The Empty Column and the Discipline of Reading a Race Track

At 10:47 p.m. on a Sunday, the third monitor in an east London flat displayed a lap-time table for a race that had ended four hours earlier. Column eleven, the gap between two cars during the safety car period, was completely empty. Not a display error. The organiser's GPS feed had lost signal for exactly the seven laps I needed most, and the system returned an empty column instead of an error message. The story was already written in my head. A driver losing two seconds a lap. A team keeping its tyres in a better window. A perfectly timed pit call. All of it plausible, all of it matching what viewers saw on television, all of it enough to build a smooth 1,400-word analysis with numbers, charts and a tidy conclusion. I wrote 1,100 words and deleted all 1,100. That is the entire content of this article. Not a race, not a championship decider, but an empty column in a spreadsheet and the question attached to it: when the data does not arrive, what is an analyst supposed to do? CONTEXT: THE TRADE OF READING WHAT WAS NEVER RECORDED A modern Formula 1 car carries roughly 300 sensors. The FIA standard electronic control unit records every timing-loop crossing to the thousandth of a second. GPS splits each lap into hundreds of measurement points, allowing the car's path to be reconstructed in three dimensions. A single session produces hundreds of gigabytes of raw data streamed from the circuit to the factories of ten teams. To an outsider, it looks like a paradise of numbers. To an analyst, it is a minefield. Because abundant data does not mean sufficient data. A sensor failing on lap 40 creates a gap that drags down every inference built on top of it. A GPS feed dropping out during a safety car period renders every model of the gap between two cars meaningless. And the worst part: the system does not raise an error. It simply leaves the cell blank. I came to the racetrack from somewhere else. Born in Vietnam, raised on football, I learned to read a match through the spaces between the lines. When there was no football, I drew football. And it turned out that drawing is also a way of understanding. My first diagrams were shaky hand-drawn lines on PowerPoint, redrawn until the sketch looked ugly enough to be believed. Every tactical diagram begins with a shaky hand-drawn line on PowerPoint. Moving to Formula 1, I carried that habit with me. Instead of measuring a defender's position, I measure a corner radius. Instead of counting passes into the box, I count the laps a driver keeps tyre temperature inside the optimal window. The geometry of space changes its units, but not its principle: the gap between two objects always carries information about the intent of whoever created it. That is why I never accept a single source. Every figure that appears in my work has passed through at least two independent sources. A number that appears only once is an unverified number. I call it double verification, and it is both the cheapest and the most expensive discipline in this trade. SOME DATA NEVER APPEARS ON ANY TIMING SCREEN The cost cap and the aerodynamic testing restrictions are two categories of data that live outside every timing table. Testing allowances are allocated in reverse order of the previous season's constructors' standings: the champion receives the smallest share, the last-placed team the largest. It is a self-levelling mechanism, and it operates in complete silence. No spectator in the grandstand sees it. No television graphic displays it. Yet it determines the pace of every team's car development for the following eighteen months. A team that understands the mechanism plans its upgrades as a staircase, spreading resources across many small packages. A team that does not understand it pours its entire allowance into one large package and then runs dry. The difference does not show up in a single race weekend. It shows up two seasons later, when it is far too late to fix. I learned to read these numbers early, not because they are exciting, but because they are the one category of data no team has any incentive to publish clearly. When a source does not want you to see something, it is usually the most important thing there is. AN EMPTY COLUMN IS NOT A NEUTRAL RESULT I learned this in the summer of 2026, when the stadiums closed and I sat rewatching 74 matches purely to count transition sequences. That summer taught me that a gap is never truly empty; it is simply waiting for the right reader. But there is a fundamental difference between a gap on a football pitch and an empty column in a spreadsheet. A gap on the pitch is a physical gap: it exists because somebody moved out of it, which means it always means something. An empty column is a technical gap: it exists because somebody failed to record something, which means it may mean nothing at all. An inexperienced analyst collapses these two kinds of gaps into one. He sees an empty column, fills it with speculation, and because his speculation matches the story people want to hear, nobody objects. I