Antonelli Tops FP2 at the Madring: What the Data Confirms and What It Leaves Blank

17:42, London time. On the third monitor in a small flat in the eastern part of the city, the second practice session (FP2) clock froze at minute 31. A red flag. The first monitor, the official timing board, remained lit: the top line bore the name Kimi Antonelli with the final time box painted purple. The second monitor ran the track-surface temperature chart, its curve climbing. The fourth monitor, the one I always leave blank until a session closes, stayed black.

I sat there for four more minutes before turning it on. I knew exactly what it would display: a long table of continuous runs, a fuel-load column, an engine-mode column, a per-lap tyre degradation column. And I knew exactly how empty it would be.

A circuit that has never appeared in any simulation database. A red flag at minute 31. A championship leader at the top of the timesheet on a soft-tyre lap. The press will call that a perfect start. I call it an afternoon in which the volume of data gathered was smaller than the volume of data lost.

Data is never in a hurry, but people always are.

The timesheet answers one question and hides ten others

The timesheet is the easiest tool to read and the easiest to be misled by in this entire industry. It answers exactly one question: who was fastest over a short window, under conditions that were not recorded. It says nothing about the fuel load in the tank, the engine mode in use, how many laps the tyres had already done, or the surface temperature at that precise second.

In my daily work I separate two kinds of signal. Absolute signal is the time you run yourself. Relative signal is the gap between you and another driver under identical conditions. The second is far more trustworthy, but only when you control the variables. At a familiar circuit you always control them, because you have three to four seasons of data to cross-check against. At the Madring, this is the first season. No baseline, no control sample, nothing to compare.

For Antonelli, the absolute signal is the session's fastest lap. The relative signal, the more valuable one, is that he was described as the more comfortable of the two Mercedes. That sentence does not say how fast the car is. It says who is controlling the car better. In a weekend where nearly every variable is unfamiliar, control is the only thing that holds.

The Madring: a circuit with no memory

A new circuit is not merely a circuit nobody knows. It is a circuit with no memory.

In simulation systems, every circuit exists as a numerical model: gradients, corner radii, surface roughness, the rate at which the asphalt evolves over time. Those models are built and refined over years, using measured data from every time a car runs. The Madring enters this weekend with an almost empty model. Engineers must lean on high-resolution offline modelling and on two on-track sessions to calibrate. Model-to-reality error at a new circuit is always larger than at an old one, and largest in precisely the factors that decide races: tyre degradation, tyre temperature over a run, and the pace of surface evolution.

This is why the history of new circuits repeats a pattern. Friday is compressed and hard to read. Qualifying scrambles the Friday order. The race produces more safety cars than usual, because teams do not yet know exactly when the tyres fall away or where overtaking is possible. The disruption does not come from luck. It comes from nobody having enough data to impose order.

One more layer. The Madring sits inside a new regulation cycle. If the internal bulletins are right, tyre behaviour and energy management under the new power units differ fundamentally from the previous cycle. A soft-tyre groove finding this season cannot transfer straight to next season, nor be compared straight to last season. The baseline is missing not only in space. It is missing in time.

Antonelli: from rookie to leader

At 60, I no longer believe in luck, only in numbers that have not yet spoken.

I have followed Antonelli since his junior days. When a young driver joins a top team, two scenarios usually appear. The first: he is slow, then explodes. The second: he is fast, then fades. Very few take the straight path. Antonelli, this season, is on that straight path, and the most notable thing is not raw speed. It is that he does not repeat mistakes.

I use three metrics to assess a young driver's first season at a top team. The share of quick laps cancelled by personal error. The variance of lap times within a single run. And the ability to regain rhythm after an incident. The last is the sharpest discriminator between a good driver and a special one. A good driver takes three to five laps to return to his old trajectory after being pushed off track. A special one returns on the next lap.

This season, Antonelli leads the championship. That is the only hard fact in the entire story surrounding this weekend, and it cannot be dismissed. Leading a race is an event. Leading a season is a structure. Structures are not built from luck.

