The Blind Test: AI with the odds hidden
In the blind arm, each AI is handed the racecard — the runners, the going, the class, the trip — and asked for one winner, with one thing withheld: the odds. No prices, no market, no favourite pointed out. It has to read the race on the form alone. This is the purest test of the question everyone actually means when they ask “can AI pick horses?”: not can it copy the market, but can it read a race.
Our own two competitors run here too, because this is their natural habitat. The Stablebet Model was built without the market price as an input, and The Favourite — which just backs the market leader — is the baseline every blind tipster has to beat.
Research, not tips. Settled honestly to industry SP, wins and losses alike. 18+ · please gamble responsibly.
The blind leaderboard
1,245 blind picks settled| # | AI | Lab | Bets | Win% | Staked | Returned | Profit | Backed the fav | Profit trend |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Claude | Anthropic | 202 | 21% | £202 | £235 | +£32.60(+16.1%) | 27% | |
| 2 | Grok | xAI | 201 | 21% | £201 | £189 | −£12.37(-6.2%) | 28% | |
| 3 | Gemini | 200 | 21% | £200 | £185 | −£15.09(-7.5%) | 23% | ||
| 4 | DeepSeek | DeepSeek | 200 | 21% | £200 | £176 | −£23.53(-11.8%) | 29% | |
| 5 | ChatGPT | OpenAI | 199 | 22% | £199 | £173 | −£25.72(-12.9%) | 25% | |
| 6 | The Favourite | The market | 120 | 26% | £120 | £103 | −£16.87(-14.1%) | 100% | |
| 7 | Stablebet Model | Stablebet | 123 | 15% | £123 | £79.50 | −£43.50(-35.4%) | 41% |
Running profit, race by race
Each line is one competitor's cumulative profit or loss in the blind arm at £1 a bet, settled to Starting Price. It builds live.
What the numbers say
The read on 7 days and 2,245 settled picks, worked out live from the board.
Every AI finds the winner in about 20.5%–21.6% of races, within a whisker of each other. The market prices those winners too short. Naming the likely winner is not the same as being paid enough for it.
They back the market favourite 23%–29% of the time, and went against a unanimous field only 4–12 times each. Most of the time, asking an AI is close to just backing the favourite.
On return per £1, every AI beats blindly backing the favourite (−14.1p). So the AI adds a little. The full, honest picture is in the standings above.
Return per £1 staked, settled to Starting Price. The Stablebet model and backing the favourite are shown as baselines. Research, not tips.
Showing these same AIs the market made every one of them worse, by an average of £14.60 per £100 staked. They stopped reading the race and started copying the favourite.
Questions
What does the blind test measure?
Whether an AI can read a horse race from the form alone, with no odds to lean on. Each model sees the racecard but not the prices, picks one winner before the off, and we settle it at Starting Price. It isolates genuine race-reading from simply echoing the market.
Why hide the odds from the AI?
Because the market price already contains almost everything knowable about a race. Show an AI the odds and it can score well just by leaning on the favourite, which tells you nothing about its own judgement. Hiding the odds forces it to form a view of its own, which is the interesting thing to measure.
Is the blind AI making money?
Watch the board above — it is derived live, wins and losses alike, and we say plainly where each competitor stands. A positive figure over a few hundred bets is well within the swings of luck, not proof of a system; the sample only starts to mean something after a much longer run. Nothing here is a tip.
How does blind compare to the informed test?
The informed arm shows the same AIs the market's implied chances as well. Comparing the two answers a neat question: does seeing the market make an AI sharper, or does it just pull it onto the favourite? The side-by-side is on the league overview.
Gamble responsibly.This page is research and entertainment, not betting advice. No AI here beats the bookmaker's margin, and nothing on it is a signal to stake. Betting should never be a way to make money. If it is affecting you or someone you know, free and confidential support is at BeGambleAware.org. 18+.