The 2026 Data Vacuum: When Every Analytical Model Returns N/A

At three in the morning in Liverpool, I opened a spreadsheet with 42 tabs and not one of them returned a result. The column headed "Advancement" read N/A. The column headed "Track validation" read N/A. The column headed "Cost cap room" read N/A. Six weeks of work, four paid data feeds, and a five-layer verification process I built after the summer of 2026 all resolved into a single, neatly formatted, empty file.

I was not shocked. I felt relieved in a way that is hard to explain.

Sports analysis has a zone where every model returns N/A. The cause is not laziness, and it is not a data paywall. The cause is that what is happening on track has never happened this way before, so every extrapolation from the past loses its validity. Formula 1's 2026 technical cycle is one such zone. The transfer market that accompanies it is another.

An analysis that returns N/A can be the most honest analysis in the room. Most writers choose to fill that gap with storytelling. I want to explain why I chose the other route, and why a blank cell in a data table sometimes carries more information value than a number inserted just to complete the row.

Context: a season with no precedent to compare against

Formula 1 enters a new rule cycle from the 2026 season. The FIA published the technical regulations in June 2026, and the changes are not marginal adjustments. The engine keeps its 1.6-litre turbocharged V6 architecture, but the energy balance shifts hard toward electricity: internal combustion output falls to roughly 400 kW, electrical output rises to 350 kW, close to a fifty-fifty split. The MGU-H is removed entirely. Fuel must be 100 percent sustainable. Cars are around 30 kg lighter, 100 mm narrower, with maximum wheelbase cut by 200 mm. Front and rear wings become fully active, replacing DRS with two aerodynamic states, X-mode and Z-mode, plus an override mode reserved for the chasing car.

Anyone who has built correlation models for the 2026 to 2026 seasons understands the problem. That dataset describes a different aerodynamic philosophy, a different energy map, a different weight distribution, and a different tyre degradation profile. Extrapolating it to 2026 means fooling yourself with a beautiful model.

At the same time, team structure is changing at a layer deeper than performance. Audi has taken over the team based in Hinwil and turned it into a works operation. General Motors has brought Cadillac in as the eleventh team, initially running customer engines before developing its own power unit. Red Bull builds its own engines with Ford. Honda has moved to Aston Martin. Alpine has shifted from works engines to Mercedes customer units. These are supply-chain changes, and supply chains do not show up in a lap-time table.

One more variable: the cost cap. Every team must split resources between racing the current season and developing the 2026 car. This is a two-objective optimisation problem with no precedent at this scale, because the last comparable regulation change was 2026, when the cost cap was still immature and an eleventh team did not exist.

And then the transfer market. The new team's seats have not been announced. Some contracts have unclear end dates. Some teams have not settled their technical philosophy, so they do not yet know what kind of driver they need. If you open a spreadsheet and fill in a column for "probability of seat change" across every team, you will write N/A on most rows. That is exactly what I did, and I left it that way.

The core: five verification layers and an N/A conclusion

My analytical framework runs on five layers. Layer one is raw data. Layer two is circumstance, meaning rules, circuit, weather, calendar. Layer three is head-to-head history. Layer four is public statements from teams and drivers. Layer five is the contradiction between the four layers above, and that is usually the layer that generates insight.

The strategy machine does not run on emotion, it runs on information. When all five layers return data that cannot be compared, the machine is not broken. It simply reports that the input is insufficient to produce an output. The operator's job is to read that signal correctly rather than lowering the confidence threshold until the model runs.

Layer one: raw data. Here I have data, and plenty of it. But 2026 data describes an aerodynamic configuration that the 2026 rules will largely invalidate. A driver who is good at managing the front tyre in high temperatures is still a good driver. But how much that skill matters in 2026 has not been measured, because the aerodynamic load distribution and the behaviour of the regenerative braking system will differ. Raw data answers the question of who is performing well. It does not answer which skills will be paid for over the next eighteen months.

Layer two: circumstance. This is the only layer I can write about with certainty, and I deliberately keep it short, because it describes the conditions of play rather than the outcome.

Layer three: head-to-head history. There is one genuinely useful data point here. Major rule changes in the past have produced reordered hierarchies. The team that performed best before a rule change does not automatically perform best after it. But that history does not tell me which team will reorder this time. It only tells me that assuming the order holds is a weak assumption.

Layer four: public statements. Every team says it is happy with progress. No team says it is behind. This is the layer with the lowest information value and the one most quoted by media, because it is the only layer that always comes with a complete, ready-made sentence.

