Own dataHistory2026AntonelliMonza

Antonelli starts from the back at Monza, and a championship leader has lined up on the last row only nine times since 1950

F1 Legend Sim · Staff · 4 September 2026 · 7 min read

Antonelli came back from the 2026 break leading the championship and having never started worse than seventh. At Monza he starts nineteenth, on a planned engine change. We looked up every time a championship leader really did end up on the last row, which is 9 since 1950, and ran 3,600 races to see what is about to happen to him.

There is a question Formula 1 cannot answer on its own: what happens when the man leading the championship is stuck on the last row on a Saturday. It cannot answer it because it barely ever happens. We went through every race since 1950, all 77 full seasons, where a clear leader was already established, and there are 1,042 of them. In 993 the leader started inside the top ten.

Outside the top ten there are 49 cases, 4.7% of the total. On the last row, counting both slots, there are 9. Dead last, 8. In seventy six years of world championship racing.

The nine times

Those 9 are the entire sample the sport has to answer the question. Here they all are:

Year	Grand Prix	Leader	Starts	Finishes	What happened
1967	Monaco	Pedro Rodríguez	P16 of 16	P5	Four laps down
1980	Monaco	René Arnoux	P20 of 20	Retired	Accident
2003	Spain	Kimi Räikkönen	P20 of 20	Retired	Collision
2003	Canada	Kimi Räikkönen	P20 of 20	P6	In the points
2004	China	Michael Schumacher	P20 of 20	P12	One lap down
2012	Abu Dhabi	Sebastian Vettel	P24 of 24	P3	Won the title that year
2016	Belgium	Lewis Hamilton	P21 of 22	P3	Podium from the last row
2017	Brazil	Lewis Hamilton	P20 of 20	P4	One place short
2021	Russia	Max Verstappen	P20 of 20	P2	The best of the eight

7 of the 9 finished the race. The average finishing position among those who did was P5. 3 reached the podium and none of them won.

Vettel is the case worth keeping in mind, because it is the only one of the nine where the title was genuinely on the line. He arrived at Abu Dhabi 2012 leading, lost his grid slot in qualifying over a fuel problem and started from the back of a twenty four car field. He finished third, and weeks later he was champion by three points over Alonso. Starting last did not cost him the title, it cost him a handful of points.

Why he starts from the back: a new engine

Antonelli will not be back there for being slow or for qualifying badly, and the same goes for the nine in the table. He starts from the back because Mercedes is fitting a new engine, and that carries a grid penalty. It is the first of the four routes to the last row, and all four have nothing to do with the pace of the car.

His is the most common one today: taking power unit components beyond the allocation the rules allow. It does not come as a surprise, it is a team decision, picking the weekend where it hurts least and accepting a start from the back in exchange for fresh machinery for the rest of the year. The other three are a pit lane start after working on the car in parc fermé, a failure or a crash in qualifying, and exclusion over a technical irregularity. Vettel at Abu Dhabi 2012 is one of those last ones: he had qualified third.

And it is not rare. Comparing qualifying position against the actual grid across 2,491 starts from 2021 to 2026, 59 drivers lined up five places or more behind where they had qualified, and 21 of them ten or more. In 2026 it has already happened 2 times. And that is a floor rather than a count: a driver who already knows he is starting from the back will not spend a set of tyres qualifying well, so his drop never shows up.

That it lands on him is not a calendar accident either: 2026 is the first year of a new engine formula, the season when teams most often fit fresh units and swallow the penalty.

What 3,600 races say

Antonelli leads 2026 on 242 points, 59 clear of Russell, his own team mate. In 12 races he has never started worse than P7, and he has started first in 6 of them. He shares the best car on the grid, which is half the explanation for where he is.

Sunday only runs once, so we ran it ourselves. We take the 2026 grid, freeze the starting order, move Antonelli to a different slot without touching the other twenty one, and simulate 3,600 races per condition across 12 circuits. The only thing that changes between one run and the next is where he starts.

Finishes	Starting P1	Starting P22
P1	23.8%	6.1%
P2	20%	6.3%
P3	13.8%	9.1%
P4	11.4%	11.2%
P5	12.6%	16.5%
P6	8.8%	17.7%
P7	6.8%	23.3%
P8	0.4%	6.1%
P9	0%	0.3%
Retired	2.4%	3.4%

3,600 races per condition. Tenth place and below do not reach 0.1% in either column, which is why they are not in the table.

The distribution shifts as a whole, but it does not fall apart. Starting first he wins almost one race in four and reaches the podium 57.5% of the time. Starting last, the wins drop to 6.1% and the podium to 21.5%. What barely moves is scoring at all: 97.6% against 96.6%.

In points per race it goes from 15.37 to 10.08. A thoroughly bad Saturday costs him a little over five points on average, with a 59 point cushion in hand.

