Cadillac F1: The Technical Feedback Loop Split Across Three Time Zones

The Canadian Grand Prix ended, and Cadillac walked into a run of ten rounds in which it brought both cars home only twice — Britain and Italy. Two out of ten. For a debut team running two race-winning drivers, that ratio demands an explanation at the level of the system, not at the level of the driver.

Otmar Szafnauer, a former Alpine team principal, said on the High Performance Racing podcast that Cadillac will not contend for a World Championship while design, aerodynamics and simulation sit in three separate places. He constructed a thought experiment: place Mercedes and Cadillac side by side with the same engine, the same chassis and the same drivers. Who wins? His answer is Mercedes.

The argument is not about cheating. It is not a regulatory dispute. It is a statement about organisation: the speed of a technical feedback loop sets the speed of car development. And a statement like that has to be tested with data rather than with the authority of the person making it.

The Operating Picture of a Debutant

Cadillac is the Formula 1 team of TWG Motorsport. It entered its first season with two experienced drivers: Sergio Pérez and Valtteri Bottas. Both have won Grands Prix. That is a start most debutants in the sport's history never had.

The facilities stretch across three locations. The design base sits at Silverstone in the United Kingdom. The aerodynamic facility sits in South Carolina. Another site in Indianapolis is still under construction. Three locations, three time zones, three HR systems, three shipping schedules for parts.

The team sits at the bottom of the Constructors' Championship. Its debut weekend produced a retirement for Bottas and sixteenth place for Pérez. Over the next three rounds, both cars finished — something Red Bull and McLaren could not manage over the same stretch. From Canada onward the curve inverted. Red Bull and McLaren showed evident improvement; Cadillac did not.

I logged these lines in my own race-tracking notes from the start of the season, not to find someone at fault, but to find which variable moved at the same time as the results. That habit comes from a long-standing rule of mine: before believing an explanation, check how much of the variance it actually explains.

Three Sites, One Lengthened Feedback Loop

Aerodynamics is a sport of reconciliation. An upgrade package is only signed off when three data sources agree: the wind tunnel, numerical simulation and the driver-in-the-loop simulator. If those three sources sit in three places, every reconciliation cycle adds latency. That latency never appears on a timing sheet. It appears in the number of rounds in which a team brings a validated upgrade to the track.

When the three sources disagree, the phenomenon is called correlation failure. A part tests well in the wind tunnel and does not work on track. This is the standard mechanism by which dispersed engineering sites pay a price — and it is a direct consequence of geography, not of the quality of the people involved.

A co-located technical base carries two advantages that never show up on a payroll. First, a short feedback loop: aerodynamicists, simulation engineers and designers share a corridor. Second, clear decision rights: when functional groups disagree, the person with final authority sits one floor down.

A three-site structure damages both. It forces every dispute up through intermediate management layers before it reaches the decision-maker. Every escalation layer is a lost week of upgrade work.

There is a third consequence that gets discussed less often. Aerodynamic staff at a satellite site tend to perceive themselves as peripheral to the centre of power. In engineering organisations, that perception is a standard exit catalyst. A team can retain people with salary, but it struggles to retain them with a career path when the path has to cross three countries.

What Convergence Data Actually Says

A debut season has one objective yardstick: the rate of convergence. A team at the bottom of the championship receives a larger aerodynamic testing allowance under the sliding scale of the Aerodynamic Testing Restrictions. That mechanism exists to help the slowest group close the gap faster.

If a team is last in the championship, holds extra testing allowance, and runs two race-winning drivers, and still regresses relative to the teams above it, the resource-level explanation collapses. What remains sits at the level of organisation.

Numbers never lie. Only the people reading them do.

A two-in-ten finish rate is descriptive data, not causal data. It proves a reliability problem exists. It does not prove that the three-site structure caused it. Between those two propositions sits a gap the indictment never bridges.

Four candidate causes exist, and none of them is eliminated in the source material. The first is chassis fragility in an early design cycle. The second is correlation failure between off-track data and the car's actual behaviour. The third is supply-chain and logistics strain when parts must cross three locations. The fourth is driver error.

The fourth carries the lowest weight. When both drivers have won races, system failure becomes more probable than individual failure. That is a probabilistic inference, not evidence. The line has to be marked clearly so that an inference is not promoted into a conclusion.

What deserves attention is the opening stretch. Over the first three rounds, Cadillac finished with both cars while Red Bull and McLaren did not. On a three-round sample that may signal competent baseline engineering. The indictment neither exploits this detail nor explains it. An argument that selects only the data points fitting its conclusion is an incomplete argument.

Costs That Never Reach the Results Sheet

The FIA cost cap limits specified operational expenditure. Infrastructure spending generally sits outside or at the edge of the capped zone. A three-site structure generates two opposing cost lines.

The first is capital cost for facilities — affordable, because it attaches to infrastructure investment. The second is duplicated operational headcount: three HR functions, three IT functions, three administrative functions, three workplace-safety functions. That second line sits inside the capped zone.

The result is a squeeze. A team can build three buildings but cannot comfortably pay for three operating machines. Every unit of salary spent on duplicated structure is a unit not spent on car performance.

This is reasoning from general principle, not a calculation from published figures. No revenue, cost or compliance figure appears in the source. That needs to be stated plainly so readers know where they stand. An honest analysis marks where hard numbers end and labelled inference begins.

One further distinction matters. The FIA Financial Regulations and football's financial fair play regimes are different systems. This indictment alleges no regulatory breach. No disciplinary decision, no sanction and no investigation is referenced. Converting an operational critique into a legal question is an illegitimate move.

