Long read

The GreatestOfAllTime, in four charts.

Start with the fairest comparison the sport gives you: teammates in the same car. Fit one cross-era rating on every comparable head-to-head, then stress-test the result from several angles before calling anyone the greatest.

Data: teammate pairings from classified race finishes, 1950 to present. Indy 500 rounds excluded.

1 · The scoreboard

One rating per driver, fit on every comparable teammate pairing from 1950 onward in a single model. Same team, same car, same weekend: about as close as F1 gets to a fair head-to-head. A rating gap of 1.0 is roughly a 73 % win probability; 2.0 is about 88 %. Bars show the 95 % Wald interval.

This is the leaderboard for the twelve drivers who actually show up in serious GOAT conversations. Everyone else is a click away — use the "explore more" toggle underneath.

Red dots are the GOAT contender pool. Rank labels are against the full field of ~390 drivers.

Fangio and Ascari sit alone at the top on raw rating; no one else gets close on this axis.

Verstappen anchors the next cluster and has the narrowest error bar of any driver on the board - more teammate pairings than most of the tier above him, despite having raced fewer seasons.

Schumacher, Hamilton, Vettel and Prost all land lower than their championships might suggest: the one-shot career rating rewards overwhelming same-car dominance, and long careers spent consistently a step ahead of strong teammates (rather than monstrously ahead of weaker ones) end up in the middle of the pack. §4 splits that trade-off out explicitly.

2 · When were they actually great?

A career-wide rating is useful for a leaderboard and bad for a life story. The five-season rolling fit gives the shape back: rise, peak, plateau, decline. Each line is the same estimator re-run on a sliding window (year ± 2 seasons).

Hollow circles = no teammate losses inside the window, so treat them as lower bounds, not a best guess.

Fangio's peak sits so early and so high that the rolling fit barely revisits it for more than a decade.

Clark and Senna never get the decline chapter, which is part of why their cases stay open-ended.

Schumacher and Hamilton are different: not just high peaks, but sustained ones.

Alonso and Vettel show how quickly a late-career rating can become a story about team context as much as driver level.

3 · Head-to-head in equal cars

A bit of fun - pick any two. The model turns the rating gap into a simple question: if these drivers had to share a car, how often would one beat the other? The uncertainty band combines the two Wald intervals as a first-order approximation, so for closely matched drivers the overlap matters as much as the point estimate.

Dumbbell of the two drivers' ratings with 95 % Wald intervals. Wider bars = less precisely estimated. If the bars overlap, the difference is not significantly different from zero at conventional levels — the "A beats B" number above is still the best point estimate, but take it with salt.

4 · Peak vs longevity

A career rating is an average. A career is a peak plus how long you hold it. Plot one against the other and the GOAT argument splits into two clusters: the sprinters (Fangio, Clark) who hit insane peaks over short careers, and the long-haulers (Schumacher, Hamilton, Alonso) who couldn't quite match the peak but held a very high level for fifteen years and change.

Grey dots: every other driver with at least three seasons in the fit. Red dots: the twelve contenders. Peak is the highest rolling-window rating in their career (the highest point from the chart in §2). Seasons are the count of valid windowed seasons, not the calendar span.

There's an obvious upper envelope where nobody lives - the top-right corner (long career and absurd peak) is empty because peak-holding for that long is just very hard. Schumacher and Alonso come the closest, but the peak-is-everything argument still begins with Fangio's y-coordinate.

The verdict

By peak

Juan Manuel Fangio 2.81 peak, untouched in 68 years

Peak × longevity

Fernando Alonso 26 seasons at elite altitude

Unfinished business

Max Verstappen Narrow SE, still climbing

The model does not collapse the debate into one uncontested name. It does narrow the field and make the tradeoff explicit:

Nobody beats Fangio on raw peak rating. 2.81 in 1956, shadowed only by Ascari's 2.77 in 1953 and, in the modern era, approached only by Verstappen's 2.48 in 2025. The model is built to reward overwhelming same-car dominance, and the fifties had enough of it to keep Fangio on top of a seventy-five-year table.
Longevity-at-altitude is Alonso, not one of the obvious picks. 26 seasons in the windowed fit, a peak of 2.20, and a consistently high level across four different manufacturers. Schumacher is the close runner-up (25 seasons, peak 2.12); Hamilton has the longest calendar career on this list but the lowest rolling-window peak (1.10) — a result reputation does not prepare you for, and one the model is willing to state.
Clark and Senna are almost certainly underrated here. Both died at peak, with the rolling fit still pointing up. The leaderboard number is what the model can justify from the data it has; the level they'd have reached with a full career is a lower bound on that, not a point estimate.
Verstappen is the active driver most likely to rewrite this chart. His standard error is already narrower than every driver above him in the table — 166 pairings against some of the strongest teammates of the era. Five more seasons of beating them will either vault him into the Fangio/Ascari neighbourhood or reveal where he actually plateaus. The model's answer for now: elite, but unfinished.

The follow-up question is next: who benefited most from circumstance, and who spent a career fighting the other way.