Baku and the Five-Name Problem: What the Data Says About Winning in Azerbaijan

Charles Leclerc has taken pole position in Baku four years running — 2026, 2026, 2026 and 2026 — and has never once stood on the top step of the podium there. Across the entire history of the Azerbaijan Grand Prix, only three of eight races have been won from pole: Nico Rosberg in 2026, Lewis Hamilton in 2026 and Valtteri Bottas in 2026. The pole-to-win conversion rate at Baku is 37.5 percent, materially below the average for most street circuits on the calendar.

That is the first figure I check whenever someone asks me who will win in Baku. It rarely answers the question, but it always says one thing: this circuit does not reward the fastest man over a single lap.

In the week before the race, Formula 1's own website published a list titled "5 drivers most likely to win in Azerbaijan." Five names, one headline, and not a single line explaining the selection criteria. I spent three days cross-referencing it against circuit data and historical results. What matters lies in the gap behind that list.

CIRCUIT CONTEXT

Baku City Circuit runs 6.003 km over 20 corners, and almost a third of that distance is the 2.2 km main straight linking Turn 20 back to Turn 1. No other round on the calendar demands a lower-downforce configuration except Monza. The rear wing is trimmed, downforce drops, straight-line speed rises — and that is the entire reward.

The cost sits in the other 19 corners. Once downforce is cut, the tyres slide more through the slow corners, surface temperature drifts above the optimal operating window, and graining appears earlier than usual. At Baku the main risk is not the rate of tyre wear. It is that the rubber is cooled along the long straight and then shocked back to temperature at Turns 1 and 2 while the car is still braking from over 300 km/h.

The pit lane is entered off Turn 20 and runs long, making pit loss one of the highest of the season — roughly 20 seconds per stop. A two-stop strategy therefore only makes sense if a neutralisation intervenes, or if degradation becomes so severe that no other option remains.

And neutralisations at Baku are close to a default. The Turn 1 to Turn 3 complex is the first choke point, where twenty cars brake from top speed into a square corner. Turn 8 — the castle section, where the road is only about 7.6 metres wide — has no run-off at all. Any error there brings a yellow flag, a safety car or a red flag.

Data is never in a hurry, but people always are. I have followed every Azerbaijan Grand Prix since 2026, and on the records I keep, no edition has finished without at least one neutralisation directly shaping the result. That is why any preview of this round should be read with a technical warning attached: probability at Baku carries wider error bars than at any other street circuit on the calendar.

DATA ANALYSIS

So on what basis were F1's five names chosen? The article does not answer that question. The fact that it does not is itself worth analysing.

There are three plausible methods for building such a list. First, current car performance — which teams have the best package in low-downforce trim. Second, personal history at Baku. Third, media reach — which drivers are searched for most. When the content sits behind the rights-holder's registration wall, the third method almost always wins.

That is why I did not start from the list. I started from the historical table.

Pole versus winner at the Azerbaijan Grand Prix: 2026, pole Nico Rosberg, winner Nico Rosberg. 2026, pole Lewis Hamilton, winner Daniel Ricciardo, from 10th on the grid. 2026, pole Lewis Hamilton, winner Lewis Hamilton. 2026, pole Valtteri Bottas, winner Valtteri Bottas. 2026, pole Charles Leclerc, winner Sergio Pérez. 2026, pole Charles Leclerc, winner Max Verstappen. 2026, pole Charles Leclerc, winner Sergio Pérez. 2026, pole Charles Leclerc, winner Oscar Piastri.

That table carries three signals. Signal one: grid position has no decisive value. Ricciardo won in 2026 from 10th, after a qualifying crash forced a gearbox change and a grid penalty. Signal two: the wins cluster around two very different driver types — those who manage tyres well in chaos, and those who can accelerate instantly onto the straight. Signal three: no driver has ever won Baku in consecutive years.

What Baku actually rewards is corner exit. Of the 20 corners, four decide lap time: Turn 1, Turn 2, Turn 15 and Turn 16. All four are slow, and all four are joined by short bursts of acceleration. At those points, low-rpm traction control matters more than peak power. A car that is strong on the straight but unsettled on exit will lose more time through those four corners than it gains back over 2.2 km.

