) = 10% - Rare but possible given the specific circumstances
P(H2: Accidental) = 30% - More common in ruck chaos
P(H3: No Contact) = 60% - Most rucks involve neither

The 40% combined probability that some contact occurred reflects Murphy's extreme reaction, which is unlikely without genuine stimulus.

The Evidence: What Actually Happened

Evidence 1: Murphy's Immediate Physical Retaliation

Murphy struck Wessels in the head immediately after the alleged incident. This is the single most important piece of evidence in the entire case.

Professional rugby players understand red card consequences intimately. Striking an opponent guarantees a red card, leaves your team a player down, and results in suspension. Murphy is 30 years old with extensive professional experience. For him to throw a punch immediately and then explain to the referee why he did it—that behaviour is stunningly irrational unless he was genuinely, severely provoked.

The probabilities:

P(Immediate strike | Deliberate grab) = 75% - Very likely to react if deliberately grabbed
P(Immediate strike | Accidental contact) = 35% - Might still strike if felt contact
P(Immediate strike | No contact) = 3% - Essentially irrational without provocation

Likelihood Ratios:

Deliberate vs No Contact: 25.0 (strongly supports deliberate over nothing)
Accidental vs No Contact: 11.7 (strongly supports accidental over nothing)

This evidence tells us: Something almost certainly happened. Murphy's reaction is 25 times more likely if Wessels deliberately grabbed him than if nothing happened. But crucially, it's also 11.7 times more likely if accidental contact occurred than if nothing happened.

Murphy's strike doesn't distinguish well between deliberate and accidental—it just tells us contact is highly probable.

Evidence 2: Murphy's Specific Language—"Grabbed and Squeezed"

Murphy didn't say "touched me" or "made contact." He specifically told the referee: "I felt the player grab my testicles and that's when I reacted immediately. That's why I reacted, because of how he grabbed and squeezed my testicles."

That language matters. "Grabbed and squeezed" implies intentional action—active verbs, purposeful behaviour. You don't accidentally "grab and squeeze" someone. You might accidentally brush against them, make incidental contact, or touch them in the chaos of a ruck. But "grabbed and squeezed" describes deliberate action.

The probabilities:

P("Grabbed & squeezed" testimony | Deliberate) = 85% - Fits deliberate act
P("Grabbed & squeezed" testimony | Accidental) = 25% - Less likely to describe accident this way
P("Grabbed & squeezed" testimony | No contact) = 20% - False claims can still be specific

Likelihood Ratio (Deliberate vs Accidental): 3.4

This evidence is critical for distinguishing intent from accident. Murphy's specific language tips the balance toward deliberate foul play rather than accidental contact. If it was just an accidental brush, you'd expect vaguer testimony: "he touched me," "I felt contact." Murphy was emphatic and specific about grabbing and squeezing.

Evidence 3: No Conclusive Video Evidence

Here's where it gets uncomfortable for the conviction. The referee and TMO reviewed footage approximately 10 times during the match. The disciplinary panel reviewed all available angles. The Bulls' coach confirmed no new video evidence was presented at the hearing.

Result: No conclusive visual confirmation.

The panel acknowledged this explicitly in their judgment. They convicted anyway based on testimony and the citing commissioner's subjective interpretation of arm positioning.

The probabilities:

P(No conclusive video | Deliberate grab) = 35% - Could be obscured in ruck
P(No conclusive video | Accidental contact) = 70% - Harder to see on video
P(No conclusive video | No contact) = 98% - Expected if nothing happened

Likelihood Ratios:

Deliberate vs No Contact: 0.36 (supports no contact over deliberate)
Accidental vs No Contact: 0.71 (weakly supports no contact over accidental)

The absence of video evidence weighs most heavily against deliberate foul play. Think about it: if Wessels deliberately grabbed Murphy's genitals with enough force to squeeze them, that's a purposeful, sustained action. Modern professional rugby has extensive multi-camera coverage. The fact that 10+ reviews by professional match officials found nothing conclusive suggests either:

(a) The action was extremely subtle (more consistent with accident than deliberate grab) (b) It was completely obscured (possible but unlikely given multiple angles) (c) It didn't happen as alleged (possible)

A deliberate "grab and squeeze" should produce more visible evidence than accidental contact. The video absence is a significant problem for the prosecution's case.

Evidence 4: The Citing Commissioner's Interpretation

The citing commissioner wrote: "The movement of Wessels' left arm, away from his body and toward Murphy's groin, is not only unnatural but also unnecessary in the context of ruck engagement. His backward glance further suggests awareness."

