## When Bets Are Not Independent
Standard EV calculation assumes each bet is independent. In practice, your betting positions are sometimes correlated — the outcome of one affects the probability of another.
## Direct Correlation: Same Event, Multiple Markets
If you bet Team A to win AND over 2.5 goals in the same match, these bets are correlated. Team A winning at 2-1 satisfies both; Team A winning at 1-0 satisfies only the first. Your EV calculations need to account for this correlation.
True joint EV ≠ EV(win) + EV(over). The actual joint EV requires knowing the joint probability distribution of both outcomes.
## Indirect Correlation: Same League, Same Season
In a league outright market, backing multiple teams to win the league creates portfolio correlation. If your model says Team A is 30% likely and Team B 25% likely (50% combined) but the true probability of "neither wins" is 55%, there is an inconsistency — the probabilities are competing.
## Hedging EV: When to Trade Out
If you backed Team A at 3.00 and they are in-play at 1.60, your current position has positive EV relative to the current price. Whether to hold or hedge depends on:
Hold EV: (Original P × 3.00) − 1 (ongoing from here)
Hedge EV: Lock profit now (certain)
The hedge is correct if the remaining uncertainty (risk of A not winning) plus the commission of the hedge exceeds the EV of holding. In most cases with large in-play price compression, the hedge is correct once sufficient profit is locked.
## Portfolio EV with Correlations
For a portfolio of correlated bets, total portfolio EV is still the sum of individual EVs (EV is additive). But the variance of the portfolio is not simply additive — it grows with correlation. Highly correlated positions concentrate both upside and downside, increasing the probability of extreme outcomes in either direction.
Create a free account to track your progress and save bookmarks.