used to be him. In July 2026, at the World Cup in Russia, I wrote a piece predicting Croatia would win on the back of 62 percent possession and six players covering more than 12 kilometres per match. Croatia did win. The article drew 4,200 reads. But when readers asked why Russia still generated so many dangerous counterattacks, I had no answer. I had missed every transition data point entirely. Russia 2026 was not merely a warning about transition. It was a warning about how we read a match. Since then I have kept my own spreadsheet recording every transition sequence. At the end of every article I add a section called Data Limitations, listing exactly what I could not measure. It is the least-read section and the one I write most carefully. FOUR TIMES AN EMPTY COLUMN CHANGED A CONCLUSION Abu Dhabi, 12 December 2026. Nicholas Latifi crashed at Turn 14, bringing out the safety car on lap 53. Lewis Hamilton led on a set of hard tyres that had already completed 38 laps. Max Verstappen pitted for fresh softs. On lap 58 the race restarted with exactly one lap remaining, and Verstappen passed Hamilton at Turn 5 to take the championship. The popular story is that an official's decision decided the title. The data says something else, and it says it through an empty column. We have pit-stop times, tyre ages, and the gaps before and after the safety car. We have no credible model for a race lasting a single lap. No historical dataset is sufficient to compute the probability of holding the lead on 38-lap-old hards against fresh softs over one lap, because that situation had never occurred in that form. Every figure produced afterwards is a reconstruction, not a record. That does not mean Mercedes and Hamilton had no choice. The fork was on lap 54, when they decided not to bring Hamilton in and instead keep track position. Had they pitted, he would have lost the lead and needed to re-pass within seven laps. Had they stayed out, they were betting the race would finish under the safety car. Both branches carried probability. Neither branch carried historical data to lean on. This is where data ends and judgement begins. An honest analyst has to say so, rather than assigning a decision a probability it never had. Singapore, 17 September 2026. Carlos Sainz won for Ferrari. Read purely as a lap-time chart, he looks like he is struggling: his pace in the final stint is markedly slower than his own pace in the first. But the gaps are the story. Sainz deliberately slowed to keep Lando Norris within DRS range, and a Norris with DRS became a barrier the two Mercedes behind could not pass in order to attack Sainz. A fence made of distance. A defensive geometry built out of the aerodynamic rulebook itself. If the data column for the gap between Norris and George Russell were missing, this entire tactic would vanish from the story. The timing table would still be enough to write an article, but that article would be wrong in substance. This is the clearest illustration of the principle: transition is not a stretch of running. It is the silence between two intentions that few people know how to read. The Red Bull penalty, October 2026. Red Bull were found in minor breach of the 2026 cost cap. The penalty: seven million US dollars and a ten percent reduction in aerodynamic testing time over twelve months. No column on any timing screen displays that penalty. No spectator sees it during a race weekend. But it determined the pace of Red Bull's car development over the next two seasons, and therefore determined the outcome of the 2026 and 2026 championship fights. THE PIT WINDOW: WHERE THE SILENCE LASTS LONGEST An average Formula 1 lap lasts roughly 90 seconds. A pit stop lasts about 2.5 seconds. But the period that truly decides a race is not those 2.5 seconds. It is the in-lap and the out-lap, two laps on which the driver runs slower than normal rhythm, on worn tyres, with nobody racing him. That is the silence I care about most. In strategy analysis, the pit window is where every model meets: track temperature, tyre degradation, the gap to the car ahead, the gap to the car behind, straight-line top speed, and pit-loss time. A race is decided by which team reads the relationship between those variables more accurately on each specific lap. The undercut works for a very simple reason: fresh tyres are faster than old tyres, and the time lost in the pit lane is smaller than the time fresh tyres claw back. Conversely, the overcut works when the old tyres remain good enough to hold the gap while the car ahead has already pitted and become stuck in traffic. There is no universal formula. Some circuits make the undercut almost always decisive. Others have pit entries so long that the overcut becomes the default option. An analyst must know which kind of circuit they are looking at before saying anything at all. This is why I spend the most time on laps where nothing happens. Not the lap with an overtake. Not the lap with a crash. But the laps where everything runs normally, because that is when a team reveals its real plan through the steadiness