This is subjective expert interpretation. Citing commissioners have specialized training in identifying foul play. They're right most of the time. But they're not infallible, and crucially, this assessment isn't corroborated by video.

The probabilities:

P(CC says "unnatural" | Deliberate) = 80% - Expert likely catches genuine foul play
P(CC says "unnatural" | Accidental) = 35% - Might misinterpret accident as suspicious
P(CC says "unnatural" | No contact) = 10% - Unlikely to completely misread innocent action

Likelihood Ratios:

Deliberate vs No Contact: 8.0 (strongly supports deliberate)
Accidental vs No Contact: 3.5 (moderately supports accidental)

The citing commissioner's assessment carries weight—it's an expert opinion from someone trained to spot foul play. But without video corroboration, we're relying on one person's interpretation of body positioning in a chaotic ruck. Arm movements in rucks are inherently "unnatural" because you're binding, pushing, reaching for the ball, and maintaining balance while other players drive into you.

Evidence 5: Wessels' Brief Testimony

The panel noted that Wessels' testimony was "brief" compared to Murphy's "long" testimony. Wessels denied the allegation but didn't provide elaborate alternative explanations.

The panel apparently interpreted this brevity as diminishing Wessels' credibility. That interpretation is statistically backwards.

The probabilities:

P(Brief denial | Deliberate) = 60% - Guilty parties might keep it short to avoid contradictions
P(Brief denial | Accidental contact) = 70% - If unaware, nothing to explain
P(Brief denial | No contact) = 80% - Innocent defendants often have little to say

Likelihood Ratios:

Deliberate vs No Contact: 0.75 (weakly supports no contact)
Accidental vs No Contact: 0.87 (weakly supports no contact)

Wessels' brief testimony is actually slightly more consistent with innocence or accident than with guilt. Think about it: if nothing happened, what should Wessels say beyond "I didn't do it"? If the contact was accidental and he wasn't aware of it, what can he explain? Innocent defendants and those involved in unknowing accidents don't have elaborate stories—they have simple denials because there's nothing to narrate.

The panel's interpretation that testimony length indicates credibility is problematic. Elaborate testimony can indicate either truthfulness (genuine memory of events) or fabrication (overcompensating with details). Brief testimony can indicate either guilt (avoiding details) or innocence (nothing to explain). Length alone doesn't distinguish.

Evidence 6: Wessels' Clean Disciplinary Record

Wessels, age 24, has nine Springbok caps and no prior disciplinary violations. The panel granted 25% mitigation based on his "good conduct and good record."

The probabilities:

P(Clean record | Deliberate) = 45% - Could be first offense
P(Clean record | Accidental) = 90% - Likely if innocent/accident
P(Clean record | No contact) = 95% - Very likely if innocent

Likelihood Ratios:

Deliberate vs No Contact: 0.47 (supports no contact over deliberate)
Accidental vs No Contact: 0.95 (weakly supports no contact)

Clean record weakly supports innocence or accident. It doesn't prove anything—first-time offenses exist—but it's more consistent with someone who doesn't commit genital grabbing fouls than someone who does.

The Posterior Probabilities: What the Evidence Actually Shows

After sequentially updating through all evidence using Bayes' Theorem (the proper mathematical framework for incorporating evidence), here's where we end up:

P(H1: Deliberate Foul Play) = 52.7%

P(H2: Accidental Contact) = 44.3%

P(H3: No Contact) = 2.9%

P(Any Contact Occurred) = 97.1%

Let's be extremely clear about what these numbers mean:

1. Contact almost certainly occurred (97.1%)

Murphy's immediate strike is so unlikely without genuine provocation (only 3% probability) that we can be highly confident something happened in that ruck. The probability Murphy completely fabricated the incident or totally misperceived nothing as genital contact is tiny (2.9%).

Murphy was provoked. Something happened to him.

2. Deliberate vs Accidental is nearly a coin flip (53% vs 44%)

Here's the critical finding: while we're confident contact occurred, whether that contact was deliberate or accidental is profoundly uncertain.

The evidence tips slightly toward deliberate (52.7%) because of Murphy's specific "grabbed and squeezed" language. But there's a 44.3% probability it was accidental contact—nearly as high as the probability it was deliberate.

This isn't "beyond reasonable doubt." This isn't even "clear and convincing evidence." This barely clears the "balance of probabilities" threshold (>50%) that rugby uses.