Two Experienced Drivers as a Diagnostic

In motorsport, a veteran driver saying the car does not behave as the simulator predicted is a classic early indicator of correlation failure. No such statement is recorded in the material I have.

That silence permits two readings. The first: there is no correlation problem at all. The second: there is one, but the internal voices were never recorded. The data cannot choose between them.

What is certain is this: with drivers of the calibre of Pérez and Bottas, any anomaly in car behaviour deserves a line in the engineering logbook, not only a line in the results table. The absence of that category of data is an evidential gap, not evidence that the problem is absent.

Debutant History as a Missing Yardstick

Formula 1 history shows that expansion teams typically need multiple seasons to converge on midfield pace. That is the standard yardstick for judging a new project, and it is entirely absent from the source material.

Without such a yardstick, every comparison becomes arbitrary. Comparing Cadillac to Mercedes compares a sixth-month-old operation with an organisation that has run continuously across decades and multiple regulation cycles. Comparing Cadillac to its own previous season is a comparison that does not exist.

The correct benchmark sits here: is the team closing the gap against its own learning curve. Answering that requires lap-time data, qualifying-gap data and a trend line across upgrade packages. The source supplies none of them.

The Contrarian Angle: The Part of the Indictment That Holds

When the world goes quiet, I start hearing what the data whispers.

The indictment of the three-site structure has a timing weakness. Indianapolis is still under construction. The team has never operated the fully assembled version of the model it is being criticised for. Judging a model in a transitional state and then drawing conclusions about its steady state is a methodological error.

The second weakness sits with the position of the critic. Szafnauer is a former director of a rival team. That gives him technical authority and simultaneously places him in a structurally adversarial position. His critique should be classified as informed expert opinion, not as a neutral audit.

The third weakness sits inside the argument itself. The scenario of identical engine, identical car and identical drivers is a stipulated condition, not an observed one. It carries large rhetorical force and small evidentiary force. It is a hypothesis awaiting a test, not a finding.

The fourth weakness is the one-sided structure of the source. Not a single statement from Cadillac leadership, an engineer or a team driver appears in the material. An indictment published without a right of reply gets absorbed by audiences as established fact. That is an asymmetry the article itself never acknowledges.

And there is a detail used as rhetoric rather than as data, already noted above. Over the first three rounds, Cadillac finished with both cars while two front-running teams could not. It is a small sample, but it is the only sample in the entire source that favours the hypothesis that the team's baseline capability is not the core deficiency.

Both readings can coexist and do not exclude each other. Good baseline capability and poor coordination structure can both be true inside one organisation. The problem is that the indictment selects only half of the picture.

Packaging and Expectation

One detail concerns how the story was packaged. The original headline puts Pérez's name first while the entire body discusses organisational structure. That is a headline optimised for traffic, not a technical summary. Readers should know that before quoting it.

The expectations set for this project also need re-reading. The team was positioned as a championship project while sitting last in its debut season. The gap between expectation and reality is wide enough to generate media pressure on its own, independent of the team's actual quality.

The mechanism has been seen many times in professional sport. A project launches in the language of the summit, operates at the floor, and only needs one authoritative voice to supply the vocabulary that turns a bad patch into a settled verdict. That voice does not create the problem. It only names it.

The Boundary Between Two Problems That Got Blended

This is the most important technical point in the whole story, and it is routinely skipped.

Two separate problems are blended into one. The first is car reliability — measurable, countable, trackable round by round. The second is organisational coherence — asserted, unmeasured, and not measurable through finish counts.

The indictment uses the first problem as evidence for the second. The data does not license that inference. A low finish rate and a dispersed structure can coexist without any causal relationship. Establishing causation would require a measurable intermediate variable: upgrade lead time, correlation-failure rate, or the number of upgrade packages withdrawn after track validation.

No intermediate variable appears in the source. That is why I place the story's central claim in the category of an untested hypothesis rather than a conclusion.

What to Track

Five signals are verifiable from public data, and they will decide which argument survives.

First, the trend in the both-cars-finishing rate round by round. Sustained improvement would falsify the structural thesis. A persistent two-in-ten pattern would reinforce it.

Second, any change to the three-site model. If design, aerodynamics and simulation consolidate into one location, Szafnauer's critique is confirmed after the fact.

Third, the appointment of a single technical authority holding cross-site decision rights. The emergence of that role would signal that the team has itself identified the coordination problem.

Fourth, an official statement of the primary objective — identity or championship.

Fifth, further Szafnauer commentary. If the critique escalates from general observation to a specific structural remedy, it is becoming a deliberate campaign rather than a single remark.

Tactics are not born on the pitch. They are born in the numbers someone chose to forget.

A Note on Method

I wrote this on the basis of a document set with no named source, no publication date, and a description of a season that cannot currently be verified independently. Every figure inside it is recorded as a citation, not as a verified fact.

That is why the entire analysis sits at the level of hypothesis. When writing about an organisation in dispute between two definitions of success, the only thing I can do correctly is mark where evidence ends, where inference begins, and where the gap sits. The rest has to be answered by time and by the timing sheets.

Conclusion

The Cadillac story is a story about a team in dispute between two definitions of success. One is measured in championships. The other is measured in identity and presence in the sport's largest market.

The available data is enough to say a reliability problem exists. The available data is not enough to say the three-site structure caused it. Readers have the right to demand a test before absorbing an indictment as fact.

What is worth waiting for is not a rapid reorganisation. What is worth waiting for is a clear answer to the question nobody wants to answer: which objective does this season serve, and who holds final authority when the two objectives pull in different directions. If the team can answer that, every resource dispute downstream will find its own arbiter.