Evidence Before Verdict: How the Cost Cap Turned Formula 1 Into a War of Data

In June 2026, at Baku, the grandstands witnessed something strange. Formula 1 cars bounced in waves, their noses slamming into the asphalt and rebounding, again and again, until several drivers later admitted they could barely see the track above 250 km/h. The phenomenon was called porpoising, after the dolphin's motion. But the interesting part was not that the cars bounced. The interesting part was that almost no one in the paddock agreed on what the bouncing actually meant. Mercedes, still wrestling with its no-sidepod aerodynamic concept, called it an urgent safety matter that justified a technical directive from the governing body. Red Bull, the team leading the championship, said most of the bouncing came from the setup choices of its rivals, and that intervening now amounted to changing the rules mid-game. Two teams looked at a single physical phenomenon and drew two opposite conclusions. Because what they were arguing about was not the phenomenon. What they were arguing about was the interpretation of data, and who holds the right to define what that data means. I have spent most of my working life in front of telemetry and lap-time spreadsheets, and I learned something uncomfortable: in motorsport, the thing that decides the final result is rarely raw speed. It is the quality of the evidence each team uses to make decisions. And since 2026, when F1 first imposed a spending cap, that quality of evidence has become the most important competitive variable of all, more important than the budget itself, which is now fixed. To understand why, we need to return to the mechanics of that financial rulebook. In 2026, the FIA brought into force a cap of 145 million dollars for a 21-race season, plus roughly 1.2 million dollars for each additional race. That figure fell to 140 million dollars in 2026 and 135 million dollars from 2026. A list of items sits outside the limit: driver salaries, the salaries of the three highest-paid executives, marketing costs, activities not directly related to racing, and heritage assets. Alongside the spending cap came a second system, the Aerodynamic Testing Restrictions, or ATR. The champion receives the least wind tunnel and CFD time; the last-placed team receives the most, on a sliding scale. On the surface, this is a rulebook about money. Look closer, and it is a rulebook about information. Because once money is capped at nearly equal levels, what remains to separate the strong from the weak is the ability to turn raw data into correct decisions, and the ability to defend those decisions against the pressure of the outside narrative. Autumn 2026 provided the first piece of evidence. On October 10, the FIA published the results of its cost cap certification for the 2026 season: Red Bull was found in minor overspend, while Aston Martin committed a procedural breach. On October 28, Red Bull signed an Accepted Breach Agreement with the governing body: a 7 million dollar fine and a 10 percent reduction in aerodynamic testing for twelve months. A financial figure suddenly became a sporting variable. And immediately, a question landed on the table: what happens to a team dominating the grid when a tenth of its development time is taken away? The answer, as I have written many times in my own notes, does not lie in the number but in the model people use to read the number. Every linear model, more money means more speed, less wind tunnel time means less performance, holds true only in a very narrow band. Outside that band, we enter what I still call the grey zone. The grey zone is not a place that lacks light. It is the place where this sport is most real. To see this clearly, let us split the problem into two layers. The first is the purely technical layer: how many wind tunnel runs a team has, how many concepts each run produces, and how many of those reach the track. The second is the organizational layer: which concept the team chooses to put on the car, and how fast it chooses. The cost cap and the ATR bear heavily on the first layer, but barely touch the second. And in the current era, the second layer is what decides championships. This explains a paradox many fans have raised with me: why did Red Bull, the team punished with reduced development time, win 21 of 22 races in the 2026 season, one of the most dominant campaigns in history? The simple explanation is that the team already had a fundamentally correct aerodynamic platform, so the development headroom that remained, even cut by 10 percent, still exceeded the entire development budget of the teams behind. The deeper explanation is that it had a decision-making system so efficient that every wind tunnel run generated higher-value information than its rivals' runs. When you hold fewer lottery tickets but pick better numbers, you still win. This is the point where I want to pause longer, because it is the core of the whole story. The cost cap did not create equality of performance. It created equality of budget and then left the quality of organization to decide the rest. In other words, it did not flatten the game. It changed the unit of measurement. Before 2026, the unit was money. After 2026, the unit is the conversion rate from data to decision. I do not believe in titles. I believe in the system that operates to produce titles. A championship trophy is the visible output of a chain of decisions that began months earlier, in departments nobody films, in meetings nobody streams. And in the cost cap era, that chain begins with a very concrete question: with the few wind tunnel runs we are granted, which hypothesis do we test first? Every new upgrade is a hypothesis. The race weekend is the experiment. And like any serious experiment, it needs a criterion for judging the result before the result arrives. The team that defines that criterion earliest gains the advantage. Mercedes across the 2026 and 2026 seasons is a painful comparative study: it advanced an aerodynamic hypothesis, the no-sidepod concept, based on its own model, and that model failed to match the track data. But instead of abandoning the hypothesis, it persisted with it longer than necessary. That was an organizational error, not a financial one. Money was not lacking. The evidence was there, but it was read wrongly. By contrast, McLaren is an example of learning speed. Midway through 2026, the team rolled out a series of upgrades that changed almost the entire aerodynamic philosophy of its car, and within a few races it climbed from the back of the midfield to the front. By 2026, it won the constructors' championship, its first since 