Jonas Vingegaard refused to work with Remco Evenepoel. The internet called it cowardice. Game theory calls it rational. Both are right — and that’s the entire problem.

With roughly 30 kilometres left in Wednesday’s third stage of the Volta a Catalunya, Remco Evenepoel launched a violent acceleration through crosswinds that split the peloton in two. Only Jonas Vingegaard could follow. The pair found themselves off the front together, a gap growing behind them, and what looked like a certain stage win materialising. Then the argument started.

For the next 25 kilometres, Evenepoel drove the move relentlessly while Vingegaard sat on his wheel with the studied contentment of a man reading a newspaper on a Sunday morning. The Belgian was visibly — then audibly — furious, waving his arm and flicking his elbow at the Dane as the gap yo-yoed between comfortable and precarious. The internet did what the internet does: split down the middle.

One camp called it a lack of panache, a betrayal of the sport’s cooperative spirit, the behaviour of a man who wins Tours de France by spreadsheet. The other camp pointed out, correctly, that Evenepoel would have destroyed Vingegaard in a sprint and that sitting was simply the intelligent move. What neither camp did was explain why both of those things can be simultaneously true, and what that means for the sport. For that, you need a brief excursion into game theory.

The Prisoner’s Dilemma on Two Wheels

In 1950, two mathematicians at the RAND Corporation formalised a problem that turns out to apply rather well to cycling breakaways. The scenario: two people acting in their own self-interest can end up in a worse position than if they had cooperated, even though cooperation would have benefited both. They called it the Prisoner’s Dilemma.

The cycling version requires no imagination. Two riders escape the peloton together. They can each choose to pull — take a turn at the front, spending energy to maintain the gap — or sit in — stay on the wheel, rest, and let the other rider do the work. Think through the four outcomes:

• Both pull: the gap holds, both arrive to contest the finish. Fatigue shared equally. Best collective outcome.
• You pull, they sit: you burn matches, they arrive fresh, they win the sprint.
• You sit, they pull: free ride for you. You win the sprint.
• Both sit: nobody drives the pace. The peloton catches you both. You get nothing.

The structure is brutal. No matter what the other rider does, you are better off sitting in. If they pull, you get a free ride and arrive fresher for the sprint. If they sit, you avoid wasting energy in a doomed move. Sitting is, in the language of game theory, a dominant strategy — it beats every alternative regardless of what your opponent chooses.

Vingegaard’s calculation on Wednesday was not complicated. Evenepoel, with his aerodynamic position and explosive sprint, would almost certainly have beaten him in a two-up finish. Working did not improve Vingegaard’s odds — it improved the odds of Evenepoel winning. Sitting was not cowardice. It was the correct solution to the problem as stated.

The difficulty is that the problem as stated is the wrong problem.

When the Game Repeats

The prisoner’s dilemma produces its dismal result only in a one-shot game — two strangers who will never meet again. The moment the game is repeated, the mathematics changes entirely.

In the early 1980s, the political scientist Robert Axelrod ran a now-famous experiment. He invited game theorists, economists, and mathematicians to submit strategies for a repeated prisoner’s dilemma tournament. Entries ranged from the highly sophisticated to the cynically ruthless. The winner, across multiple rounds, was the simplest entry submitted: a strategy called Tit for Tat. Its rules occupied four lines of code. Cooperate on the first move. After that, do whatever your opponent did last time. Punish defection with defection. Reward cooperation with cooperation. Never hold a grudge beyond one round.

Tit for Tat won not because it was aggressive, but because it was legible. Other players could read it, predict it, and respond to it. It communicated a credible commitment: I will work with you if you work with me, and I will not work with you if you sit in. This transparency created the conditions for cooperation to emerge without any explicit agreement needing to be made.

Professional cycling is a repeated game. The riders at the front of the Volta a Catalunya on Wednesday were not strangers. Vingegaard and Evenepoel share the same races for years. They will be off the front together again. The question Axelrod’s tournament raised — and that Wednesday’s stage implicitly posed — is whether Vingegaard’s rational defection has consequences that stretch beyond that particular 30-kilometre window.

A rider known to sit in faces a specific problem: the next time an escape forms and a potential partner is deciding whether to bridge across, they have to factor in the probability that their effort will be entirely unreciprocated. Rational partners adjust their behaviour accordingly. In Axelrod’s tournament, pure defectors did badly over the long run not because they lost individual encounters, but because cooperative players learned to avoid them. The punishment was not retaliation — it was exclusion.

Pogačar and Van der Poel: A Working Example

For a different model, consider the partnership — informal, unspoken, entirely self-interested — that has developed between Tadej Pogačar and Mathieu van der Poel across three spring Classics.

