Baku Does Not Listen to Prophecies: Why the Azerbaijan Race Always Breaks Every Prediction

Night in Baku. The stone walls of the Old City of Icherisheher run right alongside the track, so close that at Turn 8 — considered the hardest corner in racing — drivers have only a few dozen centimetres of margin between the wall and the guardrail. A few hundred metres later comes a two-kilometre straight along the Caspian shore, where race cars reach over 350 km/h before braking hard into Turn 1. In the eight editions of the Azerbaijan Grand Prix, six different drivers have stood on the top step. Six out of eight. That is one of the highest winner-diversity ratios of any race in modern Formula 1 history.

That ratio does not appear in the headlines spreading across platforms this week. There, an algorithm — often given the flattering name of supercomputer — has said that the race was decided before the first lap began. I read those lines and forced myself to stop. Data does not lie, but the people who clean the data do. The question is not whether the algorithm is right or wrong. The question is: who chose which numbers to put in the headline, who removed which numbers, and who benefits when readers believe a street race can be solved like a linear equation.

In nearly forty years of working with sports data, I have learned one thing: every number is a witness, and every witness needs to be interrogated. Baku's data, when properly interrogated, tells a completely different story from what the prediction tables are selling.

Context: A season compressed into a few final rounds

The championship picture this season, looking at the standings, is almost settled at the top. Kimi Antonelli of Mercedes leads by 81 points over his teammate George Russell — equivalent to nearly three full victories. The young Italian driver stands at the threshold of the first world championship of his career, with eight wins already this season.

Behind him, the real race — the race for second place — is where the most drama resides. Russell sits second but has only two wins and seven podium finishes. Charles Leclerc of Ferrari is fifth, 44 points behind Russell, with one win and four podiums. Lewis Hamilton, Leclerc's teammate, ended his winless streak this year but has just retired in Madrid due to a brake failure. Lando Norris of McLaren has two wins and five podiums. Max Verstappen of Red Bull is regaining form with five podiums in the last seven rounds, including third in Italy and second in Spain.

That is the context anyone analysing Baku must place on the table first. But there is one detail that the prediction tables consistently ignore: Baku is not a place for confirming the existing order. It is a place for breaking that order. A track that in eight editions has produced six different champions, including victories by drivers who were not the highest-rated, is itself an indictment against any apparently certain prophecy.

I have a principle when reading any prediction: before believing the conclusion, check how many genuine witnesses it was built from. A prediction table saying Antonelli will win in Baku, while he is 81 points ahead, sounds plausible. But that plausibility is a cheap plausibility, because it adds no new information. It merely repeats what the standings already say. A prediction that repeats the standings is not a prediction. It is a copy dressed up in the language of machines.

Core: A street stage that spares no one

Why Baku is the harshest test

Baku is a street circuit, and that changes every calculation. The road surface has far less grip than a purpose-built track, especially outside the usual racing line — where dust and rubber accumulate in ways that cannot be precisely predicted. Track temperature, depending on the time of day, can differ significantly between different parts of the same lap. That is why teams usually split tyre strategy into multiple scenarios instead of fixing one plan.

The safety-car risk here is higher than average. A narrow street circuit with walls on both sides turns every small collision into a situation that can change the entire picture. When the safety car appears, the gap a driver has built over an entire race can be erased in seconds. This is a decisive variable, and it cannot be modelled by an algorithm predicting the final result, because the nature of the safety car is random.

If I had to point to one number that most accurately reflects the nature of Baku, it would not be any driver's win count. It would be the six different champions in eight editions. This says that on this track, the quality of the car and the talent of the driver are not enough to guarantee victory. One more factor is needed: luck in managing unexpected situations.

Based on my experience of watching matches, across many different tracks, I have noticed that teams tend to underestimate the psychological factor at Baku. Turn 8, where there is no room for error, creates a particular psychological pressure. Drivers are not only racing each other; they are racing their own survival instinct. That often produces results no prediction model could foresee.