Both counts say the same thing

This is where the two halves of the article meet, and nobody went looking for it. The eight times a leader really did start on the last row, the ones who finished came home in P5 on average. The 3,600 simulated races of Antonelli from dead last give an average finish of P5.1.

There is no tuning behind that. They are two independent counts: one comes from sixty years of official results, the other from a simulation engine that was never shown the historical figure. Landing a tenth of a place apart is the best sign that the question has a stable answer, and that the answer is boring in the good way.

One point looks like a disagreement and is not. In the record books the best of the eight is Verstappen's second place in Russia: no leader has ever won from the last row. The simulator gives Antonelli a 6.1% chance of winning from there, roughly one race in sixteen. With only eight real cases, zero is exactly what you would expect to see.

The circuit matters more than the grid slot, and he gets the best one

What does change the answer a great deal is where the race is held. These are the odds of finishing on the podium from last place, circuit by circuit:

Circuit	Podium from last	Points per race
Monza	39.3%	12.51
Spa	33.3%	11.8
Silverstone	30.3%	11.43
Baku	27.7%	11.16
Interlagos	27%	11
Suzuka	23.7%	10.73
Barcelona	18%	9.78
Imola	14%	8.91
Zandvoort	13.3%	8.88
Hungaroring	12.7%	8.6
Marina Bay	10.3%	8.21
Monaco	8.3%	7.99

From Monza to Monaco there is a factor of almost five. The same grid, the same driver and the same starting slot produce a podium four times out of ten at Monza and fewer than one in ten at Monaco. What decides the cost of starting from the back is not how badly you qualified: it is which weekend it happens on.

And that is what makes this case something other than bad luck. Of the twelve circuits we measured, Monza is the best of all of them to come back through. A team choosing when to spend a penalty would pick exactly this, and Mercedes has picked exactly this.

Across the 300 races we simulated with Antonelli starting last at Monza, he finishes in the top five 68% of the time, reaches the podium 39.3% and wins 12.3%. The average finish is P4.2 and he takes 12.51 points per race, more than at any other track in the sweep.

His own maths came out close. Asked what was realistic, he said: "In an ideal world, I think realistically top five you can do well, but it is all driven by also your pace". The simulator gives him that top five two times in three, and on average leaves him one place ahead of where he puts himself. For once you can hold a driver's prediction against a measurement, and they agree.

If you want to try it yourself, in What If? you can force the circuit and the starting grid and see how long a slow car holds off a fast one.

Methodology

⚠️ Corrected with the final grid. This piece was written the day before qualifying, when Antonelli was expected to start last. He ended up startingnineteenth, and getting there took two corrections: three drivers took engine penalties and the FIA orders them by when each penalty was applied, which left Albon and Lawson behind him, and then Alonso picked up a penalty for irregularities in his power unit that sent him to the back and moved everyone behind up one. The simulations below are run from the LAST box, three places behind where he actually started, so if they miss what happened to him, they miss on the low side.

The historical half covers the 77 seasons from 1950 to 2026, none of them discarded, using Jolpica data (the successor to Ergast). For each Grand Prix we take whoever topped the official standings after the previous round and the slot he started from that day.

Why the official table and not a running total. The first version of this measurement added up points race by race, which looked like the careful option and hid four separate problems. Until 1990 the championship ran on dropped scores, meaning only each driver's best results counted, so the sum does not match the official table. In 1997 Schumacher was excluded from the championship, so he is in the race results and not in the standings. In 2007 Hamilton and Alonso both finished on 109 points and the order comes from the tie break. And since 2021 sprint race points sit in a separate place. The official table for each round settles all four at once, because they are already settled there.

Anyone who did not take the start is excluded. In the data, a starting grid slot of zero almost never means a pit lane start: it means the driver did not start. Without that filter the first row of the table was Jochen Rindt at Monza 1970, and what happened that weekend is that he died in practice.

The grid drops come from comparing each driver's qualifying position against his starting slot across 2,491 starts between 2021 and 2026. It undercounts penalties: a driver who already knows he is at the back and does not chase a lap time qualifies last, and the drop never registers.

The simulated half uses our database's 2026 grid, with the starting order frozen and 300 races per circuit and condition, a different seed each time. All twenty two drivers run with their real attributes, something our engine has only done since 3 September 2026, so these figures are not comparable with those in earlier articles.

Two limits of the model. The first is reliability: the engine produces 3.4% retirements while the modern real sample (22 cases between 2003 and 2026) shows 14%, so the simulator understates disaster. The second is how the field spreads out: in the model a front running car essentially never finishes below eighth, which is why "outside the points" comes out at zero across 3,600 races. Reality has more ways to ruin your Sunday than this engine knows how to simulate.

← More articles