Milan-San Remo 2025. Pogačar attacked relentlessly on the Cipressa and again on the Poggio, knowing that Van der Poel was the better sprinter and that his best chance lay in getting away alone. He could not. Van der Poel matched every move. Yet as the trio of Pogačar, Van der Poel, and Ganna navigated the descent and the coast road into San Remo, Pogačar chose to relay with the Dutchman rather than sitting in and hoping Ganna — who had been dropped and was slowly coming back — would do the same. Analysts noted at the time that sitting might have served Pogačar better tactically, since a fresher Ganna would have split the sprint and reduced Van der Poel’s finishing advantage. He pulled anyway. Van der Poel won.

“I knew Tadej was the strongest on the climbs. When it came down to the three of us, the cooperation was actually quite good, because of course we were all fighting for podium spots.” — Mathieu van der Poel, after Milan-San Remo 2025

Paris-Roubaix 2025. Pogačar’s first attempt at the Hell of the North. The two rode at the front together for the decisive stretch of the race — matching attacks, countering, setting a pace so savage that they shelled Wout van Aert and the rest of the main contenders. By the time they hit sector 9 with 38km to go, only the two of them remained at the head of the race. Then Pogačar misjudged a corner, crashed, and needed a bike change. Van der Poel rode away solo and won his third consecutive Roubaix. Pogačar chased back to second, 78 seconds down. The cooperation was genuine and mutual — neither sat in while the other drove. The fact that it ended with Pogačar on the tarmac was misfortune, not betrayal.

Milan-San Remo 2026. Pogačar crashed before the Cipressa and chased back through the help of teammates. He caught the lead group just in time, and between the Cipressa and the Poggio, the three leaders — Pogačar, Van der Poel, and Pidcock — worked together to keep the peloton at bay. Van der Poel was eventually dropped on the Poggio under Pogačar’s repeated accelerations. Pogačar won.

Across all three races the pattern is consistent: Pogačar races aggressively, does his share of the work, and does not sit in when sitting would be the locally rational move. Van der Poel does the same. What they have established — without ever needing to discuss it — is a mutual reputation for genuine engagement. This is not altruism. It is rational behaviour in a repeated game — exactly what Axelrod’s tournament predicted would emerge.

The Problem with Being Right

Back to Wednesday, and to Vingegaard’s defence. He operates almost exclusively in Grand Tour racing, where the prisoner’s dilemma rarely takes this pure form — breakaways at the Tour de France are usually hopeless for GC contenders, and the relevant game is one of sustained attrition over three weeks, not sprint dynamics. His reputation as a wheel-sucker in Classics and short stage races costs him very little in the context of where he actually wins. The people keeping Axelrod’s ledger are not the ones deciding the Tour de France.

There is also a further complication in Wednesday’s specific case. When Evenepoel crashed inside the final kilometre, Vingegaard — who could, by the logic of pure self-interest, have sprinted away to a stage win — sat up and waited for the peloton.

“I didn’t want to take advantage of a situation like that, so I decided at that moment to just wait for the bunch. I hope he’s okay.” — Jonas Vingegaard, post-stage, Volta a Catalunya Stage 3

This is not the behaviour of a man running a pure defection strategy. The line Vingegaard is drawing is between competitive tactics and sportsmanship — a distinction that the prisoner’s dilemma, being a mathematical framework and not an ethical one, cannot accommodate.

The fans calling him out are not wrong that something has been lost. Cycling’s mythology runs on panache — on riders who pull when they probably shouldn’t, attack when the odds are against them, and create races that are remembered not as optimisation exercises but as spectacles. Fausto Coppi, Eddy Merckx, Marco Pantani: none of them were playing the dominant strategy. They were playing to the gallery, and the gallery was the point.

The fans defending him are not wrong either. He was facing a rider who would have beaten him in a sprint. He read the situation correctly and responded logically. Criticising that is like criticising a chess player for not sacrificing their queen for aesthetic reasons.

So Is It a Good Thing?

Jonas Vingegaard may well be the most rational rider in the peloton. Whether this is a good thing depends entirely on what you think cycling is.

If cycling is a competition to be won by the most efficient allocation of effort, he is right. If it is a repeated game played among humans who remember, he is right in the short term but potentially accumulating a reputational debt that compounds slowly. If it is a spectacle — something people watch because it is beautiful and unpredictable and occasionally absurd — then optimisation is a partial answer to a much larger question.

The prisoner’s dilemma was always about more than prisoners. Its unsettling insight is that rationality at the individual level does not guarantee good outcomes at the collective level. A peloton of Vingegaards — all of them correctly calculating that they should sit in — produces no breakaways, no panache, no Via Roma sprints born from twenty kilometres of genuine mutual effort. It produces a series of bunch sprints decided by lead-out trains, which is its own kind of game and a much less interesting one.

Pogačar pulls when he probably shouldn’t because he has understood — consciously or otherwise — that he is not just playing against the rider next to him. He is playing against his own legacy, against the sport’s appetite for spectacle, against the long-run equilibrium that makes cycling worth watching. He is, in Axelrod’s terms, cooperating on the first move. And over time, cooperation tends to win.

Vingegaard sat in on Wednesday. He was right to. The fact that we are still arguing about it is, perhaps, the most interesting result of the day.