Why Derby Matches Defy Statistics

League tables are supposed to settle arguments before kickoff. Points difference, expected goals, squad value – these numbers predict outcomes with real accuracy, most weekends. Then a derby arrives, and the model wobbles. 

Betting platforms with dedicated football markets, like Jawhara bets, price these fixtures every week and routinely see derby odds swing harder than any form-based logic would suggest. 

The gap between what a spreadsheet expects and what actually happens on derby day has puzzled analysts for years. Statistics describe most matches well. Derbies are the exception the data quietly struggles to explain.

When the Table Says Nothing About the Result

Form guides are built on one assumption: recent results predict future ones. Derbies are where that assumption gets tested hardest, and it doesn't always survive.

Take the 2024-25 Manchester derby. City won both league meetings that season, 3-0 and 2-1, looking every bit the stronger side. Then United beat them 2-1 in the FA Cup final – the same two squads, the opposite result, months apart.

Derby

Historical Head-to-Head

Where Recent Form Broke Down

Manchester Derby (City vs United)

City dominant across recent league seasons

Won both 2024-25 league meetings, yet lost the FA Cup final 2-1 to the same opponent

Old Firm (Celtic vs Rangers)

Celtic riding a long unbeaten domestic run

September 2025 meeting: Celtic won 3-0, extending the streak regardless of Rangers' summer rebuild

Milan Derby (Inter vs AC Milan)

Inter lead 32 wins to Milan's 24 across the last 71 meetings, 15 draws

April 2025 ended 2-1 to Inter, reinforcing rather than reversing the trend

North London Derby (Arsenal vs Tottenham)

Arsenal lead 89-68, with 55 draws across 211 meetings

Arsenal won 3-2 in April 2025; five months later, the same title-chasing side could only draw 1-1

El Clasico (Barcelona vs Real Madrid)

Widely regarded as club football's most-watched fixture

League position rarely predicts the margin, regardless of which side is topping the table that year

Merseyside Derby (Liverpool vs Everton)

One of the most fixture-dense rivalries in English football

Everton have repeatedly avoided defeat against Liverpool sides dominating the rest of the league

Kentucky Derby (horse racing)

Post-time favorite has won only 40 of the last 150-plus runnings

2022's Rich Strike went off at 80-1 and won anyway

College rivalry week (US college football)

Programs often meet with wildly different season records

Harvard Sports Analysis found rivalry-game margins are not statistically more volatile than any other week

That last row matters more than it looks. A researcher at the Harvard Sports Analysis Collective compared rivalry and non-rivalry college football games and found the standard deviation from the predicted scoring margin was nearly identical – 15 points for rivalries, 16.2 for everything else.

Why Derbies Resist Statistical Modeling:

• Local pride overrides squad-depth calculations that decide most ordinary fixtures.
• Managers rotate tactics specifically to neutralize a known rival, discarding their usual game plan.
• Fan intensity measurably affects player decision-making under pressure.
• Referees officiate differently amid heightened crowd noise and occasion.
• Set pieces and individual moments of composure outweigh expected-goals models.
• Former players facing old clubs often over- or under-perform relative to career norms.
• Betting markets price in emotion as much as form, creating volatile early odds.
• Short head-to-head sample sizes make historical trends statistically unreliable.
• Media narratives add pressure that standard models were never built to capture.
• Neutral fixtures rarely carry the same anytime-goal intensity a derby generates.

None of this means derbies are random. It means the variables driving them sit outside the columns a spreadsheet usually tracks.

The Numbers That Do Move: Cards, Goals, and Chaos

If match outcomes resist prediction less than fans assume, some other numbers genuinely do shift on derby day, and by a meaningful margin.

Disciplinary data is the clearest example. Cards shown in derbies run over 30% higher than the league average, and roughly 65% of major derbies go over 2.5 goals – both signs that intensity, if not destiny, really does change.