Antonelli and the shadow of an unconfirmed title

Kimi Antonelli is in a position any driver would dream of. He leads by a safe margin, and in theory, simply avoiding major mistakes is enough for him to take the title. But that very position creates a different kind of pressure — the pressure of not being allowed to make a mistake. When you are leading, every risk is asymmetrical. One more win changes little, but one retirement can open the door for those behind.

Eight wins in a season is an impressive achievement. But looking closely at history, I realise that a driver's win count is not a reliable predictor for a specific track. In a season with eight wins, mathematically, most races have seen a familiar result. That means the remaining rounds — including Baku — are where randomness can speak loudest.

What I notice about Antonelli is not the 81-point figure. It is how he manages the run-in phase. A young driver approaching a title for the first time often undergoes a marked psychological adjustment in the final rounds. They start calculating instead of attacking. They start defending instead of conquering. And at a track like Baku, that shift in racing attitude can make a bigger difference than any technical factor.

Russell: The challenger hidden by his own teammate

George Russell is in a paradoxical position. He sits second in the standings, but the 81-point gap to his teammate makes his position more formal than substantive. Only two wins while his teammate has eight is a clear sign of the performance gap, even though both drive Mercedes cars.

Seven podiums is a figure that shows remarkable consistency. Russell is not someone who frequently makes mistakes, and that is precisely his strength. But consistency and the ability to win are two different qualities. At a track where six different champions have been crowned in eight seasons, consistency has value, but it does not guarantee the summit.

For Russell, Baku is a chance to prove he is not merely the man behind his teammate but also a driver capable of deciding results himself. In a team where the number one driver is approaching a title, the role of the number two driver always carries invisible pressure. Russell knows this, and he also knows that each remaining round is a chance to reshape his position within the team.

Verstappen: When the car, not the driver, is the variable

Max Verstappen is in a clear phase of regaining form. Five podiums in the last seven rounds, with third in Italy and second in Spain, show he has rediscovered the feeling of winning. But there is one important detail in this picture: commentary around Verstappen frequently mentions the possibility that his car could betray him at a crucial moment.

That is a form of analysis I am always wary of. When good results are attributed to driver talent and poor results to the car, that is not analysis — it is an interpretation that favours an already-built reputation. Verstappen is one of the outstanding drivers of his generation, but that does not mean every result of his can be attributed to two simple causes.

At Baku, the question is not whether Verstappen has enough talent to win. He does. The question is whether Red Bull can give him a car competitive enough for Baku's particular track conditions. My phone has rung several times this week with sources from the technical world, and what I heard was not absolute confidence. But what comes from technical sources always needs cross-checking before publication, and in this case, I choose to keep a little scepticism to myself.

Hamilton and Leclerc: Two stories in the same garage

Ferrari enters Baku with two drivers at completely different stages of their careers. Lewis Hamilton, who ended his winless streak this year, is still seeking stability in his new colours. His retirement in Madrid due to a brake failure is a reminder that even the greatest drivers depend on the machine beneath them.

Charles Leclerc is in a different position. He sits fifth, 44 points behind Russell, with one win and four podiums. But there is one detail few mention: Leclerc has never won at Baku. That is a remarkable statistic, because Baku is one of the tracks where the skill of pushing beyond limits is prized, and Leclerc is often rated as one of the drivers with the highest such ability.

What is interesting about Leclerc is how he faces statistics unfavourable to him. Drivers who have never won at a particular race are often assigned non-existent pressure by the media — as if they were trying to prove something to some abstract entity. That is a game of media, not of sport. And the best drivers are those who clearly understand that difference.

Norris and Piastri: Strength from behind

McLaren enters Baku with two drivers in good form. Lando Norris has two wins and five podiums, while Oscar Piastri continues to establish himself as one of the most promising young drivers. What stands out about McLaren is not the individual numbers, but the consistency of performance across both drivers.

In a season where Mercedes dominates at the top, McLaren is doing what many teams cannot: maintaining competitiveness without a superstar driver. That is a systemic achievement, reflecting the quality of both the technical and strategic departments. At Baku, where strategic preparation can make a big difference, McLaren is a team I would not be surprised to see produce a notable result.

