Every bet has an expected value — the average amount you'd win or lose per bet if you could make the exact same bet thousands of times. It's the single number that tells you whether a bet is worth making. Positive EV: bet. Negative EV: don't, no matter how likely it looks.
The calculation
EV = (your probability of winning × the profit if you win) − (your probability of losing × the stake). Say you back an outcome at decimal odds of 3.00 (a $2 profit on a $1 stake) and you rate it 40% likely: EV = (0.40 × $2) − (0.60 × $1) = $0.80 − $0.60 = +$0.20 per $1. Positive — a bet worth making, even though it loses 60% of the time.
Why EV beats win rate
A 60%-losing bet can be highly profitable, and a 90%-winning bet can be a slow leak — it all depends on the price. Chasing win rate pushes you onto short-priced favourites with no value left; chasing EV puts you on whatever the market has mispriced, which is often the underdog. Winning feels good; positive EV is what actually compounds.
Where the probability comes from
EV is only as good as your probability estimate. That's the hard part — and the whole reason to build a model, de-vig a sharp line, and check your calibration. Get an honest probability, compare it to the price, and bet only when EV is positive past a sane threshold.
That's the Momus loop end to end. See it on the track record, or read how Momus finds value on Polymarket.