Metric

Derby Day vs. Average

Cards shown

Over 30% higher than a typical league fixture

Both teams to score

Hits in roughly 58% of major rivalry matches

Over 2.5 goals

True in about 65% of high-profile derbies

Betting market movement

Odds swing more sharply pre-kickoff due to fan-driven volume

Former-player performance

Around 60% of players significantly over- or under-perform against ex-clubs

Goal-margin variance (college football)

15-point standard deviation for rivalries vs. 16.2 for non-rivalries – essentially the same

Broadcast audience

Derbies consistently rank among a league's most-watched fixtures of the season

Managerial risk-taking

Coaches more frequently alter formation specifically to counter a known rival

Two of those rows tell almost opposite stories, and that's the point. Discipline and goal counts move; the identity of the winner mostly doesn't.

What This Means for Anyone Reading the Data:

• Cards and goals genuinely spike, so discipline and goalmouth markets track a real, measurable derby effect.
• Result unpredictability, by contrast, is largely a myth once sample size is controlled for.
• Betting markets often overprice emotion rather than any real shift in underlying quality.
• Managers plan specifically around a rival's known weaknesses, adding tactical noise rather than chaos.
• Former players carry psychological baggage that shows up in individual output, not team results.
• Media coverage inflates perceived unpredictability far more than the numbers support.
• Fans remember rare upsets vividly and forget the dozens of routine, form-following results.
• Comparing rivalry and non-rivalry fixtures is the only reliable way to isolate a genuine derby effect.

Card counts and goal totals are the honest part of the derby myth. The result itself is usually the least surprising thing about the day.

Why the Myth Persists Anyway

If the numbers are this clear, the anything can happen cliché should have died out by now. It hasn't, and the reason has more to do with memory than with math.

Rich Strike winning the 2022 Kentucky Derby at 80-1, or Donerail's 91-1 shock back in 1913, are the stories people actually retell. The hundreds of routine, favorite-wins-comfortably derbies in between rarely get mentioned at all.

Common Belief

What the Data Actually Shows

Form goes out the window in derbies

True for cards and goals, not for who wins (Harvard study)

Anything can happen

Derby margins are about as variable as any other week statistically

Underdogs always have a puncher's chance

True in famous isolated cases (Rich Strike, Donerail), not on average

Rivalry hatred decides matches

Influences discipline and effort, not necessarily the final scoreline

Bookmakers get derbies wrong

Betting volume and fan emotion move odds, not the true win probability

Managers can't set up trap games

Coaches frequently do plan targeted tactics – that's real signal, not myth

Every derby produces chaos

Only the highest-profile fixtures show consistently elevated card and goal stats

Home advantage disappears in derbies

Home teams still win derbies at rates close to their season-long average

Fans decide it more than tactics

Officiating and discipline shift, but tactical planning remains the deciding factor

The gap between belief and data isn't dishonesty – it's how human memory naturally weights a good story over a boring, accurate one.

Why the Myth Refuses to Die:

• Humans remember outliers far more vividly than routine, predictable results.
• A single famous upset can define the perception of an entire fixture type for decades.
• Media coverage rewards drama, amplifying rare chaos over the dozens of predictable derbies nobody wrote about.
• Betting markets commercially benefit from the public perception of unpredictability.
• Fans of the losing side need anything can happen to stay emotionally invested going forward.
• Real statistical noise, like cards and goals, gets conflated with the separate question of who wins.
• Small per-fixture sample sizes make it easy to cherry-pick proof in either direction.
• Football folklore rewards a good story more consistently than a boring, correct prediction ever will.

The myth survives because it's more fun to believe than the spreadsheet. That doesn't make the spreadsheet wrong.

Conclusion

Derbies really do behave differently – more cards, more goals, sharper odds swings – but the idea that results themselves turn unpredictable doesn't hold up once the sample size is large enough. What changes is intensity, not destiny. 

The romance of anything can happen survives because rare upsets are unforgettable and routine wins are not.