3. The panel's decision barely meets the conviction standard

Rugby disciplinary proceedings require conviction to meet "balance of probabilities"—more likely than not, above 50%. The 52.7% probability of deliberate foul play technically exceeds this threshold.

By 2.7 percentage points.

Meanwhile, there's a combined 47.3% probability that Wessels did NOT commit deliberate foul play (44.3% accidental + 2.9% no contact). That's substantial reasonable doubt.

4. Law 9.27 requires deliberate action

This is the legally critical point. Law 9.27 prohibits "grabbing, twisting, or squeezing the genitals." Those verbs—grab, twist, squeeze—imply intentional action. You can't accidentally "grab" someone in the sense the law means. You can accidentally make contact, brush against someone, or touch them incidentally. But "grabbing" is a purposeful act.

The relevant probability for a Law 9.27 conviction is therefore 52.7% (deliberate), not 97.1% (any contact). Accidental contact, no matter how uncomfortable for Murphy, shouldn't satisfy Law 9.27's intentionality requirement.

The Sensitivity Analysis: How Uncertain Are We?

To test robustness, we ran the analysis under different assumptions:

Scenario	Assumptions	P(Deliberate)	P(Accidental)
Base Case	Balanced estimates	52.7%	44.3%
Higher Accidental Prior	Rucks make accidents common	35.7%	62.6%
Strong Language = Intent	"Grabbed & squeezed" strongly indicates deliberate	71.8%	24.2%
No Video Favors Accident	Deliberate acts should show on video	38.9%	57.3%
Most Favorable to Wessels	Combined skeptical assumptions	12.9%	81.5%

Range of P(Deliberate): 12.9% to 71.8%

That's an enormous range. Depending on how you weight the video evidence absence, Murphy's language, and the prior probability of accidental contact in rucks, the posterior probability of deliberate foul play varies by nearly 60 percentage points.

Critically: in ALL scenarios, some form of contact (deliberate OR accidental) remains highly probable (94.4% to 98.3%). Murphy was almost certainly provoked. But whether that provocation constituted a deliberate Law 9.27 violation is genuinely uncertain.

The panel convicted based on the narrow edge of one interpretation. Under equally plausible assumptions, Wessels is more likely innocent of deliberate foul play.

What Went Wrong: The Conflation Error

Here's what we think happened at the disciplinary hearing:

The panel correctly identified that contact almost certainly occurred. Murphy's reaction is too extreme, too immediate, and too specifically described to be complete fabrication. Something provoked him.

But the panel then conflated "contact occurred" (97.1% probable) with "deliberate foul play" (52.7% probable).

Those are not the same thing.

Murphy saying "he grabbed my testicles" might be an accurate description of what it felt like from his perspective during the chaos of a ruck, even if the contact wasn't deliberate from Wessels' perspective. Murphy's subjective experience of being grabbed doesn't necessarily mean Wessels deliberately grabbed him—it might mean Wessels' hand ended up in the wrong place during ruck engagement, and Murphy experienced that as grabbing.

The citing commissioner's interpretation of "unnatural arm movement" might accurately describe that Wessels' arm moved toward Murphy's groin area—but that doesn't prove the contact was intentional rather than an artifact of ruck positioning.

The panel appears to have reasoned: "Murphy was clearly provoked (true) → therefore Wessels committed deliberate foul play (doesn't follow)."

The middle hypothesis—accidental contact that Murphy experienced as genital grabbing—was apparently never seriously considered. The panel treated it as binary: either Wessels deliberately grabbed Murphy (guilty) or nothing happened and Murphy made it up (innocent). That framework forced a guilty verdict once they accepted contact occurred.

But the three-hypothesis framework reveals the truth: contact occurred (97% confident), but whether it was deliberate is essentially a coin flip (53% vs 44%).

The Precedent Problem

This conviction creates troubling precedent for rugby discipline:

Kyle Sinckler (2021): Franco Mostert accused Sinckler of biting during the British & Irish Lions tour. Allegation dismissed by the disciplinary committee—no clear evidence.

Bongi Mbonambi (2023): Tom Curry accused Mbonambi of racial slur at Rugby World Cup. World Rugby investigated but couldn't proceed—no audio/video evidence.

Jan-Hendrik Wessels (2025): Josh Murphy accused Wessels of genital grabbing. Convicted despite no conclusive video evidence, based primarily on testimony.

