Compare future cash states
OpenStax's expected-value framework compares future probability-weighted payoffs. When a live offer appears, the immediate decision is between cash available now and the future cash distribution from continuing, including any void, partial, or promotional states.
Betfair's current Cash Out help is one operator example: an offer is not guaranteed to be available or accepted, and cash out can affect accumulator features. Archive the offered amount, time, leg states and product rules.
Worked conditional comparison
Suppose an illustrative accumulator can return GBP 90 gross if its only remaining live state succeeds. A current model assigns that state probability 0.40, and all other modelled states return zero.
Conditional expected cash return from holding = 0.40 * 90 = GBP 36
Under the same expected-value calculation, an accepted GBP 32 cash-out offer is GBP 4 below the modelled hold return before uncertainty and other payoff states. If the illustrative probability were 0.30, expected hold cash would be GBP 27 and the comparison would reverse.
| Input | Value |
|---|---|
| Final gross return | GBP 90 |
| Conditional success probability | 0.40 |
| Model expected hold cash | GBP 36 |
| Accepted cash-out cash | GBP 32 |
The values are illustrative, not a recommendation. The quality of the decision depends on the conditional probability and complete contract.
Add missing states
Include a voided final leg, partial settlement, dead heat, non-participant, abandonment, result correction and promotion eligibility wherever the product permits them. Betfair's general Sportsbook rules demonstrate why multiples and cash-out features can have product-specific conditions.
Then run low, central and high probability estimates. If the choice changes across a plausible range, record the decision as uncertainty-sensitive. Expected value is probability-weighted payoff under OpenStax's framework; it is not a promise about one result.
Decision checklist
- Confirm the offer is accepted, not merely displayed.
- Freeze all remaining event and market definitions.
- Estimate conditional probabilities from current information.
- Map every cash payoff and promotion effect.
- Compare cash out, hold and any executable hedge.
- Save the decision before the outcome.
Do not cash out or continue solely to recover a prior loss. If gambling decisions are causing distress or financial harm, stop and use the NHS support routes.
Next step
Use Hedging Accumulators for the next part of this topic.
Derive the binary break-even probability
For a simplified state with final gross return R on success, zero cash on failure, and immediate cash-out offer C:
Hold and cash out have equal modelled cash value when p * R = C.
Break-even conditional probability p = C / R.
Using the illustrative GBP 32 offer and GBP 90 final return:
p = 32 / 90 = 0.3556, or about 35.56%
Under OpenStax's probability-weighted payoff method, holding has higher modelled cash in this two-state example when conditional success probability exceeds 35.56%; cash out has higher immediate cash value when it is lower. Add any non-zero failure return, void state, commission or promotion before using the threshold.
Use the same expected-value calculation across a sensitivity table and save the probability source. An estimate close to 35.56% is less robust to probability error than one far from the threshold, even when both lie on the same side.
Continue learning
- Next guide: Why Accumulators Lose
- Related guide: Acca Builder Tools Compared
Assumptions and limitations
The worked example has one binary remaining event and no commission or tax. Real offers can change rapidly, execution can fail, and conditional probabilities are uncertain.

