Original dataDuelsTeammatesFormula 1

Twelve teammate duels, in exactly the same car

6 August 2026 · 9 min read

Almost nothing in Formula 1 can be compared cleanly. The cars are different, the strategies are different, the eras are different. There is exactly one exception, and it is the one the paddock has been using to settle arguments for decades: the teammate. Same car, same garage, same engineers. We took twelve famous pairings and ran five seasons each, with the machinery exactly equalised.

Why the teammate is the only honest judge

When one driver wins twenty races and another wins none, the first question is what car each of them had. We measured that in two thousand simulated careers: of the drivers who never sat in a top-three car, not one — literally not one — won a championship. The car contaminates any comparison across teams.

Inside the same team, that contamination disappears. Which is why the intra-team record is the statistic team principals look at most and headlines mention least. So we ran the experiment the other way round to how it is usually done: instead of counting titles, we put each pair in two identical cars and gave them five seasons of sixteen races.

The twelve duels

Each driver’s attributes come from our database, which covers the full grid since 1988 and is calibrated against the real results of every season. Sorted from closest to most lopsided:

DuelYearSeasons wonPointsWins
Senna vs Prost1988 · McLaren2 of 5330–3407.6 – 8.4
Leclerc vs Vettel2019 · Ferrari2 of 5343–3308.4 – 7.6
Alonso vs Hamilton2007 · McLaren3 of 5345–3289.2 – 6.8
Hill vs Villeneuve1996 · Williams4 of 5350–3139.4 – 6.6
Norris vs Piastri2023 · McLaren4 of 5342–3129.8 – 6.0
Norris vs Ricciardo2021 · McLaren5 of 5359–31910.6 – 5.4
Hamilton vs Rosberg2014 · Mercedes5 of 5368–30213.0 – 3.0
Hamilton vs Rosberg2016 · Mercedes5 of 5366–30413.2 – 2.8
Verstappen vs Ricciardo2018 · Red Bull5 of 5373–30013.2 – 2.8
Alonso vs Massa2010 · Ferrari5 of 5379–29114.6 – 1.4
Schumacher vs J. Verstappen1994 · Benetton5 of 5384–28315.2 – 0.8
Alonso vs Fisichella2005 · Renault5 of 5391–28615.2 – 0.8

The biggest surprise: Senna and Prost are still level

The most famous duel in history is also the closest of the twelve, and it falls by a hair on the side most people do not expect: Prost takes three of the five seasons, with 340 points to 330 and 8.4 wins to 7.6. In any other context we would call that a draw — change one race and you change the result.

And that is exactly the interesting part. The legend says Senna was from another planet and Prost was the calculator who lived off other people’s mistakes. Strip out the noise — the cars, the regulations, Suzuka — and what is left is two drivers separated by almost nothing. We pull that thread further in our cross-generation comparison.

The three duels decided by a single point

Besides Senna-Prost, two more end up essentially level: Leclerc and Vettel in 2019 (343-330, with Vettel taking three of the five seasons) and Alonso and Hamilton in 2007 (345-328). Neither of those being close is a coincidence: both are duels that in real life ended in open warfare — one broke a team apart, the other took a championship down with it.

What the numbers say is that neither driver was right. They were so close that the difference came down to the day rather than the talent — and that is precisely why the conflict was inevitable.

And the ones that were not duels but hidings

At the other end there are four pairings where the result never changes: Alonso against Fisichella (15.2 wins to 0.8), Schumacher against Jos Verstappen (15.2 to 0.8), Alonso against Massa (14.6 to 1.4) and Hamilton against Rosberg (13 to 3). Five seasons out of five, no exceptions.

Rosberg deserves a footnote, because his is the one that ages worst in the collective memory: he was world champion in 2016 and here he loses all five seasons. That is not a contradiction, it is the whole argument of this article. A championship is won in one specific year, with its retirements and its weekends; an intra-team duel repeated five times measures something else.

What this does not prove

To be clear: this is not a verdict on who was the better driver. It is what our ratings say once you take the car away. Those ratings are calibrated against real results, so they are not made up — but they are not revealed truth either: a driver whose team treated him badly for two years carries that label in the data.

Nor does it simulate what really decides an intra-team duel in real life: who gets the new part when there is only one, who picks the strategy first, who has the veteran engineer. None of that is here. What is here is the clean question — same car, sixteen races, five times — and its answer.

If you want to run your own duels with any pair from the 863 drivers in the database, it takes two clicks in the what-if mode. And if you would rather live it from the inside, in the driver career mode your teammate is exactly that: the only rival racing your own car.