Probability of Touch vs Probability of ITM

Side-by-Side

What Each One Measures

Take a 30-day SPY 525 call when SPY is trading at 510. Two distinct probability questions:

• POT(525): what is the probability that SPY touches 525 at any point during the next 30 trading days? This is a path-dependent question. The underlying may touch 525 on day 5, then drop below, finish at 510 on day 30, and POT is satisfied.
• POT-ITM(525): what is the probability that SPY closes above 525 on the expiration day? This is a terminal-only question. The path the underlying takes during the 30 days does not matter; only the final closing print does.

The two are related but distinct. For OTM strikes, POT is consistently larger than POT-ITM. The intuition: any path that ends ITM also touched the strike at least once, so {ITM at T} is a subset of {touch K during (0, T)}. The probability of touching is therefore at least the probability of ending ITM, and typically much larger because many paths touch and revert.

The Reflection-Principle Approximation

For driftless Brownian motion, the reflection principle gives the exact relationship: POT(K) = 2 · POT-ITM(K) for K above spot. The result follows from the symmetry of Brownian paths: for every path that touches K and ends ITM, there is a mirror path that touches K and ends below K, and those paths have equal probability.

For real markets with drift (interest rates, dividends), the relationship is approximate: POT(K) is approximately 2 · POT-ITM(K), with a small correction term involving drift. For typical 30-day equity options at moderate volatility, the 2x approximation holds within 5%.

This is operationally useful: if a broker reports POT-ITM (delta is ~POT-ITM under Black-Scholes), you can quickly estimate POT as 2 · delta. A 25-delta call has approximately 50% probability of being touched during its life, even though only 25% probability of finishing ITM.

Where They Agree

For deep ITM strikes, POT and POT-ITM converge: if the underlying is already past the strike, both probabilities approach 1. The 2x-relationship holds for OTM strikes; it does not hold near the strike. For an ATM strike (spot = K), POT is essentially 1 (the strike is touched immediately) while POT-ITM remains near 0.50, so the two metrics diverge most sharply at the money.

Both probabilities are computed under the risk-neutral measure when extracted from option prices. Both adjust for spot drift, dividend yield, and the same volatility input. Both can be model-extended (Heston, jump-diffusion, local vol) for non-Black-Scholes pricing.

Where They Diverge

• Stop-loss design. If you place a stop at strike K on a long position, POT(K) is the probability the stop fires during the position's life. POT-ITM is irrelevant for stop placement because the path matters, not just the terminal value. Using POT-ITM for stop probability systematically underestimates stop-out risk by ~50%.
• Touch-and-go barriers. Knock-in and knock-out barrier options price directly off POT-style probabilities. POT-ITM cannot price these contracts because barriers are path-dependent.
• Premium-collection sizing. A 25-delta short put has POT-ITM of ~25% but POT of ~50%. The "probability of having to manage the position" is closer to POT than POT-ITM. Many premium-sellers under-estimate the management frequency by anchoring on delta.
• Dealer-hedging analysis. Dealers care about spot moves through the option's life, not just terminal value. POT is more predictive of dealer-hedging-flow severity than POT-ITM.

Worked Example

SPY at 510, 30-day expiration, 525 call. Volatility 14.5%, rate 4.5%, dividend yield 1.3%:

• Delta = 0.27 (BSM, with continuous-dividend discount)
• POT-ITM at 525 = 0.25 (terminal probability of finishing above 525, equal to N(d2))
• POT(525) = 0.51 (path probability of touching 525 at any point)
• 2 · POT-ITM = 0.51, almost exactly equal to the path-touch probability for this OTM strike (drift correction is small at 30 days)

The retail-trader heuristic "delta = probability of ITM" is approximately true (delta = 0.27 vs POT-ITM = 0.25). But the trader who interprets "probability the call gets touched" as 27% is off by roughly half - the actual touch probability is ~51%.

Why They're Often Confused

Retail platforms and broker tools often report delta and call it "probability." This is approximately true for POT-ITM but is widely misread as "the probability anything interesting happens to this option." For OTM short premium positions, the relevant probability for risk management is POT, not POT-ITM, and the gap is approximately 2x.

The confusion is amplified because option premium is computed based on terminal-value probabilities (POT-ITM), so the "fair value" intuition aligns with POT-ITM. But position management (when do I hedge? when do I exit?) is path-dependent, so POT is the more relevant metric.

Further Reading

• Probability Analysis: POT, POT-ITM, and Risk-Neutral Density
• Expected Move
• Risk-Neutral Density
• Black-Scholes Model
• Pricing Model Landscape

References

• Hull, J. C. (2022). Options, Futures, and Other Derivatives, 11th ed. Pearson. Standard reference for option probabilities and barrier-option pricing.
• Karatzas, I. and Shreve, S. E. (1991). Brownian Motion and Stochastic Calculus. Springer. Standard stochastic-calculus reference for the reflection principle.
• Glasserman, P. (2003). Monte Carlo Methods in Financial Engineering. Springer. Standard reference for numerical first-passage simulation.
• Sinclair, E. (2013). Volatility Trading, 2nd ed. Wiley. Practitioner-oriented treatment of POT-vs-POT-ITM in position management.

This is one of the model-vs-model comparison pages. For the full landscape of pricing models and their relationships, see the Pricing Model Landscape.