Probability of profit is the market-implied chance that a position finishes above breakeven at expiration, computed from the option chain's own prices. It is a model-based estimate under the market's implied distribution, not a promise, and it says nothing about how large wins or losses are.
For a lognormal model, the probability that the stock finishes beyond a breakeven level is N(d2) evaluated at that level: the same term that prices the option, reused as a probability. Three implementation details separate a careful POP from a sloppy one:
Magnitude. A 71% POP credit spread loses more when wrong than it gains when right; the probability and the payoff have to be read together, which is why credit-to-width floors sit next to POP in an entry discipline. POP is also an at-expiration number: a position can spend weeks tested and underwater and still finish profitable, or get stopped by an exit rule before its POP resolves.
Only as far as it is audited. The honest test is calibration: bucket closed positions by their entry POP and compare each bucket's realized outcome rate against the forecast, with confidence ranges. For premium selling, winning slightly more often than forecast is the expected signature of the volatility risk premium; winning less often than forecast, beyond the confidence range, is the side that matters. A recorded held-to-expiry outcome for every closed trade, even ones exited early, is what makes that comparison honest, and Options Scanner runs exactly that comparison as a monitored gauge rather than a one-time study.
Because the options market prices that probability into the premium. A high-probability spread pays a small credit against a large width, so frequent small wins and rare full-width losses can sum to roughly zero before costs. Edge, where it exists, comes from the premium being priced above the movement that follows, not from the win rate itself.
The common closed-form approach models the underlying as log-normal at expiration, using implied volatility and time remaining, and reads off the probability of finishing beyond the breakeven. Platforms differ in the drift assumption and in whether they measure at the breakeven or the short strike, so two POP numbers for the same trade rarely match exactly.
A short strike's delta approximates the probability of expiring in the money, so one minus the delta is a rough probability of the strike holding. It is a useful mental shortcut that drifts from the model-based number as skew, time, and distance from the money grow.