Settings

Language, shortcuts & timezone

Language

Timezone

Time zone changes apply on this device. A new device may start at GMT+8; check match times before reading.

Skip to content
Referrals

Settings

Customize your preferences

How to Read an Accumulator Win Rate Without Mixing Tickets and Legs

By Opta Tips Team Updated

Reconstruct a fictional accumulator ledger to separate ticket wins from leg wins, compare fold counts and keep win rate distinct from financial return.

Two panels show twelve winning leg entries out of sixteen, or 75%, and one winning four-leg ticket out of four, or 25%, in a fictional sample.

An accumulator win rate should count complete tickets under a stated settlement rule. A leg win rate counts individual selections. Both can be correct, and they can tell very different stories about the same record. Before comparing a headline percentage, establish what one observation represents, which tickets have finished, and whether the record mixes doubles with much longer combinations.

This guide builds a small fictional ledger that you can reproduce. It explains how to calculate the two rates, why neither establishes profitability, and how to compare records without rewarding a change in ticket format. None of the examples is an OptaTips performance claim or a forecast of future results.

Write the counting rule before looking at the winners

For a simple starting point, restrict the sample to conventional, fully settled win-or-lose accumulators with cash stakes. Exclude cash outs, refunds, void selections, half settlements and promotions from this first calculation. Those cases need additional labels. Give each submitted ticket a persistent identifier and list its legs separately beneath it.

In this restricted sample, ticket win rate equals winning tickets divided by decided tickets. Leg win rate equals winning leg entries divided by decided leg entries. Multiply either fraction by 100 to express a percentage. A ticket with three winning selections and one losing selection contributes zero winning tickets, three winning leg entries and one losing leg entry.

The distinction follows the contract. Pinnacle’s multiple-bet explanation describes a combined wager that requires the selections to win and calculates its return from the combined prices. Check the terms of the actual operator and product; the figures below assume ordinary independent-fixture accumulators with no special settlement features.

“Decided” is a reporting definition here: the ticket’s final win-or-loss outcome is known. It does not mean every published recommendation was necessarily placed, or that an account could obtain every displayed price. Keep recommendation records and actual betting records separate when those populations differ.

One ledger, two accurate percentages

Assume four fictional four-leg tickets, each staked at 10 units. Every leg has decimal odds of 1.50. All sixteen leg entries refer to different fixtures; no selection is repeated across tickets. The results are:

  • Ticket A: four winning legs; ticket wins.
  • Ticket B: three winning legs and one losing leg; ticket loses.
  • Ticket C: three winning legs and one losing leg; ticket loses.
  • Ticket D: two winning legs and two losing legs; ticket loses.

There are twelve winning legs out of sixteen, so the leg win rate is 75%. There is one winning ticket out of four, so the accumulator win rate is 25%. No arithmetic contradiction exists. The first percentage describes components; the second describes the combined conditions that had to succeed together.

The winning ticket returns 10 × 1.50 × 1.50 × 1.50 × 1.50 = 50.625 units before rounding. Across the four tickets, stakes total 40 units. Gross returns total 50.625 units and net profit is 10.625 units. Profit divided by stakes is 26.5625%. These are fictional, unrounded arithmetic results, with no fees, taxes or promotions. A real ledger should retain the operator’s actual settled amounts.

This example happens to make a profit. Change every leg price to 1.40 while keeping exactly the same outcomes, and the sole winning ticket returns 38.416 units. The same 25% ticket win rate now accompanies a loss of 1.584 units on 40 staked, or −3.96%. Price changes the financial result without changing either count of winners.

A high leg hit rate does not specify the ticket chance

The observed 75% leg result above is a description of sixteen completed entries. It is not automatically a forecast that the next selection has a 75% chance. Even if a well-supported model assigned each leg that probability, multiplying it across a ticket would require an additional dependence assumption.

Under the explicit hypothetical assumption that four legs are independent and each has probability 0.75, their joint win probability is 0.75⁴ = 31.640625%. The general multiplication rule uses conditional probabilities; independence allows the simpler product. OpenStax’s probability chapter explains that distinction.

