ChatGPT, Claude, Gemini... The AIs of recent years are a constant topic of conversation and are increasingly becoming everyday helpers. Whether it's questions that you would have googled five years ago, holiday inspiration for a trip you haven't yet planned or the simple and popular help for school and university - artificial intelligence provides support. How good is it when you want to predict the outcome of a sporting event? We asked ChatGPT for an assessment of the four remaining cycling classics.
Pogačar defends Tour of Flanders title
According to ChatGPT, there is a clear favourite for the cobbled classic: Tadej Pogačar. With a calculated chance of winning of 60 to 65 per cent, there is hardly any way around the newly crowned winner of Milan-Sanremo according to the AI. With a victory in the Tour of Flanders, he would defend his title from the previous year and celebrate his third triumph in the Belgian classic by winning in 2023.
Pogačar's biggest rival is probably three-time winner Mathieu van der Poel. With a calculated chance of winning of 20 to 25 per cent, the Dutchman is much less likely to win than the Slovenian according to the AI, but van der Poel should not be written off. The decision is likely to be between these two riders, explains ChatGPT.
Favourites' chances of winning
- Tadej Pogačar: 60-65%
- Mathieu van der Poel: 20-25%
- Wout van Aert: 8-10%
- Mads Pedersen: 4-5%
Van der Poel takes fourth victory in a row at Paris-Roubaix
The AI also believes in defending the title in the classic Paris-Roubaix. With a calculated chance of winning of 35 to 40 per cent, the chances of the current best cobbled rider and three-time winner Mathieu van der Poel are slightly better than those of Tadej Pogačar (25 to 30 per cent), who is only missing this monument for his career grand slam.
In general, however, the outcome of the French classic is much more open than the Tour of Flanders. The main reason for this is the notorious "pavés", which harbour a high risk of defects and crashes, meaning that the chance factor plays a greater role. Also, unlike the Tour of Flanders, there are no extreme climbs. The cobblestones decide the winner. Jasper Philipsen, Wout van Aert and Mads Pedersen are therefore also among the favourites, but their chances of winning are no higher than ten percent.
Favourites' chances of winning
- Mathieu van der Poel: 35-40%
- Tadej Pogačar: 25-30%
- Jasper Philipsen: 8-10%
- Wout van Aert: 8-10%
- Mads Pedersen: 5-7%
Pogi dominates Liège-Bastogne-Liège
Unsurprisingly, ChatGPT also rates last year's winner of Liège-Bastogne-Liège as the most likely to win (I'm starting to recognise a pattern). With a calculated probability of 50 to 55 per cent, three-time winner Pogačar will also be crowned for a fourth time at this year's oldest cycling monument. However, the chances of winning are not based on the evaluation of past successes, but on the profile of the race. Liège-Bastogne-Liège has the most metres in altitude of all the classics and is therefore made for climbing specialists.
The chasing pack for the last spring classic of the year is therefore different to that of the Tour of Flanders and Paris-Roubaix. Pogačar's biggest rival and clearest challenger is Remco Evenepoel, who also has more than justified hopes of a third triumph with a 25 to 30 per cent chance of winning. He is one of the few riders who can follow Pogačar on the long climbs.
Favourites' chances of winning
- Tadej Pogačar: 50-55%
- Remco Evenepoel: 25-30%
- Tom Pidcock: 5-8%
- Mattias Skjelmose: 4-6%
- Ben Healy: 3-5%
Tour of Lombardy: Same procedure as every year
Tadej Pogačar has won every edition of the Tour of Lombardy since 2021. ChatGPT also believes in a sixth consecutive title - and calculates a win probability of 65 to 70 per cent for the Slovenian! The route profile with its many long Alpine climbs is perfectly suited to the climbing specialist, which is why only one rider has a halfway realistic chance of an almost surprising victory: Remco Evenepoel. The Belgian is the only rider with a similar profile, which is why his chances of winning are between 15 and 20 per cent.
Favourites' chances of winning
- Tadej Pogačar: 65-70%
- Remco Evenepoel: 15-20%
- Tom Pidcock: 4-6%
- Isaac des Torro / Adam Yates (UAE doubles team): 3-5%
- Primož Roglič / other climbers: 2-4%
Classification of the predictions
According to ChatGPT, the predictions presented here are based on current betting odds, the performance of the riders in the past and other factors such as route profile, teams and tactics. Especially with regard to the Tour of Lombardy, which takes place late in the year, the probabilities can still change quite a bit.
It is also a fact that there are still probabilities, which should not encourage sports betting and cannot predict the outcome of the races. When asked who the favourites were for Milan-Sanremo, Mathieu van der Poel was given the highest probability of victory (30 to 35 per cent). Pogačar was also among the favourites with a probability of victory of 25 to 30 percent, but Tom Pidcock, who came second by a few centimetres, only had a probability of victory of three to five percent.
So it remains to be seen to what extent the predictions presented here for the outcomes of the remaining four classics will be correct. The fact that Paris-Roubaix is the only race in which Pogačar is not the biggest favourite according to the AI should not stop the Slovenian from doing everything in his power to win the last missing Monument and join the elite circle around Eddy Merckx, Roger De Vlaeminck and Rik Van Looy with the Classics Grand Slam.
Lukas Niebuhr
Working student
