In the eighties four cars in ten finished. Today more than eight do
F1 Legend Sim · Staff · 5 September 2026 · 6 min read
In the nineteen-eighties, four out of every ten cars that took the start actually finished. Today more than eight do. It is the largest change Formula 1 has lived through and it almost never gets told, because you cannot see it: it happens in the workshop, not on track. We counted all 1,161 race results since 1950.
Finishing used to be the exception
| Decade | Cars that finish |
|---|---|
| 1950s | 50% |
| 1960s | 47% |
| 1970s | 51% |
| 1980s | 44% |
| 1990s | 50% |
| 2000s | 68% |
| 2010s | 80% |
| 2020s | 85% |
Four consecutive decades below 52%, and then a climb that has not stopped. The worst year of all was 1984, with 36% of cars finishing. The best is 2024, at 89%.
Put another way: a driver in the eighties went out knowing that not finishing was the likelier outcome. One today goes out knowing he almost certainly will. They are two different sports with the same name.
What changed, and when
The jump is not gradual, it is concentrated. Between the nineties and the two thousands the figure moves from 50% to 68%, and keeps climbing from there. It lines up with simulation arriving in component design, with limits on how many engines a team may use in a season, which forces them to be built to last, and with safety rules that retire a damaged car before it fails outright.
And what they broke
The source records why each car retired, and read decade by decade it tells the whole story without needing any explanation:
| Decade | Most common cause | Second | Retirements |
|---|---|---|---|
| 1950s | Engine (24%) | Accident (13%) | 973 |
| 1960s | Engine (21%) | Withdrew (11%) | 1080 |
| 1970s | Engine (21%) | Accident (19%) | 1695 |
| 1980s | Engine (21%) | Spun off (9%) | 2223 |
| 1990s | Engine (22%) | Spun off (20%) | 1941 |
| 2000s | Engine (17%) | Collision (15%) | 1169 |
| 2010s | Collision (21%) | Accident (12%) | 842 |
| 2020s | Retired (48%) | Collision (10%) | 427 |
The engine was the leading cause of retirement for six straight decades, always around 20%. Then in the 2010s collisions overtake it. That change of places is the whole article in one line: cars stopped breaking on their own, so what stops them now is each other.
The race almost nobody finished
| Grand Prix | Started | Finished |
|---|---|---|
| Monaco Grand Prix, 1996 | 22 | 3 |
| Detroit Grand Prix, 1984 | 26 | 5 |
| San Marino Grand Prix, 1985 | 26 | 5 |
| South African Grand Prix, 1993 | 26 | 5 |
| British Grand Prix, 1975 | 26 | 6 |
Monaco 1996 is the outright record: twenty-two cars started and three finished. Olivier Panis won it in a Ligier, the only victory of his career and the last of his team's. That year's championship was not decided by much either: run a hundred times, it is the closest season of them all, with the next tightest five points back.
And the figure that best sums up the first seventy years: of the 1,161 races held, in only 18 did every car that started also finish. The first time it happened was in 2005: fifty-five years of Formula 1 without a single race in which the whole grid made it home.
And this explains something else
Reliability is the cause that best explains a figure people do discuss: that fewer and fewer different drivers win. Eleven won across sixteen races in 1982; three won across twenty-two in 2023.
It is not that there was more parity back then. It is that the favourite kept breaking, and when the favourite breaks the win falls to hands that would not otherwise have held it. With today's grid finishing at 85%, that door is shut.
The other half of this story is in how many different drivers win each year. And there is a third face to it, the one belonging to the driver coming from the back: when half the grid broke, surviving was worth ten places without passing anybody. That is why the big comeback is dying out, from 95% of races featuring one in the eighties to 34% this decade.
A warning about our own simulator
This is worth saying because it affects what we do here: in our engine, a car's reliability barely changes the result. It carries a weight of 0.04 in the retirement formula, so the difference between a very reliable car and a mediocre one is almost imperceptible. We measured it across six seeds.
It is a known limitation and we have not fixed it yet, because touching that constant moves the balance of every game mode. But it means that if you simulate a season from the eighties, the engine will give you a tidier race than the real one was. On the season pages you can see both side by side: what actually happened, and what comes out when you replay it a hundred times.