The Wessels conviction is inconsistent with recent precedent. In both the Sinckler and Mbonambi cases, serious allegations were dismissed because conclusive evidence was lacking. In the Wessels case, the panel convicted despite explicitly acknowledging no "irrefutable visual confirmation."

The difference? Murphy struck Wessels immediately, creating circumstantial corroboration that something happened. But "something happened" isn't the same as "deliberate Law 9.27 violation."

If this precedent stands, rugby now convicts on testimony alone without video corroboration, even when match officials reviewed footage 10+ times during the game and found nothing conclusive. That's a low evidentiary bar for a nine-week suspension that damages a player's career and reputation.

What the Bulls Should Argue on Appeal

The appeal should focus on the accident hypothesis that the panel apparently never considered:

1. Insufficient evidence of intent

While contact likely occurred (97%), only 52.7% confidence exists that it was deliberate—barely above the conviction threshold. The 44.3% probability of accidental contact represents reasonable doubt that should favor the accused.

2. Law 9.27 requires deliberate action

"Grab," "twist," and "squeeze" are volitional verbs requiring intent. The panel conflated "contact occurred" with "deliberate foul play," but Law 9.27 doesn't prohibit accidental genital contact during ruck engagement—it prohibits deliberate grabbing.

3. Video evidence standard

The absence of conclusive video after 10+ reviews by professional match officials is unprecedented for a Law 9.27 conviction. The Joe Marler case (2020) had clear video plus Marler's admission. The Wessels conviction without video creates a dangerous precedent that testimony alone suffices.

4. Wessels' testimony misinterpreted

The panel apparently treated Wessels' brief testimony as indicating guilt or low credibility. The statistical analysis shows brief denial is actually more consistent with innocence or unknowing accident (likelihood ratio 0.75). The panel's reasoning was backwards.

5. Precedent inconsistency

Sinckler and Mbonambi cases dismissed without evidence. Why different standard here?

The appeal isn't arguing Wessels is definitely innocent—it's arguing that the evidence doesn't support deliberate foul play to the required standard when a substantial probability of accidental contact exists.

The Bigger Picture: What This Means for Rugby

The Wessels case sits at the intersection of several pressures on modern rugby:

Taking sexual misconduct seriously: The #MeToo era rightly emphasizes believing victims and taking allegations seriously. Murphy's complaint deserves investigation and due process.

Maintaining evidence standards: But taking allegations seriously doesn't mean abandoning evidence requirements. The accused still deserves protection against wrongful conviction, especially when testimony alone produces barely-better-than-chance probabilities of deliberate wrongdoing.

Dark arts vs modern standards: Rugby historically tolerated aggressive "dark arts" in rucks—gouging, grabbing, niggling. Modern rugby rightly cracks down on this behaviour. But distinguishing deliberate foul play from accidental contact requires better evidence than subjective interpretation of ambiguous video.

Regional politics: The all-Welsh panel judging a South African player accused by an Irish-based player feeds existing tensions about Northern Hemisphere vs Southern Hemisphere rugby governance. Perception of fairness matters even when bias isn't present.

The solution isn't to ignore allegations like Murphy's. It's to recognize that "something happened" and "deliberate foul play" are different claims requiring different evidence. The first can be established on testimony and circumstantial evidence. The second requires more—video corroboration, clear intent indicators, or admission.

Could We Be Wrong?

Absolutely. Our model makes assumptions:

Conditional independence: We assume evidence pieces are independent (Murphy's strike and testimony provide separate information). They're not—both stem from the same alleged event. This could inflate our posteriors.

Likelihood calibration: Our probability estimates (e.g., 75% chance Murphy strikes if deliberately grabbed) are informed judgments based on professional player behavior, but they're not empirically derived from large datasets of genital grabbing incidents (because such datasets don't exist).

Prior uncertainty: The 10% prior for deliberate foul play could be too low (making deliberate foul play seem less likely than it is) or too high (if we're overweighting Murphy's reaction in setting the prior).

Accidental contact base rate: We set 30% prior for accidental contact, but this is essentially a guess. If accidental genital contact in rucks is actually very rare (say, 5%), then the posterior probability of deliberate foul play would increase substantially.

If you disagree with our calibrations, build your own model. If you have better data on base rates for accidental ruck contact or likelihood ratios for specific evidence, update the posteriors. That's how Bayesian reasoning works.

But the core insight remains: the panel appears to have conflated "contact occurred" with "deliberate foul play," and those probabilities are dramatically different (97% vs 53%).