And then there was the Italian round. Nineteenth to first. Stripped of emotion, that phrase is a concrete sequence of actions: a correct strategy, a stable run of quick laps, overtaking skill on a surface with changing grip, and almost certainly one or two safety cars exploited properly. Any driver can do one of those four things in a race. Doing all four in the same race is a difference in kind.

I hold to my old view: one race is not enough to declare a great driver. But one race is enough to eliminate a hypothesis. After Italy, the hypothesis that Antonelli is only quick when everything goes his way has been eliminated.

Soft tyres and finding the groove: reading it correctly

The phrase found a groove quickly on the soft tyre sounds gentle. In data, it is not gentle at all.

The soft is the compound with the narrowest thermal window in a weekend's tyre allocation. To exploit it, a driver must reach optimal surface temperature on the first lap, hold it in a very narrow band, and not push past the threshold. Too cool, and you lose grip. Too hot, and you lose it on the next lap. The so-called groove is the state in which a driver finds that balance point on the opening lap instead of taking three or four laps to adjust.

At a new circuit, the surface has not been coated with enough rubber. Grip is low and uneven. That narrows the soft tyre's thermal window further. A driver who finds the soft-tyre groove in such conditions shows two things at once: sensitivity to low grip at the throttle, and the ability to read tyre temperature through feel. The second is far harder to teach than the first, and it is what separates Antonelli from most drivers his age.

But here is the limit. One-lap pace on the soft is a predictor for qualifying, not for the race. In a race, the soft holds for only a few opening laps before degrading. What decides a weekend has never been the fastest lap, but the average pace over a continuous twenty-lap run. And that very dataset is the one missing.

I cross-checked three independent timing sources after the session. All three gave the same picture: one quick lap, a run sequence too short to conclude from, and a large hole in the most important data. The timesheet shows off a driver. The long-run table stays silent.

The long-run gap: the central problem of the weekend

Among the entire body of information around this Friday, exactly one sentence is worth writing down: very little long-run data was gathered on a brand-new track. Every other conclusion depends on that sentence.

What is long-run data and why does it matter so much? It is a continuous sequence of fifteen to twenty-five laps, run with a fuel load approximating race conditions, to measure two things. First, the car's stable average pace across the run. Second, the shape of tyre degradation: does the tyre lose grip gradually, or fall off a cliff past a certain temperature threshold. Those two shapes demand entirely different strategies. Gradual degradation allows a longer stint. Threshold degradation forces an earlier stop than planned and turns the race into a question of timing.

Without long-run data, teams enter the race with a higher degree of uncertainty than usual, and uncertainty always leans toward the wrong decision. A two-stop strategy may be right or wrong, and you only learn the answer after paying the price. This is why new circuits produce higher safety-car rates and higher strategic-error rates. Nobody can predict what they have never seen.

I recall the season I followed Brentford. Brentford does not read the future; it simply reads data more carefully than others. That principle applies here in reverse. With no data to read carefully, nobody holds an information edge. The red flag levelled the field's ability to gather long-run data, and that levelling does not favour Mercedes. It only creates uncertainty for everyone.

What stands out is that Antonelli was among the few who built an edge before the red flag came. A quick soft-tyre lap in FP2, before minute 31, is a real result. It is just real within a very short window.

Track evolution: the enemy of every prediction

The track is evolving quickly, and it is only set to get more rubbered in as the weekend progresses. This is the second key sentence of the story, and its destructive power is greater than people assume.

At a new circuit, surface evolution can reach two to three seconds per lap across a weekend. That means the Friday timesheet order has very low predictive value for Saturday. A team that peaks too early in a session will be overtaken by teams that choose to run later. A driver fastest in FP2 can drop to fifth in qualifying without doing anything wrong.

