The 1996 F1 season is the closest we have measured in 30 years

Of the 30 seasons we have re-run one hundred times each, 1996 is the closest. Hill takes the title in 46 of every 100 and Schumacher in 44: two points apart, with the next tightest season five points back. The odd part is who the two are, because they were not team-mates and one of them had a far better car than the other.

The number, and why it is surprising

Every season is run a hundred times with its real grid, and out comes a spread of champions. In almost every year one name dominates that spread. In 1996 there are two, and they are all but level:

Look at the last column before the titles one. Hill and Schumacher finish on the same average position, P1.7, across sixteen hundred simulated races. And they do it from cars that are nothing alike: the Williams FW18 is the tenth most dominant car of the 37 we have measured, and the 1996 Ferrari was the first year of a rebuild that had not paid off yet.

The fight there was, and the one the engine sees

In reality the 1996 championship was a Williams internal affair. Hill against Villeneuve, the two blue cars, eight wins in sixteen races for the older man and the title settled in Japan. Schumacher finished third, a long way back, with three wins remembered more for how they happened than for what they were worth in the standings.

The model tells a different story. It says the real fight, the one latent in the machinery, was Hill against Schumacher. And that Villeneuve, in the same car as Hill and over the same calendar, comes out on 10 of every 100.

Both can be true at once. One describes what happened on sixteen Sundays; the other, what happened on average when you run those Sundays a hundred times.

Villeneuve is the perfect control

This is where 1996 becomes useful beyond the anecdote. Villeneuve had exactly the same car as Hill, the same team and the same races. Any difference between them is the driver and nothing else, which is the kind of clean comparison Formula 1 almost never allows.

And the difference is large: 46 against 10, with P1.7 against P2.6 on average. Villeneuve arrived as the reigning IndyCar champion and finished runner-up in his first European season, which is no small thing. But the model says the 1996 Williams won the title far more often with Hill in it than with him.

That contrast is what gives the tie above its meaning. The car does not explain everything, because Villeneuve had the best one and it was not enough. Nor does the driver, because Schumacher had the level and not the car.

Why it comes out as the closest of the 30

Because it is the year where the two forces cancel each other most precisely. Hill had the machinery advantage and Schumacher the driving one, and in 1996 those two advantages offset almost exactly: two points out of a hundred. In every other measured season, one of the two wins out more comfortably.

What this does not say

It does not say Hill did not deserve the title, or that Schumacher did. It says that if you re-run 1996 many times with the same machinery, the championships split almost evenly between the two. Hill won the one that was actually raced, which is the only one that counts.

And there is a limit worth keeping in mind, because it weighs heavily in 1996: the engine spreads mechanical failures according to each car’s reliability across the year rather than placing them in a specific race. Over a sixteen-race season that is a lot of slack. We go into it in the five seasons this simulator cannot explain, where 1996 misses the list narrowly.