Layer five: contradiction. In the 2026 cycle, contradiction sits mainly between statements and allocated resources. A team that says it is focused on 2026 while still bringing upgrade packages to the final three races of 2026 is saying two different things. That kind of contradiction produces a signal, but not a result. It speaks to intent, not to pace.

The output of five layers is N/A. And I leave it as N/A.

Why I do not fill the gap with storytelling

This industry has a highly effective defence mechanism: when there is no data, people switch to storytelling. A story does not need five verification layers. It needs a character, a conflict, and an open ending.

In a transfer window, that mechanism runs at full power. A rumour is published, then quoted by three other outlets, then loops back to the original source as if the original source were quoting itself. After four iterations, the rumour has enough "basis" to become a row in another outlet's summary table. I have watched this process construct a complete transfer in readers' minds before anyone signed anything.

The defence is not to deny rumours. The defence is to rank rumours by evidence structure. A rumour has three kinds of pieces.

The first piece is money. Release clause structure, payment terms, who pays wages, who holds image rights. This is the hardest piece to fake, because it requires two signatures.

The second piece is organisational movement. A team hires a new technical director. A team opens an office in another city. A team adds simulation engineers. These moves do not say who will sign, but they say what kind of driver that team is preparing for.

The third piece is time. Contracts have specific end dates. Payroll has cycles. New teams have launch dates. Time is the one piece that cannot be negotiated.

When the three pieces do not match, you are reading a story. When they match, you are reading a structure. In the 2026 transfer market, most of the rows I checked fell into the first category.

There is one technical detail in driver contracts I always read first: the performance clause. A contract may run three years on paper, but a performance clause can turn it into one year if results fall below a threshold. This clause exists in both football and Formula 1, and it is usually skipped in contract summaries because it carries no attractive headline number. Yet it determines the real timing of a vacant seat.

The contrarian angle: N/A is a form of conclusion

Readers often treat N/A as an analyst's failure. I think the opposite is true in most cases.

An analytical framework only matures after reality refutes it. If my framework has never been wrong, I have no evidence that it works. And the clearest way a framework demonstrates maturity is not producing more predictions, but knowing when to refuse a prediction because the input is insufficient.

There is a paradox here. The more data you have, the easier it is to become overconfident in extrapolation. A model run on the 2026-2026 dataset can produce a very pretty number for 2026. That number is not mathematically wrong. It is epistemologically wrong, because it assumes the relationships between variables remain fixed while the rules of the game have changed.

This is where I have to bring up an old lesson. My mistake is named Kanté, and I do not want to forget it.

In the summer of 2026, a local sports site in Liverpool asked me to write a prediction piece for the World Cup final between France and Croatia. My article contained two errors. I misspelled a midfielder's name. And I recorded three tackles for him when the correct figure was four. The match finished 4-2, and the site was mocked by readers for a week. I deleted the piece, reopened the entire tournament dataset, and built a five-step process: cross-check the source, review the footage, verify the count, ask a specialist, and wait thirty minutes before publishing.

The lesson was not that I miscounted a tackle. The lesson was that I published a number I had not cross-checked, simply because I needed the article to have a highlight. That is the same mechanism as filling a gap with storytelling. The writer fears white space, so they fill it with something approximately right.

Since then, every statistical claim I write carries a source. I write more slowly. But errors of the "I heard" variety have largely disappeared.

There is a toxic version of this lesson I have to warn myself about. If carefulness turns into self-torture, I will never publish anything. What I do now is limit the self-review section to a short paragraph, enough to keep discipline without turning the piece into a long confession.

A cross-discipline comparison: when the stadiums were empty

In 2026, European football leagues played behind closed doors. I spent most of that period collecting data on home advantage. The results were not uniform across leagues, but a trend appeared: home advantage fell in many leagues, and the size of the fall differed between leagues. That suggests most home advantage comes from the stands rather than the turf.

The empty-stadium season and the home-advantage problem is a clean example of changing a single variable and watching the whole system shift. The 2026 rule cycle is not clean in that way. It changes many variables at once, across many layers, within a single season. So it does not produce a natural experiment. It produces a mixture in which no variable can be isolated.

Players change, stands change, but the advantage problem stays exactly where it was. The question is always the same: where does the advantage come from, and can it be transferred. The difference is that football offers data from many leagues to compare against. In Formula 1, regulation cycles occur a few times per decade, and no two are alike. The sample is too small for statistical inference.