However, it must be said clearly that strength from behind does not mean the ability to overturn the order at the top. McLaren is racing in a season where the title is nearly decided. Their value lies in creating pressure on the teams above, and in a race with such high randomness as Baku, that pressure can be converted into concrete results.

Lawson and Hadjar: The story of a shared seat

Among the stories I have followed this season, the story of Liam Lawson and Isack Hadjar is perhaps the most systemic. Lawson had a short three-race spell standing in for Hadjar at Red Bull while Hadjar was injured, with results of seventh, fourteenth and sixth. He then returned to Racing Bulls, while Hadjar returned to his seat.

This is a model I have long been familiar with in the sports industry: a multi-team structure that allows drivers to be rotated between different seats within the same system. Red Bull has two teams — the main team and Racing Bulls — and that creates a talent supply channel that can flex according to circumstances.

What stands out about Hadjar is his return after a wrist injury. Drivers returning after time off due to injury often face what I call rustiness in racing rhythm. Their reflexes are slightly slower, their feel for the car has not fully returned, and their ability to read race situations has not reached the necessary sensitivity. That is a factor prediction tables often ignore, but it has a real impact on results.

Gasly and Alpine: The frontier of the race

Pierre Gasly and the Alpine team represent the middle tier of the race — where competition happens with different goals. Not race wins, but points that can make a difference in the constructors' standings race. That is a competition that gets less attention but has important meaning for the teams' financial futures.

At Baku, a track with high randomness, midfield teams always have a chance. The history of this race shows that — six different champions in eight seasons is a clear sign. That chance does not come from midfield teams having better cars, but from random variables that can erase the gap in pure performance.

Contrarian angle: The prophecy industry and the price of certainty

What caught my attention most this week was not any specific prediction, but their form. An algorithm called a supercomputer gave predictions for the entire race, from twenty-second place to first. The predictions were presented with the precision of a confirmed result sheet, rather than with the humility of a probabilistic forecast.

Here, I want to return to a principle I built after years of working with data. Data only becomes rebellion when someone is brave enough to believe in it. But belief in data must come with an understanding of data's limits. An algorithm predicting the result of a street race is not a prophetic tool. It is a probabilistic tool, and probability only means something when presented with corresponding uncertainty.

When a prediction table presents results with high certainty in a context of high randomness, that is not a prediction. It is a media product designed to attract attention. I do not say this to criticise those who make such content. I say it to remind myself, and to remind readers, that the sports industry operates on a foundation where manufactured certainty sells more copies than genuine humility.

There is a blind spot in these prediction tables I want to point out. They consistently ignore the most decisive factors of a street race: the risk of technical failure, the recovery ability of drivers after injury, and the influence of the safety car. The brake failure that forced Hamilton to retire in Madrid is one example. Hadjar's wrist injury is another. These factors are not in the prediction models, because they are unpredictable events. And precisely because they are unpredictable, they are the factors with the greatest impact on the final result.

I wonder whether we are witnessing a shift in how the sports industry produces and consumes information. Previously, predictions were made by experienced commentators who understood the limits of their knowledge. Today, predictions are made by algorithms with no ability to assess their own reliability. This difference is not small. It is the difference between a human who knows they can be wrong and a machine programmed never to admit it.

And this is what I want to emphasise. At a track where six different champions have been crowned in eight seasons, certainty is a more dangerous illusion than scepticism. The sceptic knows they can be wrong. The certain one does not.

Takeaway

Baku will not read the prediction table before the race begins. The walls of the Old City of Icherisheher will not care about Antonelli's 81-point gap, and Turn 8 will show no mercy to anyone who believes the result is settled. The race will be decided by factors no algorithm can fully model: one brake applied too late, a safety car appearing at just the right moment, a driver returning from injury finding his feel at the most important lap. If you are looking for certainty from sport, you are looking in the wrong place. Sport does not exist to confirm what we already know. It exists to show us what we cannot predict — and that is exactly why we still sit in front of the screen every weekend, waiting for something the prediction table cannot say in advance.