On the BettingPros Prop Bet Cheat Sheet, select everything (Cmd+A) and copy. The book names have to come through or the fees can't be charged, so grab the whole page rather than just the rows.
MLB and NFL only for now. Manual only — it never fires on its own, so it never spends credits without you clicking. Leave it off most days and switch it on for the one day a week you want it (Saturday or Sunday, your call). Your key stays in this browser only.
A bet has to clear all of these. Widening them until something appears is how you end up on the rows that only look good.
Kalshi runs on a curve rather than a flat rate: 7% × price × (1 − price) per contract, which is 7% × (1 − price) of what you put up. A short favorite costs about 2.3%, an even-money bet 3.5%, a +400 longshot 5.6%. The fee is charged on top of the contract cost, so net EV is (gross − fee) ÷ (1 + fee). Kalshi also rounds each order's fee up to the next cent, worth up to 0.4% on a $2.50 ticket and not modelled here.
Kelly scales a bet to its edge. The haircut shrinks that edge first, because the fair values behind it move and the biggest ones move most. What survives gets clamped to the range you actually bet in.
| Team / Player | Sport | Market | Odds | Fair Value | Market EV | Bet Now | Fee | Net | Stake | Notes |
|---|
For a single pick from anywhere. Enter the other side's fair value too if you can get it — that is what exposes a bad devig.
Log a bet from the board, then come back after the game starts and enter where the line closed. Closing line value is the only thing that tells you whether an edge was real before you have hundreds of settled bets.
| Date | Bet | Sport | Your price | Fair then | Net EV | Stake | Result | Closed at | CLV |
|---|
Each row is only meaningful once it holds a few dozen bets. Below that, treat it as noise you are watching rather than a result.
| Split | Bets | Beat the close | Mean CLV | Mean net EV |
|---|
CLV here compares your price to the closing price on the same side, in implied-probability terms. It is not devigged, so a positive number is evidence rather than proof — but a screen that consistently beats the close is finding real prices, and one that does not is finding noise no matter what the EV column said.
Expect a few hundred bets before any of this separates from variance.