Find the gap between books.

Sportsbook A prices the Lakers at +120; Sportsbook B prices the Celtics — the other side of the same game — at +105. Both can't be right: the two implied odds add up to less than 100%, so the books disagree in your favor. Stake $100 on the Lakers (+120 returns $220) and $107 on the Celtics (+105 returns ~$220). You've risked $207 and collect ~$220 whoever wins— a locked ~$13 profit on either outcome. That's the point of an arb: profit on both sides, not one.

That's arbitrage. The catch: arb windows close in seconds. The books re-price as soon as the disagreement is spotted by the market. Players who only ever place arb bets get limited fast on the aggressive books — the books know arb players don't lose, and they don't want to take that action.

Whet checks for arb windows when you open the Arbitrage tab, over the newest completed batch of odds from the last ten minutes, and re-checks every 30 seconds while the tab is open. For each window it finds, it computes the optimal stake split across the two books. It reads player props only, so game lines are not covered.

Arbitrage doesn't mean “risk-free.” Limits drop. Books void bets when they think a line was mispriced. Bonus funds don't always count toward the lock. Two markets can be correlated in ways the calculator doesn't see (player props with shared rosters). Real arb in 2026 is a managed-risk strategy, not a printing press.

Whet does not track any of that for you. There is no per-book limit history, no void-rate signal, no flag detection. The scanner prices the window; managing your accounts is on you.