How we make money by knowing when NOT to bet — backtested on 943 real UFC fights since 2024-02-24.
Most bettors lose money because they bet on every fight. Our system is profitable not because it predicts every fight correctly — it's profitable because it knows which fights to skip.
We backtested three different approaches across 943 real UFC fights since 2024-02-24, using real Vegas odds. Each strategy bets a flat $100 per fight. The difference? Which fights you choose to bet on.
Every prediction gets classified into one of four signals based on model confidence, the size of the edge vs Vegas, and whether the model disagrees with the favorite.
The highest-conviction tier. Many of these are underdog calls where the model sees value Vegas missed — they win less often than favorites, but the odds more than compensate over a large sample.
Solid model conviction without the full value-spot setup. Playable on their own or as parlay anchors — sized smaller than Strong Bets.
These often agree with Vegas but at a slightly better implied price, or carry a reliability caveat (a fighter with little UFC data, or a strong disagreement with a heavy favorite). Smaller positions.
This is where the magic happens. Betting fights the model can't separate is how bankrolls die. The system's ability to say "don't bet" is what keeps the overall strategy profitable.
Out of 943 backtested fights, the system flagged 94 as Skip — fights where the model had no clear edge. Bettors who ignored this signal and bet anyway would have lost $1,982.
By not betting on those 94 fights, the remaining 849 actionable bets produced +$9,511 in profit at a +11.2% ROI.
Skipping the skips locks in the win and avoids the loss. The aggregate live numbers are on our track record page.
Every prediction includes a bet signal. Filter by Strong Bet, Bet, Lean, or Skip.