The Bottom Line

Jan-Hendrik Wessels was convicted of genital grabbing and banned for nine weeks based on:

Josh Murphy's testimony (credible but conflict of interest)
Citing commissioner's subjective interpretation (expert opinion, no video corroboration)
Murphy's immediate strike (proves something happened, doesn't prove deliberate grab)
No conclusive video evidence (weights against conviction, not for it)
Wessels' brief testimony (actually slightly favors innocence/accident)

The Bayesian analysis shows:

97.1% probability contact occurred - Murphy was almost certainly provoked
52.7% probability contact was deliberate - barely above conviction threshold
44.3% probability contact was accidental - substantial reasonable doubt
Conviction based on conflating "contact" with "deliberate foul play"

Rugby's disciplinary system requires only "balance of probabilities" (>50%). The conviction technically meets that standard. But barely. With enormous uncertainty. And an alternative hypothesis (accidental contact) that's nearly as probable as the conviction hypothesis (deliberate foul play).

Is that good enough for a nine-week ban that will keep Wessels out of rugby through the Springbok northern hemisphere tour? That damages his reputation with accusations of sexual misconduct? That sets precedent for conviction without video evidence?

The mathematics says: maybe, but only just barely, and you'd better be damn sure you're not confusing "something happened" with "deliberate foul play."

The panel wasn't sure enough. Their own judgment admits no "irrefutable visual confirmation." They convicted anyway.

The Bulls are right to appeal. Not because Wessels is definitely innocent—he might be guilty. But because 52.7% confidence in deliberate foul play, when there's a 44.3% probability of accidental contact that doesn't violate Law 9.27, isn't a strong enough foundation for conviction in a serious disciplinary matter.

The evidence shows Murphy was almost certainly provoked. It doesn't show—at least not with sufficient confidence—that Wessels deliberately grabbed him rather than accidentally contacted him in ruck chaos.

That distinction matters. The panel missed it. The mathematics doesn't.

Methodology note: Analysis constructed using Bayesian probability framework with three hypotheses (deliberate foul play, accidental contact, no contact). Priors calibrated from Law 9.27 violation base rates (~0.05-0.1% of matches), false accusation research (2-8%), and accidental contact estimates in ruck contexts. Likelihood ratios derived from professional player behavior patterns, match official review processes, expert testimony standards, and disciplinary precedent. Evidence includes Murphy's immediate physical retaliation, specific testimony language, video evidence absence, citing commissioner interpretation, Wessels' testimony brevity, and clean disciplinary record. Posterior probabilities computed through sequential Bayesian updating. Sensitivity analysis conducted across five scenarios with varying assumption sets. This is subjective probability quantification grounded in available evidence, not objective certainty. Model assumptions detailed in technical appendix.

Sources

All information accessed October 2025 from publicly available reporting and official rugby sources.

Case Documentation:

SA Rugby Magazine: "Revealed: Images that convicted Jan-Hendrik" (Oct 25, 2025)
Reuters: "Springbok Wessels handed nine-match ban for genital grab" (Oct 23, 2025)
BBC Sport: "Wessels receives nine-game ban for genital grab" (Oct 23, 2025)
SuperSport: "Disciplinary decision: Jan Hendrik Wessels" (Oct 22, 2025)
Daily News: "Bulls coach Ackermann: There was no new evidence at the hearing" (Oct 25, 2025)

Historical Precedents:

CNN: "Joe Marler banned for 10 weeks after grabbing opponent's genitals" (March 12, 2020)
Wales Online: "Joe Marler breaks silence on grabbing Alun Wyn Jones' genitals" (June 29, 2020)
BBC News: "Joe Marler ban: Is genital grabbing a problem in rugby?" (March 13, 2020)

Disciplinary Framework:

World Rugby: "Regulation 17: DISCIPLINE - FOUL PLAY" (May 4, 2020)
World Rugby: "Laws of the Game - Law 9: Foul play" (November 3, 2020)
World Rugby: "Writing Guide: The Citing Narrative" (Citing Commissioner Training)

Statistical Methodology:

Research literature on false accusation base rates (Lisak et al., FBI statistics)
Professional rugby disciplinary precedent analysis
Bayesian inference methodology applied to legal/evidentiary contexts

Technical Note: Likelihood ratios and prior probabilities represent informed estimates based on rugby disciplinary patterns, professional player behavior research, and statistical best practices for uncertain evidence. Sensitivity analysis demonstrates robustness across assumption sets. Full technical methodology and probability calculations available in supplementary materials.