Combine three bets into an accumulator and the odds multiply cleanly. Three prices at 1.80, 1.90 and 2.00 become a single 6.84. Your edge does not multiply in quite the same way. It compounds, and compounding is much less forgiving when one leg is weaker than the others.
The formula
Combined odds are the product of each leg's decimal price. Combined implied probability is the product of each leg's implied probability:
combined odds = odds₁ × odds₂ × odds₃ … combined probability = (1 ÷ odds₁) × (1 ÷ odds₂) × (1 ÷ odds₃) …
At 1.80, 1.90 and 2.00, the combined price is 6.84 and the combined market-implied probability is 14.62%. The individual implied probabilities are 55.6%, 52.6% and 50.0%, which are multiplied together to produce the combined figure.
Where the edge actually comes from
Expected value on an accumulator depends on your own combined probability estimate compared with the combined price, just as it does for a single bet.
Suppose you rate the three legs at 58%, 55% and 50%. The first two give you a genuine edge over the market's 55.6% and 52.6%, while the third is priced fairly at exactly 50%.
- Your combined true probability — 58% × 55% × 50% = 15.95%
- Combined odds — 6.84
- Expected value — 15.95% × 6.84 − 1 = +9.1%
That is a real positive edge. It is smaller than either individual edge in percentage-point terms because the third leg contributes no additional value, but it does not hurt the combination either because it is priced fairly.
One weak leg is enough to erase it
Now change only the third leg. Instead of a fair 50% probability, suppose you believe its true probability is 45%.
- Your combined true probability — 58% × 55% × 45% = 14.36%
- Combined odds — 6.84 (unchanged)
- Expected value — 14.36% × 6.84 − 1 = −1.8%
Two legs have a genuine edge, while the third has a modest negative edge. The entire accumulator has now moved from profitable to unprofitable.
That is the important difference between adding and multiplying probabilities. A weak leg does not simply remove its own small share of the value. Its probability is multiplied through every other leg in the accumulator, so its shortfall affects the entire combination.
Combined probability decays faster than intuition expects
Even four legs that each have a genuine edge can produce a surprisingly low overall hit rate.
Suppose each leg has a true probability of 55% against a fair 50% market price:
0.55 × 0.55 × 0.55 × 0.55 ≈ 9.2%
A four-leg accumulator made entirely from bets where you genuinely have an edge still loses roughly nine times out of ten.
That is not a problem with the bets or with the mathematics. It is simply what happens when several probabilities below 100% are multiplied together.
The result is that an accumulator's win rate tells you surprisingly little about whether it was a good decision. A long run of losing accumulators can come from genuine positive-EV selections just as easily as from selections with no edge at all.
Same-game legs are rarely independent
Everything above assumes that the probability of each leg is unaffected by the others. That is a reasonable approximation when the legs come from different matches, but it can be a poor assumption when they come from the same game.
For example, "this team wins" and "over 2.5 total goals" are not independent events. A team winning comfortably often goes together with more goals being scored, so the two outcomes are positively correlated.
Simply multiplying their standalone probabilities can therefore give the wrong combined probability. This is one reason same-game parlays are normally priced with a correlation adjustment rather than using straightforward multiplication.
Treat a same-game combination with more caution than a combination built from unrelated matches. The advertised price may already reflect part of the relationship between the selections, so an apparent value gap deserves closer scrutiny.
Sizing a bet with a low hit rate
A four-leg accumulator built entirely from genuine edges can still lose most of the time.
That makes flat staking on accumulators different from flat staking on singles. A stake that feels comfortable on a bet with roughly a 50% chance of winning can produce much larger swings when the same stake is placed on something that wins only around one time in ten.
The same principle used for single bets still applies: as the probability of winning falls, stake sizing becomes increasingly important because each result has a larger effect on the bankroll.
What this means for building one
- Every leg needs to clear the market on its own. A strong first two legs do not compensate for a weak third - the third leg's own probability multiplies through the other two, dragging the whole combination down by exactly its shortfall.
- More legs is not automatically worse, but it is automatically less frequent. Longer accumulators built entirely from real edges are still correct decisions in expected-value terms; they simply need a much longer run of attempts before the result looks anything like the expected value, for the reason above.
- Check the combined implied probability before staking, not after. It tells you upfront how demanding the combination actually is, in the same units as any single bet's implied probability - no different math, just multiplied through one more time per added leg.
None of this makes accumulators inherently worse than singles. A combination of genuine edges is still a positive expected-value proposition.
The difference is that the gap between "this was a good decision" and "this paid off" remains wider for longer. The win rate is lower by construction, so judging an accumulator strategy by whether last week's tickets landed tells you very little.
A better question is whether each individual leg had value when it was placed. That is the part of the decision you could actually assess in advance.
Run your own legs through the accumulator calculator to see the combined price and combined implied probability side by side before you commit to any of them.