Tsunoda, Monza, and the Mistake That Was Not in the Throttle

At the press conference after leaving Monza, Yuki Tsunoda said something I had to rewind three times to make sure I heard correctly. He admitted he had forgotten to activate Straightline Mode, particularly from Turn 8 to Turn 10, and that Monza had more activation points than the previous year. Not late braking, not lost downforce at the second Lesmo. A Formula 1 driver admitting he forgot a button press. And that button press decided his weekend. I have been covering Formula 1 since 2026, setting a record of 406 consecutive live race broadcasts. I watched Senna wrestle a manual gearbox at Monaco, watched Schumacher adjust front brake bias with a hand crank in the cockpit. But I had never seen a driver lose because of a machine-learning algorithm. Until this Monza. To understand what actually happened, context comes first. The 2026 generation car with a nearly 50:50 electric-to-combustion power split, active aerodynamics, and a new control interface. Tsunoda entered Racing Bulls as a stand-in for Liam Lawson, who was promoted to Red Bull after Isack Hadjar's injury. He had never run the new car in the real world before Zandvoort — only hours in the simulator. At Zandvoort, he finished ahead of Arvid Lindblad, just outside the points. A positive data point. But at Monza, everything inverted. He was outpaced by Lindblad throughout practice and qualifying. Not one lap, but a trend. And in the race, he lined up in the wrong grid slot at the restart, prompting Audi to protest that the team was forced into another formation lap. To be clear: these two failures differ in nature. One is a comprehension failure of the new generation. The other is a procedural discipline failure. Both deserve separate analysis. Starting with the technical side, because it is the least discussed. The central problem of the 2026 generation is not a hardware upgrade but the complexity of the driver interface. The car demands active management of Straightline Mode activation points, energy harvesting, and lift-and-coast behavior, all counter-intuitive to the cornering-speed maximization instinct drivers are trained on for an entire career. Monza makes everything worse. It is a circuit with few braking events and long full-throttle sections, meaning energy recovery is extremely poor. Energy must be rationed like food in a siege. Each SM activation costs energy that cannot be recovered in the next section. Miss one point, and you not only lose speed there — you distort the entire energy budget of the lap. But the most notable technical detail, and the one that made me sit down and write, is the machine-learning deployment-mapping constraint. According to Tsunoda himself, an optimal lap requires specific throttle positions 'to avoid confusing the algorithms that map ideal electrical power deployment.' Read that again. The driver is no longer merely slow or imprecise. A wrong input can mis-train or mis-trigger a software layer. This is an entirely new failure mode, qualitatively different from any previous car generation. You are no longer fighting the car. You are fighting the machine-learning model the car is training. And there is a two-variable optimization conflict the old generation did not have in this form. The out-lap: if you do not use energy, tires stay cold. To reach tire temperature, you must drive aggressively. But driving aggressively to warm tires drains the energy needed to charge the battery. You must reach 100% battery, but how do you get there without trading away tire temperature? The previous generation had one variable. The 2026 generation has two, pulling in opposite directions. This is the part data does not speak to. No telemetry table shows you a driver hesitating 0.3 seconds before pressing SM because he fears draining energy too early. But I have been in this paddock long enough to hear that hesitation in the rhythm of an engineer's voice on the radio. Data only says part of it; the rest lies in knowing how to listen. Moving to the procedural side, because that is where Audi is knocking. Lining up in the wrong grid slot at the restart is an operational discipline failure. It has nothing to do with pace, nothing to do with energy. It is about rehearsing procedure. A professional team has a checklist. The driver has an engineer reading the grid slot over the radio. So why did it happen? Here we must read Audi's intent. A works team lodging a protest over a formation lap is no small matter. In the cost-cap era, when gaps between midfield teams are compressed to the point where one position can decide millions in prize money, administrative protests become a legitimate points-harvesting lever. Audi did not protest because it was wronged. It protested because Tsunoda's tenth place has value. If the protest succeeds, P10 could become a penalty. Points shift from Racing Bulls to Audi's direct competitors. A procedural error becomes a constructors'-standings consequence. This is what television viewers do not see: the race does not end when the chequered flag waves. Now the hardest part to say. Tsunoda's self-assessment that 'I think I've done a good job' is not supported by the record. He was outpaced by Lindblad in practice and qualifying, made a grid-slot error, and faces a protest. There is a gap between self-awareness and performance reality worth flagging, not to criticize, but to frame the problem correctly. But here is the counter-intuitive point: Tsunoda's problem may not be Tsunoda. Max Verstappen has said these cars are 'numb' to driver input, and that stand-ins like Tsunoda or Lawson being relatively quick is evidence of that. Tsunoda rejects the framing, calling it 'a very individual thing.' Analytically, both are partly right, and both have economic motives dressed as technical critique. Top drivers have an interest in arguing the car undervalues their skill. Stand-in drivers have an interest in defending the car's difficulty. This is not a pure technical debate. It is a labor market rewritten in the language of telemetry. But Verstappen's own evidence has a hole. Tsunoda was quick at Zandvoort when he had never run the new car outside the simulator. That could prove the car is easy to adapt to. Or it could prove Zandvoort has an energy-recovery profile so different that lack of real-world mileage went unpunished. Monza, where every weakness of a new driver converges, is evidence to the contrary — the car is not easy at all. Two races. A sample too small to conclude anything about either car quality or driver capability. But a small sample has never stopped the media from building either an epic or an indictment. And here is the biggest execution blind spot: simulator-to-track correlation risk. Tsunoda entered the 2026 generation with zero real-world mileage. A simulator does not teach you the feel of cold tires at 9 degrees Celsius on Monza asphalt. Does not teach you how battery performance degrades as cockpit temperature rises. Does not teach you how the machine-learning algorithm reacts differently with 20 other cars on track instead of you alone in a closed room. The team that understands this first gains an advantage. In the 2026 generation, car setup is no longer a mechanical configuration. It is a trained model. And a model is only as good as the data feeding it. The team that logs and models huge volumes of energy traces across circuits accumulates compounding advantage. This is a game Racing Bulls, with midfield resources, may lose before the race even begins. I once audited AC Milan's tracking dataset for the 2026-17 season and found home xG higher than away xG while actual goals were level. The cause: a sensor delayed by 0.2 seconds. Every analysis must carry a note on measurement conditions. Same here. No telemetry, no GPS data, no lap-time gap cited. Conclusions must be downgraded in confidence accordingly. So what gets verified next race? Watch Tsunoda's SM activation points through the main straights. If he activates correctly at the first and second long straights, the issue is comprehension. If he still misses, the issue is the interface. Watch the start procedure at the first restart. If it repeats, it is a team systems failure, not an individual one. And watch how Verstappen drives his car. If a top driver also starts missing activation points, the 'numb cars' argument collapses — and inverts: not that the car dulls skill, but that every driver is relearning from scratch. Every collapse has a precondition; few bother to look beforehand. At Monza, the precondition was not in Tsunoda's brake pedal. It was in an algorithm not taught enough, a driver not run enough, and a protest waiting to be judged. Empty grandstands do not kill the race, but they take away something numbers cannot measure. And in this car generation, what is taken away may be the patience for a young driver to relearn how to race from zero.