Eleven legends, one car, 8,250 seasons: the greatest-of-all-time league
F1 Legend Sim · Staff · 22 August 2026 · 9 min read
It is the argument that never dies, and it is usually settled with memories. We settled it with an experiment instead: eleven candidates, the exact same car in every duel, everyone against everyone, 150 championships per duel. That is 55 duels and roughly 8,250 full seasons. The result has a predictable part, a surprise, and an awkward confession about what a simulator can and cannot know.
Why comparing eras is impossible — and what you can do instead
The Senna of 1988, the Schumacher of 2004 and the Verstappen of 2023 never shared a track at their peak. Different cars, different rules, different tyres, different rivals. Any direct comparison has to neutralise all of that, and that is exactly where opinion usually replaces evidence.
A simulator does not solve the problem, but it does narrow it. Give all eleven the same car — the same one, not a similar one — and make them run the same championship against the same grid hundreds of times, and the machinery stops being an excuse. The only thing that differs between one side of the garage and the other is the driver.
The league: eleven drivers, one shared car per duel
Every pairing ran 150 championships of 18 races as team-mates, both in the best car of a real season and against that year’s grid. The right-hand column is the share of championships each driver finished ahead of his rival, added up across all ten opponents.
Each duel’s season is picked so that neither of the two duellists raced in it — if a name is in the grid, the engine runs him twice. Seven seasons came up: 1990, 2002, 2006, 2007, 2014, 2020, 2026. The reasoning is in the methodology at the end.
| # | Driver | Starting rating | Duels won | In the rain |
|---|---|---|---|---|
| 1 | Max Verstappen 2023 | 98 | 83.5% | 87.9% |
| 2 | Michael Schumacher 2004 | 97 | 80.7% | 81.3% |
| 3 | Ayrton Senna 1993 | 96 | 71.3% | 68.9% |
| 4 | Alain Prost 1990 | 94 (+1) | 70.1% | 70.3% |
| 5 | Lewis Hamilton 2020 | 96 (-1) | 69.0% | 70.4% |
| 6 | Fernando Alonso 2005 | 94 | 56.8% | 57.9% |
| 7 | Sebastian Vettel 2013 | 92 | 39.9% | 38.1% |
| 8 | Juan Manuel Fangio 1957 | 90 | 34.5% | 38.6% |
| 9 | Jim Clark 1965 | 88 | 23.5% | 19.5% |
| 10 | Niki Lauda 1977 | 86 (+1) | 11.6% | 12.5% |
| 11 | Jackie Stewart 1971 | 86 (-1) | 9.1% | 4.5% |
The «starting rating» is the overall each driver carries into the simulator; in brackets, the places the league gains or loses against that order.
The confession: most of this was already in the starting rating
Let us begin with what an honest article has to say first. The order that comes out of 8,250 seasons is almost exactly the order of the rating each driver walks in with. There are three swaps, all of a single place, and all three — as we are about to see — fall inside the margin of error.
That is not a flaw in the experiment; it is its ceiling. A simulator does not know who was better. It knows what we told it about each of them. If you were hoping the machine would dethrone somebody, the machine cannot do that, and anyone claiming otherwise is selling you something. What it can do — and this turns out to be more useful than it sounds — is tell you how much each rating point is actually worth in results, and where the differences stop existing.
The real finding: there are tiers, not places
Here is what the 8,250 seasons do add. At this number of repetitions, two results less than about 3 percentage points apart are indistinguishable — the gap fits inside the noise. Measured with that ruler, the table stops being a list of eleven places and becomes a set of tiers.
At the top, Verstappen and Schumacher are 2.7 points apart against a ±2.8 threshold: they are tied. Head to head, Verstappen finishes ahead in 54% of championships — a coin toss with a slight lean.
The next tier does the same thing between Senna, Prost and Hamilton: 1.3 and 1.1 points between them, against a ±3.4 threshold. Another dead heat, this time three-way.
What the model does separate is the tiers from each other: Schumacher to Senna is 9.4 points, Hamilton to Alonso 12.2, Alonso to Vettel 16.9. All far beyond the margin.
Translated: inside this model, «who is number one?» has no answer, and neither does «who is third?». What does have an answer is which group each driver belongs to. That is more useful than a ranking, because it tells you which arguments are worth having and which are arguments about nothing.
The surprise: rain does not rescue Senna. And we know why
Senna’s wet-weather legend is one of the few things almost everyone agrees on: Monaco 1984 charging through the field in a Toleman, Estoril 1985 winning by over a minute, Donington 1993 passing four cars on the opening lap. So we ran the whole league again in the wet, expecting to watch him climb.
He fell. From 71.3% to 68.9%, third place to fifth. Against Hamilton the result flips outright: Senna takes 53% of championships in the dry and only 42% in the rain.
Before publishing that as «the simulator debunks the myth», we went to find out why. The answer does not flatter the simulator: there is no wet-weather attribute. Our engine has no field that says «this driver is extraordinary in the wet». What it does when it rains is reweight the formula — the car counts for less, the driver and his reflexes count for more — and use the reactions attribute as a stand-in. And there Senna carries a 94, the lowest of the top five: Verstappen 98, Schumacher 97, Prost 96, Hamilton 95.
So the model is not debunking anything. It is exposing a hole of its own: Monaco 84, Estoril 85 and Donington 93 have nowhere to live in our data. A driver whose greatest virtue was precisely that walks into the simulator without it. We are logging that as what it is — a gap in our data, not a verdict on Senna — and it may well be the most valuable thing the 8,250 seasons produced.
The one who does gain in the wet is Juan Manuel Fangio: against Vettel he goes from 42% of championships in the dry to 51% in the wet — from losing the duel to winning it. Same mechanism, opposite direction: rain shifts weight away from the car and towards the driver, and his is a hands-first profile rather than a machinery-first one.
What these numbers do not say
Three warnings, and we mean them:
One. The older drivers finish at the bottom — Clark, Lauda and Stewart between 10% and 24% — and that does not mean they were worse. It means their starting ratings are lower. Rating a 1965 driver and a 2023 driver on one scale is a debatable editorial call, and it is ours, not the engine’s.
Two. The experiment measures repeated race pace. It does not measure developing a car, handling title pressure, or building a team around yourself — several of the reasons Schumacher and Hamilton are who they are.
Three. Qualifying barely counts. In our formula race pace weighs far more than a Saturday lap, so a driver whose greatest weapon is one perfect lap — Senna again, with a 99 in qualifying, the highest of the eleven — is competing on the ground that suits him least.
So who is the greatest?
Inside this model and with these ratings: Verstappen and Schumacher tied at the top, and a three-way tie behind them with Senna, Prost and Hamilton. Outside this model the question remains open, and that is probably for the best. What we gained is not a verdict: it is knowing that the argument is worth having between tiers and never inside one — and finding, along the way, one thing our simulator cannot do.
If you want to build your own equal-machinery duels, What If mode lets you pick both drivers, the car and the season. And the driver profiles carry the attribute breakdown all of this comes from.