Claude, ChatGPT, and Gemini all leaned on G2 Esports and Karmine Corp for LEC Week 1, but they treated the weekend very differently: Claude built a moderate four-leg accumulator, ChatGPT played it safest, Gemini chased handicap value, and the human expert went full underdog on a five-leg long shot.
The AI picks were not wildly different in their team hierarchy, but they were very different in confidence. Claude and Gemini accepted some volatility for a return that felt worth chasing. ChatGPT clustered its faith around G2 Esports. Craig Robinson, representing the human side, bet as if the odds were quietly wrong.
This is a prediction exercise built around a fictional $10 stake, not betting advice. The numbers matter because they reveal risk appetite, and in esports betting that is almost as interesting as the actual picks.
The Odds Snapshot Shows Three AI Models Playing It Safer
Craig Robinson’s five-leg accumulator carried the highest combined odds at 19.72, which would turn a $10 stake into $197.23 if every leg landed. That is the kind of bet that looks ridiculous before the first match and legendary if it pays off.
Gemini’s three-leg bet was next, at 6.88 combined odds, for a potential return of $68.76. Claude’s four-leg bet sat at 5.36, returning $53.60. ChatGPT was the most conservative, with three legs at 2.29 combined odds, for a potential return of $22.87.
The key distinction is simple: the AI models mostly avoided Craig’s boldest underdog angles. They were not afraid to back Karmine Corp or G2, but they were more willing to let the odds market do some heavy lifting. Craig’s approach was a value argument, not a confidence argument.
The Experiment Was Built Around One Plain Prompt
The LEC returns with the 2026 Summer Split in Paris, and this week’s setup was designed to be readable: give each predictor the same weekend fixtures, the same odds snapshot, and a fictional $10 bankroll. Then ask who builds the better accumulator.
“I am giving you a fictional $10 to place a bet on this weekend’s LEC games. You can build an accumulator, or pick an outright. You can determine your risk profile. You will be competing against a human, plus two other AI chat bots.”
That prompt is deliberately loose, which makes the results more revealing. Claude, ChatGPT, and Gemini each had to decide what “value” meant. One model interpreted value as a safer G2-heavy path. Another saw enough upside in Karmine Corp and G2 to chase better returns. The human interpreted value as “the public is overpricing G2.”
That is the interesting part. These are not just predictions; they are models of confidence. The AI arms race is increasingly about turning models into analysts, advisors, and assistants, but this kind of experiment shows how much personality still leaks through when the task leaves room for judgment.
Claude Backed G2 and Karmine Corp as the Weekend’s Core
Claude built a four-leg accumulator worth 5.36 combined odds. Its legs were: G2 Esports to beat Movistar KOI at 2/5, Karmine Corp to beat Team Vitality at 1/4, G2 Esports to beat Team Vitality at 1/6, and Karmine Corp -1.5 handicap against Movistar KOI at 13/8.
Claude’s reasoning was the most conventional. It treated G2 as the team to beat after winning both LEC Versus and the Spring Split, while also leaning on Karmine Corp as the strongest challenger. The model’s most aggressive leg was Karmine Corp winning cleanly against Movistar KOI rather than simply winning the match, which is how it boosted the accumulator without going fully rogue.
Claude also made a sensible scheduling call by skipping the Sunday KOI vs Vitality and G2 vs Karmine Corp matches. By then, those teams would have already played twice, meaning form could shift and the original odds could feel stale. That is a small detail, but it is exactly the sort of thing that separates a cautious accumulator from a reckless one.
ChatGPT Went With G2 in Every Leg
ChatGPT chose the safest route of the three AI models. Its three legs were all G2 Esports wins: G2 vs Movistar KOI at 2/5, Team Vitality vs G2 at 1/6, and G2 vs Karmine Corp at 2/5. The combined odds were 2.29, producing a potential $22.87 return on the fictional $10 stake.
That is a medium-risk accumulator in name only. It is basically a G2 confidence bet wrapped in accumulator language. The model’s logic was that G2 had the strongest recent pedigree, including Spring Split title wins and a clean EWC qualifier sweep against Karmine Corp.
The obvious weakness is the first leg against Movistar KOI. KOI pushed G2 to five games in Spring, so Friday is the match most likely to ruin the bet before the weekend has even properly started. ChatGPT’s pick is tidy, but it asks G2 to be dominant across the entire weekend, including against a team that has already proven it can make trouble.
Gemini Chased Value Against Team Vitality
Gemini made the most interesting AI pick. Its three legs were Karmine Corp -1.5 handicap vs Team Vitality at 1/1, G2 Esports -1.5 handicap vs Team Vitality at 4 Key Numbers
3 metrics
Comments (0)
Loading comments…