In data terms, this is baseline drift. Every comparison you make at one moment loses its value when the baseline moves. At a stable circuit, drift within a session is a few tenths. At the Madring, it can be five times larger. This turns the Friday timesheet into a temporary document, not a conclusion.

There is an under-discussed strategic consequence. When grip rises fast, the value of the soft tyre falls. A quick soft lap in FP2 can look impressive, but by qualifying, when the surface has gained two or three seconds of grip, the gap between compounds narrows and teams can run the medium almost as fast. The soft-tyre advantage is diluted by the track's own evolution.

For Antonelli, that does not erase the lap's value. It merely repositions it. The lap proves adaptability, not a gap. And in a weekend where the surface changes hourly, adaptability is a more durable asset than a gap.

Overnight improvement: a structural Mercedes signal

There is one sentence in the bulletin I read more slowly than the others: Antonelli and Mercedes tend to make good steps overnight into the weekend.

This is a claim about a pattern, not about data. But if the pattern is real, it is one of the most valuable structural signals a team can own. The reason lies in how a team operates between days.

After each session, data goes back to the factory. Engineers compare the car's actual on-track behaviour with simulation predictions. The gap between the two is called correlation error. A team with low correlation error can trust its simulation and decide quickly. A team with high correlation error must slow its decision-making, test more options, and often arrives at the next session with a sub-optimal set-up. The ability to make good overnight steps is a direct measure of that correlation capability.

At a new circuit this capability matters even more. You have no historical data to compensate. You only have the model, and its quality determines your learning rate. A team that learns faster over twenty-four hours will move ahead in qualifying.

At Mercedes, the engineering group is among the strongest in the sport's recent history. That they are described as in the best place on Friday, at a circuit with no data, fits the hypothesis that their edge lies in learning speed rather than raw car speed. But a hypothesis remains a hypothesis. The bulletin contains no numbers to confirm it. And an unconfirmed hypothesis is worth only half its weight.

Internal benchmark: the more comfortable Mercedes

Among the few qualitative facts of Friday, one locates relative strength inside the team: Antonelli looked the more comfortable of the two Mercedes.

This is the kind of information I value above a quick lap. The gap between two drivers in the same team, same car, same tyres, same track conditions is the cleanest indicator motorsport data can provide. It removes almost every equipment variable from the equation. What remains is the driver.

That Antonelli is rated higher at this stage of the season is not a great surprise. He leads the championship, and leading at mid-season usually indicates a very high level of consistency. But the phrase more comfortable is also worth reading closely. Comfortable describes the feel of operating the car, not absolute speed. A driver comfortable with a car can often extract more from it in fluctuating conditions, and that ability is worth most in the race, not in a single lap.

Notably, the bulletin does not name Mercedes' other driver. This is a small but meaningful detail. When someone is described as the more comfortable of the two, the information conveyed is about the named driver, not the unnamed one. That silence is itself data, and I read it as a sign that Mercedes' internal contest is tilting one way.

I must also state the limit clearly. A session interrupted by a red flag is too small a sample to conclude anything about internal order. FP2 is not qualifying. Different fuel. Different engine mode. Different pressure. Across my career I have seen no fewer than ten cases of a driver faster in free practice losing to a teammate in qualifying the same day. Small samples always lie, and they lie loudest when people want to believe them.

Dissecting the 19th-to-1st in Italy

The only hard event in the entire file around this weekend is the Italian round: Antonelli from 19th to 1st. I want to dissect it with data rather than emotion, because this is the kind of result the press inflates and analysts overlook.

A recovery from 19th to 1st in a modern race is almost always the product of three reinforcing factors. The first is superior race pace, expressed as consistent overtaking ability and high average pace across stints. The second is strategy, specifically choosing the right pit timing and exploiting the right safety-car window. The third is tyre management, especially the ability to extend a stint when rivals are forced to pit earlier.

Of those three, the second contains the most luck. You do not control when the safety car appears. You control only your decision when it does. The first and third contain the least luck, and they are the durable ones. A recovery from 19th to 1st requires all three, which means it cannot simply be attributed to luck.

