BTTS and Win Accumulators: Model the Joint Score Event

One joint event

For a home-win-and-BTTS selection, qualifying scores include 2-1, 3-1, 3-2 and every retained higher home win in which the away team scores. Scores 1-0, 0-0, 1-1 and 1-2 do not qualify.

P(Home win and BTTS) = sum of P(home goals = h, away goals = a) for h > a and a >= 1

Do not calculate P(Home win) * P(BTTS). A home win changes the relevant distribution of score states, so the two events are not independent.

Product and settlement contract

bet365's related-contingency guidance gives one current example in which related selections may be rejected as an ordinary multiple but accepted in a separately priced builder. DraftKings' market rules provide another product-specific example.

Retain the quoted combined price; do not reconstruct it from standalone win and BTTS prices. Confirm event period, own goals, awarded results, voids and data source using the accepted operator rules. Betfair's football rules illustrate current football-specific definitions.

Worked score-grid comparison

Suppose a complete illustrative score model assigns 0.18 probability to all cells in which the home team wins and both teams score. If the quoted product price is 5.20:

Raw break-even probability = 1 / 5.20 = 0.1923, or about 19.23%

The illustrative model probability of 18.00% is below that raw threshold. This transparent example does not estimate a real match and ignores model uncertainty and limits.

Build a multi-match accumulator

OpenStax's independence condition and the product rules cited above support this sequence:

1. Derive each match's joint combination from its own complete score grid.
2. Validate the joint product on chronological held-out matches.
3. Map shared teams, competitions, lineups, weather and model error across fixtures.
4. Multiply match-level probabilities only where independence is justified.
5. Preserve each separately quoted builder price and the final accepted accumulator price.
6. Enumerate void and repricing states.

OpenStax's independence condition applies between complete match events just as it does between individual legs. Expected value then requires all net payoffs under OpenStax's framework.

Next step

Use Btts Accumulator Tips for the next part of this topic.

Joint-event invariants

Check the home-win-and-BTTS probability against its parent events:

• It cannot exceed P(home win).
• It cannot exceed P(BTTS Yes).
• It excludes every away win, draw and win-to-nil score.
• Home-win-and-BTTS plus home-win-to-nil must equal P(home win), when the score grid fully partitions home wins.

These set relationships follow from OpenStax's treatment of intersections and mutually exclusive events. Use them as unit tests for the score mapper.

For an away-win-and-BTTS product, mirror the grid rather than relabelling the home calculation. Test neutral venues and provider home-away ordering explicitly.

When one same-match component is voided, do not divide the product price by a standalone price. Preserve the operator's revised receipt because the original combination was jointly priced. Compare model probability with the revised product only after the new event contract is known.

Continue learning

• Next guide: Bundesliga Accumulators
• Related guide: Cards Accumulators

Assumptions and limitations

The score-grid value and price are illustrative. Model tails, lineup information, low-score dependence and product repricing can materially change the comparison.