One win from four tickets is not proof that a 31.64% model is either accurate or inaccurate. A tiny realised sample can vary substantially. Nor should an analyst estimate each probability from this very sample, multiply the estimates and present the result as an independently tested prediction.

For dependence within one ticket, read the separate explanation of accumulator correlation and conditional probability. The present task is defining a record’s units and denominators. Identical labels such as “four-fold” do not establish that either the selections or successive tickets are independent.

Separate ticket formats before pooling them

Suppose a second fictional record contains ten doubles, of which six win, and ten six-folds, of which one wins. The overall rate is seven wins from twenty tickets: 35%. The format-specific rates are 60% for doubles and 10% for six-folds.

If the next reporting period contains ninety doubles and ten six-folds at those same illustrative rates, the combined rate becomes 55 wins from 100 tickets, or 55%. Each format’s rate is unchanged. The rise from 35% to 55% comes entirely from changing the mix.

This is why a combined percentage should sit beside a breakdown by original fold, the number of tickets in each group and the reporting dates. Where settlement reduces a ticket’s active selections, retain both its original and effective fold. Do not quietly move an awkward case between categories to make a comparison look cleaner.

Other useful partitions include market type, pre-match versus in-play publication, and whether the record covers every issued ticket or a selected service. Use partitions that answer a defined question. Splitting a small sample into dozens of attractive subgroups can make the record less informative, especially if the winning subgroup was chosen after seeing the results.

Use the same population for the rate and the money

A ticket count gives every ticket one observation. A financial return calculation weights each ticket through its actual stake and settled return. Those measures should be reported together, but they cannot substitute for each other. Ten losing one-unit tickets and one winning ten-unit ticket do not have the same financial structure as eleven equal stakes.

Keep one column for stake, one for gross return including returned stake, and one for net profit. For an ordinary losing cash ticket, gross return is zero and net profit is minus the stake. For an ordinary winning ticket, net profit is return minus stake. Sum those amounts over exactly the tickets used for the accompanying financial claim.

Do not attach an all-time win rate to last month’s profit or compare a published tip’s price with a follower’s later execution price without explaining the difference. If actual stakes or settled amounts are unavailable, the ticket outcome count may still be verifiable. The financial result is then unknown rather than zero.

The current accumulator page is an entry point for reading the site’s available combinations. This worked example does not establish that a particular site’s display includes all the ledger fields described here, or that access to one product includes any other product. Inspect the current public description and the scope of the specific record.

Keep unfinished tickets visible

Choose whether the report groups tickets by publication date or final settlement date. A publication cohort answers what happened to tickets issued during a defined period; it may still have pending cases after that period ends. A settlement cohort answers which tickets finished during the period; it may contain tickets issued earlier.

For either approach, give the cutoff time and list the pending count. Pending tickets do not belong in a final decided win-rate denominator, but removing their existence from the report conceals how complete the result is. An accumulator with a late fixture should not appear to disappear from a daily report simply because other fixtures have finished.

A practical publication-cohort display might read: twenty submitted tickets, seventeen decided, five winning, three pending; decided win rate 5/17 = 29.41%. It should not present five wins from twenty as the final success rate, and it should not imply that seventeen represents every ticket submitted.

Two further checks deserve separate treatment. When different tickets repeat a selection, the ticket count can exaggerate how many independent outcomes were observed; the worked guide to shared selections across accumulator tickets follows that exposure through a probability tree. When a selection is void, the original ticket and the final combination can have different folds; the guide to void legs and win-rate denominators reconciles those statuses. Neither adjustment justifies changing a published result silently. Their purpose is to make the same record easier to inspect.

A record that another reader can reconstruct

Before relying on an accumulator headline, ask for the ticket IDs, issue dates, original leg lists, exact markets, quoted prices, final statuses and cutoff. For financial claims, add stake and settled return. Retain a correction history when an outcome changes, so that a revised total can be reconciled with the earlier report.

Then reconstruct one ticket and one total. Check that the component results explain the ticket status, that every included ticket appears once, and that numerator and denominator describe the same population. Report the ticket rate, component rate and financial result with their separate definitions. A percentage becomes useful when a reader can follow it back to the observations that produced it.

Reader discussion

Comments 0

No comments yet.

Leave a comment

Your email address will not be published. Required fields are marked *