At 60, I have watched enough races to know that luck does not accumulate. A lucky driver can win one race. A driver winning from 19th needs more than luck. He needs a car that can overtake, a strategy team that can read a race, and a set of tyres that can carry him to the flag. Combining all three in one afternoon is a result that belongs to capability, not luck.

Now, the part I must state clearly. One race is a sample of one point. No self-respecting data analyst draws a conclusion about a driver from one data point. So I do not conclude that Antonelli is a great driver. I only record what can be confirmed: the hypothesis that he cannot race from the back has been rejected by the data.

The transfer market: value being re-priced

The transfer market is a contest in which whoever prices correctly wins.

In the context of a young driver leading the championship with a top team, the driver market freezes around him. This is a pattern I have seen many times. When a young driver secures a race seat at a top team and starts leading, team management tends to switch immediately into protection mode. That means extension clauses are offered earlier, contract buyout clauses are priced higher, and rival teams are locked out of the race for the signature.

The internal bulletin mentions no concrete transaction. There is no information on contracts, extension clauses, or any negotiation. This means I cannot model any chain reaction in the seat market. But I can record the current state: Antonelli's position at Mercedes is being reinforced by results, and results are the best form of contract insurance a driver can hold.

There is another aspect the bulletin does not mention, but market data always does. An Italian driver at a top team carries enormous domestic commercial pull. That value does not appear on a timesheet, but it appears in every contract negotiation. I have no data to quantify it in this specific case, and I will not invent a number. But I know it exists, and it grows every time this driver wins.

A contrarian angle: Friday cannot predict Saturday

This is the part where I must speak against the way the story is being told.

The story being told is that Mercedes are in a good position and Antonelli is in good shape. I do not dispute the second. I dispute the first, because it is unverified.

There are four reasons to doubt the predictive value of this Friday result.

The first is sample size. One session, cut short by a red flag at minute 31, with no long-run data. This is the smallest sample a weekend can produce.

The second is baseline drift. The track is evolving quickly, and every FP2 lap will lose value in proportion to that rate of evolution. A gap measured at time T no longer holds at time T plus three hours.

The third is engine mode. A soft-tyre lap in FP2 may be run in a higher engine mode than race mode, and with a lower fuel load than race conditions. Both inflate the car's true capability. You cannot know the degree of inflation without fuel-load and engine-mode data. And at a session on a new circuit, teams tend to conceal both.

The fourth, and most serious, is the reliability void. The bulletin contains not one item on engine, gearbox, or any component that can fail. A car leading on Friday can lose the entire weekend to a technical fault on Saturday. This is a standing risk in motorsport, and it appears in no prediction built on a timesheet.

I am not saying Mercedes will lose. I am saying the question of whether Mercedes can stay ahead is an open question, and the very framing of that question is a sign of how uncertain this is. When a driver is genuinely dominant, people do not ask the question. They wait.

The real industry story is the circuit, not the driver

On rereading the whole bulletin, I realise the element with the greatest industry value is not Antonelli. It is the Madring.

A new European round, especially in Spain, is an event whose significance extends beyond one racing weekend. It speaks to calendar-expansion strategy, to widening media-rights markets, to attracting new sponsors in the Spanish-speaking world. In my industry, a new circuit is worth more than a new driver, because a circuit lasts longer.

A new circuit also generates a downstream chain. It creates demand for data-analysis services, for new simulation models, for historical data accumulated over time. In the past five years, the revenue share of sports data analytics has grown substantially, and new circuits are one of the main drivers. Every new circuit is an empty database waiting to be filled. Whoever fills it first owns the information edge.

Every football cycle imitates the data of the previous cycle, but nobody learns. The same principle holds in motorsport. Teams copy each other's analytical models but rarely build a differentiated analytical capability. Those that do, in recent history, are usually outside the traditional group. That is why I wait to see which team handles the Madring best. The answer to that question says far more about the future of the standings than one Friday session's result.