Watching esports taught me football; watching football taught me money flow. In esports, data is nearly complete because every action passes through a server. In football, data comes from cameras and human coders, so error lives in the labelling stage. In Formula 1, data comes from sensors with high precision, but its meaning depends on the regulations in force. Three environments, three levels of cleanliness. The principle is identical: the quality of a conclusion cannot exceed the quality of its input.

The Vietnamese league story

In Vietnam, the data story has a different shape. The national top flight produces enough matches to generate a sample, but the public data infrastructure remains thin. Analysts must code by hand, build their own datasets, and own their own margins of error.

The Vietnamese transfer market carries one structurally notable feature: smaller clubs often take players on loan with an obligation to buy. In accounting terms, the cost does not land immediately. In sporting terms, the small club is developing semi-finished products for a bigger club. When the loan matures, the small club must buy a player it has already used for a full season, at a price fixed in advance, while its wage bill has already been locked by that very obligation. The following year, there is no room left to buy in another position.

This is the kind of structural problem that statistical tables do not display. If you look only at goals and assists, you see a player performing well. If you look at contract structure, you see money flowing in a predetermined direction, and a club paying for its own success with next season's spending room.

That is also why I always read release structure and wage structure before reading match statistics. Do not ask who plays well; ask which side the system is standing on. A good player in an unfavourable system produces ordinary numbers. An ordinary player in a favourable system produces beautiful numbers.

Referees, VAR, and the same transparency problem

There is an intersection between Formula 1 and football that I rarely see discussed: both are sports where official decisions directly affect outcomes, and both are struggling with the problem of explanation.

Formula 1 runs a stewards system with published documents. Viewers can read the decision, the reasoning, and the penalty. Football has VAR, but the in-stadium explanation mechanism remains weak. Fans in the ground often cannot hear why a decision was made, while television viewers hear part of it through commentary. The result is two audiences with two different levels of understanding of the same event, and the forgotten group is the one that paid to be there.

When fans are not given an explanation, they build their own. Those explanations tend to lean heavily on assumed intent. And once assumed intent becomes the default, every technical argument becomes an argument about belief. That is an environment where data cannot correct anything, because no one agrees on which data is valid.

Transparency in sport requires publishing the process, not stopping at publishing the result. A decision with a stated reason generates argument about the reason. A decision with no stated reason generates argument about the person. The second kind of argument never ends.

Why I write slowly

I have a bad habit of procrastinating. I often promise a publication date and then move it, because I want every number verified to the maximum. Once I spent nearly a week on a small statistical table.

The paradox is that this very habit saved me in the 2026 cycle. Because I write slowly, I had enough time to realise that my spreadsheet was not short of data. It was short of the conditions under which data means anything.

But I also know my limits. Delaying too long turns carefulness into a form of avoidance. What I do now is set a data freeze date before submission. After that date, I add no new data, edit no table, reopen no closed tab. I accept that the analysis will contain gaps and I mark those gaps explicitly.

This matters to readers. A piece that states plainly that the data is insufficient for a conclusion is more useful than a piece that reaches a conclusion and buries its assumptions in the fourth layer, where few readers ever go.

Conditional forecasts: three scenarios for 2026

I do not issue unconditional forecasts. I issue three scenarios with activation conditions and verification deadlines, and I will return to compare them when the season closes.

Scenario one. If a team has full works power unit resources and started its 2026 chassis programme at least two quarters earlier than the rest, that team has a high probability of sitting in the leading group in the first half of 2026. Observation condition: the date it stopped updating the previous season's car. Verification deadline: round eight of the 2026 season.

Scenario two. If a customer-engine team keeps a stable technical platform while its engine supplier concentrates resources on the 2026 package, that team may lose relative position early in the cycle and recover later. Observation condition: the pace gap between the customer team and the works team using the same engine, measured race by race. Verification deadline: round twelve.

Scenario three. If the cost cap forces teams to choose between updating the old car and developing the new one, the team with the more accurate simulation model will reallocate resources more efficiently. Observation condition: the number of upgrade packages brought to the track in the second half of the previous season. Verification deadline: the end of the 2026 season.

These three scenarios share one property: each can be refuted. That is what I want. A prediction that cannot be refuted is not a prediction; it is a sentence with no expiry date.

Takeaway

A data vacuum is a form of output. The writer's job is to record its exact depth, leave the white space intact, and state clearly what is missing. When the 2026 season closes, I will reopen this 42-tab spreadsheet and check every row. That is when my framework gets another chance to mature, and also when I will learn precisely where I was wrong.