2026. On the surface, that is a story about development speed. Seen from the organizational layer, it is a story about a team daring to re-read its own data and concluding that the old hypothesis was dead. Aston Martin offers another angle, and perhaps a cautionary one. Early in 2026, the team repeatedly reached the podium with Fernando Alonso, prompting talk of a transformation. Then performance faded as rivals caught up and overtook it in the development race. That is the mark of a team that optimized well for one specific rule set but had not yet built a continuous learning system. The cost cap did not punish it for a lack of money. It exposed a gap in conversion capability. Here I must be careful, because it is easy to slide into a conclusion that is too tidy. The tidy conclusion is this: the cost cap rewards good organizations and punishes poor ones. Reality is more complex. Some well-organized teams still fail, and some average ones succeed thanks to a correct technical platform. Ferrari in 2026 is an example of the reverse case: it had a car fast enough to fight for the title early in the season, but its chain of strategy and reliability decisions threw the opportunity away. That was not a money problem. It was a problem of the quality of evidence used in decisive moments, when the clock is running down and there is no time to re-verify. At this point, I want to widen the frame one beat beyond the track, because the story of the cost cap is not only a technical story. It is a chapter in the financialization of this sport, a process I once observed in another discipline and always view with a wary eye. When a sport shifts from selling tickets to selling commercial rights, data, and narrative, the pressure of financial reporting begins to weigh on the sporting decisions themselves. The cost cap is an attempt to restrain that pressure, but it also creates a new intermediary layer: people who manage numbers rather than manage the car. And as that layer grows, the distance between data and decision can widen rather than narrow. I once wrote that a club or a racing team going public turns the emotion of its fans into money. That is not inherently bad. But it raises a question of priorities: when this quarter's revenue depends on keeping a sponsor happy, is the correct technical decision still prioritized over the easy-to-narrate one? In the cost cap era of F1, that question becomes sharper, because every expenditure must be justified before a spreadsheet. There are twenty cars on the track, but the real race takes place between two brains. One brain manages the wind tunnel, the CFD, and the upgrade schedule. The other manages the story a team tells itself about whether it is on the right path or the wrong one. The cost cap made the first brain more important, but it did not make the second disappear. On the contrary, when every team has roughly the same amount of money, the temptation to self-deceive grows, because there is no longer a shield of "we lack the budget" to explain failure. That is why I always emphasize the discipline of evidence. When I was a journalism student in Turin, an editor told me that tactical analysis written by a woman was merely decorative. I spent 240 minutes reviewing footage and drew 14 pressure maps before resubmitting the piece. Since then, my principle has been: no data, no argument. That principle applies identically to F1. Without lap data, there is no conclusion about performance. Without wind tunnel data, there is no conclusion about development direction. And precisely for that reason, I must state clearly what many commentaries skip: most of what fans read about the cost cap is speculation presented as fact. When a team is penalized, the media immediately draws a performance-decline curve. But that curve does not exist in the data. It exists in expectation. The difference between the two is the entire job of the analyst. Here, it is time for me to argue against myself. If I say the cost cap did not flatten the game, then the opposing side has a strong point: look at the overall standings, and the gap between the leading and trailing teams in recent seasons, measured as a percentage of lap time, has narrowed compared with the previous decade. That is a fact. And if I ignore it, I am doing exactly what I criticize in others. The strongest version of the opposing argument is this: the cost cap prevented the old-style budget polarization, when one team spent three times what another did. It forced the big teams to prioritize, and that prioritization, over the long run, creates a more competitive grid. I accept most of this argument. But I reject the conclusion drawn from it, that the cost cap is a flattening machine. A flattening machine must make teams similar in their ability to generate speed. The cost cap does not do that. It makes teams similar in the money they may spend, then leaves organizational capability to decide the rest. The gap narrowed not because teams became alike, but because the entry threshold, the minimum organizational capability needed to compete, was pushed higher. This is the blind spot I want to name. When people talk about the cost cap, they usually talk about money. When they talk about money, they implicitly assume money is the only variable that can be managed. But the variable that actually changed is time, specifically the speed of decision-making. In an environment where every team is limited in its testing, the team that shortens the loop from hypothesis to conclusion wins. That loop is measured in weeks, sometimes in days. And it appears in no balance sheet. I remember a winter night in 2026, when football and nearly all of world sport stopped because of the pandemic. I used that stretch to build a dataset of pressing phases, logging hundreds of goals to find transition patterns. When stadiums reopened in silence, I wrote about how home teams lost a share of their pressure without a crowd. The lesson I carried from that period is simple: when the environment changes, old patterns lose their reliability, and one must rebuild the model from scratch rather than patch it. The cost cap era of F1 is an environment that has changed in exactly that sense. There is one detail I always remember when thinking about that change. In 2026, Adrian Newey left Red Bull after nearly two decades, and by September of that year he had signed with Aston Martin. In the same year, Lewis Hamilton announced a move to Ferrari from the 2026 season. These two events are usually read as transfer news. Read from the system layer, they are signals that teams are re-pricing the most valuable asset of the new era: human judgment