of its pace. THE DRIVER MARKET AND THE CELLS THAT CANNOT BE FILLED On 1 February 2026, Ferrari announced that Lewis Hamilton would join the team from the 2026 season. In September 2026, Aston Martin confirmed that Adrian Newey would start work from March 2026. Both announcements are public data: dated, sourced. But the question fans care about most, why that team and why that moment, has no data cell at all. This is the territory where I am most careful. Managers have an incentive to leak information in their client's favour. Teams have an incentive to stay silent until everything is signed. Journalists have an incentive to publish ahead of rivals. Those three incentives combine to create an environment in which the most reliable source is usually the one saying nothing. A prolonged silence in the driver market is a signal. But it is a weak signal, and it is only worth writing when at least two independent sources are silent together. If only one source is silent, the likeliest explanation is that they are busy. THE DISCIPLINE OF READING SILENCE I have a process. It is not pretty, and I have no intention of making it pretty. Step one: list every number required to answer the question. Not every number available, but every number necessary. Step two: flag those numbers that already have at least two independent sources. Step three: for numbers with a single source, or none, write out why they matter. Step four: if the gap sits somewhere that could reverse the conclusion, convert the conclusion into a question. A botched pit stop is not a mistake. It is data the system is trying to send you. If a team loses 4.2 seconds on a tyre change when their normal stop is 2.4 seconds, that is a signal. If a driver loses 0.3 seconds in sector three while being fastest in sectors one and two all session, that is a different signal. The error is not the bad number. The error is calling it a mistake before asking what the system is trying to say. In my spreadsheet, every cell carries a colour code. Green means double-verified. Yellow means a single source. Red means no source. Before writing, I count the red cells. If the red cells sit where they cannot reverse the conclusion, I keep writing. If they sit where they can, I stop. That is why, that Sunday night, I deleted 1,100 words. Column eleven sat exactly where it could reverse the conclusion. Without it, every sentence about safety car strategy would be speculation dressed up with numbers. THE BLIND SPOT OF BEING CAREFUL But if I stopped the article here, I would repeat my own mistake in a different form. Data discipline has a blind spot: it makes analysts believe that once data is complete, the conclusion will simply present itself. That is not true. Some things cannot be measured, and they are frequently the things that decide. How many hours a driver slept before a race. How long it takes a race engineer to trust his own instinct. A 7 a.m. meeting where somebody decides not to change the front wing because of something a junior said. No sensor records that moment. The shaky lines on PowerPoint do not originate in data. They originate in an impulsive decision at two in the morning, when the person making it is no longer lucid enough to argue with themselves. I draw first and look up the numbers second. Admitting that matters more than hiding it behind a spreadsheet packed with figures. And there is another, larger professional risk: when an analyst publishes a Data Limitations section, a portion of the readership concludes that the piece is untrustworthy. But honesty about a gap does not reduce the value of a conclusion. It limits the scope of its application. Those are two different things. The real danger lies on the other side. Readers grow used to clean articles, with no gaps, no risk, no doubt. After a while they begin to treat that cleanliness as the standard. And at that point, the honest analyst becomes an irritant. I choose to be an irritant. WHAT TO READ NEXT The empty column is still on the third monitor. I will not delete it. I will leave it in the file, flagged red, and return to it when a second source appears. At the end of this season, I will cross-check. I want to know whether the story I almost wrote that night matches a conclusion built from complete data. If it matches, I have proved nothing beyond the fact that I was lucky. If it does not, I have one more reason to keep the discipline. An empty dataset does not tell you that you are wrong. It tells you that you do not yet have the right to conclude. Those are two different sentences, and in a sport where every decision is made within seconds on incomplete information, the distinction between them may be the only difference between an analysis and a carefully written guess. Next time you read a smooth analysis of a race weekend, try counting how many empty columns were quietly filled in without anyone telling you. Method note: every factual claim in this article is drawn from public FIA records and official team communications. Sections that could not be cross-verified were left blank rather than inferred.