That is why late-braking, tidy-turning, early-on-throttle drivers tend to go well at Baku. Sergio Pérez is the clearest case: two wins and a further podium in 2026 with Force India, in a period when he never had the fastest car. Pierre Gasly finished third in 2026 with AlphaTauri. Lance Stroll finished third in 2026 with Williams at 18 years old. Sebastian Vettel finished second in 2026 with Aston Martin.

That list of names does not overlap with the championship standings. It overlaps on one shared trait: keeping tyres inside a stable temperature window while repeatedly switching between hard braking and hard acceleration.

Qualifying data at Baku says something else again. Because the main straight runs close to 2.2 km, the tow effect in qualifying is larger than anywhere else. A car running behind a rival from Turn 16 to Turn 20 can recover two to three tenths purely from the aerodynamic benefit of the slipstream. The starting order at Baku therefore reflects track timing more than pure pace.

This partly explains Leclerc's four consecutive poles. His engineering group puts the car on track at exactly the right moment, but that is an advantage worth one lap, not 51. In the race the car has to run in dirty air, the tyres have to absorb continuous thermal cycling, and the strategy has to survive at least one neutralisation.

Now, back to the five names. If the criterion was low-downforce car performance, the list would contain only the teams with the best straight-line efficiency. If the criterion was personal history, Leclerc, Verstappen, Piastri and Pérez would all be on it. If the criterion was media reach, the list would look identical to every other list of the season, with only the round name changed in the headline.

I cross-checked three independent data sources — official F1 statistics, lap-time databases and the tyre supplier's thermal records — and could not find a single index capable of reproducing that list. The list is unverifiable. And a prediction that cannot be verified cannot really be called a prediction.

One more thing stands out: the regulation cycle. The current season sits at the end of the cycle before the 2026 reset, when new power units, active aerodynamics and sustainable fuels will change almost the entire performance baseline. In such a phase, the gaps between teams compress, because everyone has reached the ceiling of the current concept. A high-variance circuit like Baku flattens the order once more. Every technical cycle copies the data of the cycle before it, and nobody learns.

THE CONTRARIAN ANGLE

There is a correlation many readers mistake for causation: drivers who appear on pre-race prediction lists tend to win a lot of races. That is true, but not because the list has predictive power. It is true because the list is built from drivers who were already winning. The cause sits in the car, and the list is only a consequence.

At Baku that correlation is weaker than anywhere else. A simple calculation: if a five-driver list were right in six of the last eight races, it would hit 75 percent. But look at the table above — only three of eight races were won from pole. Three of eight. Any five-name list that accurately described Baku's reality would have to include at least two drivers starting outside the top three.

This type of content lacks variables more than it lacks names. At Baku the decisive variable is when the safety car appears and who still has a spare set of tyres to exploit it. A driver running seventh when the yellow flags come out on lap 30 has a better shot at winning than the leader who has already burned through his softs. No five-name list can encode that, because it depends on an event that has not happened yet.

One point I want to make clearly: I am not against making predictions. I am against making predictions without stating the criteria, because then the reader has no way of judging where the prediction went wrong. A good prediction must be capable of being proven false. Otherwise it is just a well-dressed sentence.

At 60, I no longer believe in luck, only in numbers that have not spoken yet. And the numbers at Baku say the error bars around any prediction are wider than a five-name list can express.

SIGNALS FOR THE NEXT ROUND

Second practice will answer more than any list. Watch the average pace across the long runs on the Medium compound: if the gap between the fastest car and the fifth-fastest is under four tenths per lap, the race will be decided by strategy rather than outright speed. Watch where cars queue at Turn 20 on the final qualifying run, because at Baku the man who finishes ahead is usually the one who starts on the inside. And watch which spare tyre sets each team keeps back after qualifying.

Baku always answers in a different way from how the question was asked. That is its entire character, and it is also why it is worth watching more than any list.