in an environment where the machines have a limited number of runs. When you are allowed only a certain number of wind tunnel runs, the value of the person who knows which concept to test rises. Not because the wind tunnel matters less, but because it becomes a scarce resource, and every scarce resource raises the value of whoever allocates it. This is basic economic logic, but its consequences in sport are anything but basic. It means a team can buy every piece of equipment and still lose to a team with less equipment that knows how to use it. The cost cap cannot buy judgment. It can only buy the conditions for better judgment. And this is where I return to the grey zone, the concept I use to describe situations that fall outside every model. The cost cap is a model. Like any model, it has boundaries. Inside those boundaries, it works well: it prevents reckless spending, it forces teams to be more transparent, it creates a playing field with rules. Outside them, it is blind. It cannot measure the value of a meeting held at the right moment, of the decision to stop developing a wrong direction, of keeping a good engineer instead of letting him leave for a rival. Those things decide championships, and they are invisible to a spreadsheet. If I had to pick one season to illustrate this entire argument, I would pick 2026. It was the season in which Max Verstappen won his fourth consecutive title, but McLaren won the constructors' championship, its first in 26 years. A season in which both results are true at once, and neither can be explained by money alone. Verstappen won through his ability to maximize results on days when his car was not the fastest. McLaren won through its ability to sustain a stable development curve across the whole season. Two kinds of victory, both inside one shared financial rulebook. This leads me to an observation I consider central to any analysis of the current era. The cost cap moved the competition from owning more resources to using resources efficiently. But it did not remove uncertainty. It merely pushed uncertainty somewhere else. Previously, uncertainty lay in the budget: the team that spent more could buy more chances to trial and error. Now, uncertainty lies in judgment: the team that chooses the right hypothesis to test first will have more chances to succeed. Both are uncertainty. Only the unit of measurement changed. And precisely because uncertainty has not vanished, naive prediction models keep failing. People still ask me who will win next season. I usually answer with a line many find annoying: I do not predict the champion. I predict who will run out of development headroom first. Because in an environment where every team is limited, the team that collapses is not the weakest one. The team that collapses is the one that hits its development ceiling earliest, the one that has used up its good ideas and has no new hypothesis left to test while rivals still have road ahead. This is a different reading from the conventional one, and I find it more useful. It does not require me to know in advance who is faster. It only requires me to track each team's learning speed race by race, and compare that speed with its rivals'. A team can be slower across the first three races yet learn faster across the next three. If I only look at the standings, I will miss it. If I look at the learning curve, I can see it before it shows up on the scoreboard. That is why I keep every draft with a timestamp and an analysis log. Not to defend myself against criticism, but to defend myself against the habit of self-deception. When I write a prediction, I record the date and the evidence that led to it. Months later, I read it back and check. Sometimes I am right, sometimes I am wrong. But what matters is that I know why I was right or wrong, and that is the only way a model gets better. I apply that discipline to F1 with the same severity. When a team announces an upgrade, I do not read its press release first. I read the lap data first. I compare speed on the straights, in the slow corners, in the fast corners. I look for where the improvement appears and where it is absent. A correct upgrade must leave a trace in the data. If there is no trace, either the upgrade is not working, or it is working in a dimension I have not yet measured. Both possibilities are worth pursuing, and neither can be resolved by following the story. The cost cap makes this work more important, because it makes every decision more expensive in opportunity terms. When you have unlimited money, a wrong decision costs only more money. When you are limited, a wrong decision costs a slice of budget that could have gone to another direction. Opportunity cost becomes real cost. And when opportunity cost is real, the quality of evidence becomes a strategic asset. I want to close this analysis with a deliberate counter-example, because I do not want my model to look too perfect. There are seasons in which the team with the best system did not win. In 2026, the title fight between Verstappen and Hamilton was decided at the final race, in laps that no model, however good, could predict precisely. That is a reminder that the grey zone always exists, and sometimes it occupies the entire stage. A good system raises the probability of success; it does not guarantee it. Anyone who says otherwise is selling you a model, not a fact. The same is true of the cost cap. It is a good governance tool, not a law of physics. It changes the distribution of advantage; it does not eliminate luck, and it does not turn this sport into an equation. Those who expect it to produce a perfect grid are expecting something no model can deliver. When I look toward 2026, the arrival of a new power unit rulebook with a near-even split between the internal combustion engine and the electrical system, along with active aerodynamics and new manufacturers entering, I see a period in which the financial rulebook itself will be tested. A new technical rulebook creates a nearly level starting point in knowledge, and when knowledge is level, advantage comes from learning speed. In such an environment, the team that builds the fastest learning loop will be the first to break away from the pack. And this is the thought I want to leave behind. The question is no longer which team has the most money. The question is which team can prove its conclusion with evidence fastest, and dares to throw that conclusion away when the evidence turns. In a sport where every team is limited in how many times it is allowed to be wrong, the most precious quality is not speed on the track. It is the speed of honesty toward one's own data.