Narrative bubbles and the ignored gap

There is a phenomenon I see repeated every season. After a good Friday result, the narrative jumps. After a poor Saturday, the narrative collapses. The human subject of the narrative does not change across those two days. Only expectations change.

I call this the narrative bubble. It is not an ethical judgment on the press. It is a property of the sports information system, where publication speed exceeds verification speed. The narrative bubble is not wrong on facts. It is wrong on weighting.

In this case, the narrative bubble has a real basis. Antonelli really does lead the championship. Mercedes really was quick in Friday's session. But the real basis of the bubble sits at the season level, while the conclusion drawn sits at the weekend level. This is a common data-hierarchy error, and it is the one I try hardest to avoid.

The biggest ignored gap in this story is technical risk. Nobody discusses it on a Friday, because nothing has broken yet. But reliability is a variable with a distribution, and within a new regulation cycle that distribution is wider than usual. A new power unit has not accumulated enough running hours for teams to know its limits precisely. Every weekend is a test, and every test is a chance for a surprise.

I do not introduce reliability into analysis to make it more pessimistic. I introduce it because analysis without it is incomplete. A model without a technical-risk variable is a model that will certainly be wrong in the long run, however right it may be in the short run.

Treat people with data, not data with people

Across many years working with sports data, I have learned one thing I consider most important. Data does not explain people. People are the skeleton; data is the ruler.

In Antonelli's case, data tells me he was quick in Friday's session, that he leads the championship, and that he recovered from 19th to 1st in a race. Those facts cannot explain how a 19-year-old feels at the top of the timesheet at a strange circuit. No index measures that, and I will not invent one.

Conversely, people do not explain data either. A driver's pressure does not change lap time. A team's mood does not change the shape of tyre degradation. This is why I never write about a race's emotions before checking the numbers. Emotion can be right, but it cannot be verified. Data can be dry, but it can be verified.

My rule in reporting is to keep the data skeleton as the structure and let human detail fill the gaps. A driver can describe his feelings, but those feelings only count when they match the data. When data and feeling conflict, I trust the data. When data is missing, I do not substitute feeling for it. I simply record that the data is missing.

For this weekend, the data is missing in exactly the most important part. And so the most honest conclusion is an open one.

The signals I will track for the rest of the weekend

I do not bet on a driver. I bet on signals. Here are the signals I will track on my monitors, and the conditions under which they change the conclusion.

The first signal is long-run pace. If Antonelli delivers a run of fifteen laps or more with a stable average and low degradation, Mercedes' edge is real. If his long run is merely average, the FP2 quick lap is a pretty number and nothing more.

The second signal is the scale of reordering in qualifying versus Friday. The larger the reordering, the lower the predictive value of Friday's session, confirming we are dealing with a low-signal circuit.

The third signal is the gap between the two Mercedes in qualifying and in the race. If the other Mercedes closes in on or beats Antonelli, the comfort narrative reverses.

The fourth signal is reliability information. Any component change, any penalty, can decide a weekend with no relation to speed.

The fifth signal, and for me the most important long term, is which team handles the Madring best. The answer speaks to learning capability, not luck. And learning capability is the most durable thing in this sport.

An open conclusion

A new circuit. A red flag. A championship leader on top of the timesheet. Those three facts are real, and I do not deny them. I merely refuse to call them a conclusion.

What the data confirms is Antonelli's adaptability on a surface nobody knows. What the data leaves blank is Mercedes' race pace over a long run. That gap will be filled within the next twenty-four hours, and by then every prediction made today will be tested by the one thing nobody can argue with: lap time.

Across more than thirty years of following Grands Prix, I have learned that the winner is not the one who predicts correctly. The winner is the one who keeps a cool head long enough to read the data before it becomes a result. The race at the Madring will tell us what Friday's session did not yet say. And for a man who sat in front of four monitors all afternoon, that remains the most interesting part of the weekend.