Define the outcome set
| Team-level market outcome | Meaning under the simplified contract |
|---|---|
| Home scores next | The next qualifying goal is credited to the home team |
| Away scores next | The next qualifying goal is credited to the away team |
| No more goals | No qualifying goal occurs before the market period ends |
The three states must cover the entire settlement space for the simplified team market. A player next-goalscorer product needs additional participation, own-goal, no-scorer and field states under its saved rules; DraftKings' current soccer rules illustrate why those states must be checked product by product.
Worked competing-rate example
Suppose an illustrative remaining-period model uses home goal mean 0.70 and away goal mean 0.50, with independent constant-rate Poisson processes. The combined mean is 1.20.
Using the OpenStax Poisson framework:
P(no more goals) = exp(-1.20) = 0.3010
P(at least one more goal) = 1 - 0.3010 = 0.6990
P(home scores next) = (0.70 / 1.20) * 0.6990 = 0.4078
P(away scores next) = (0.50 / 1.20) * 0.6990 = 0.2913
Check: 0.4078 + 0.2913 + 0.3010 = 1.0001 after rounding.
The rates are invented. The calculation demonstrates a complete distribution; it does not describe an actual match and does not allow the independence or constant-rate assumptions to go untested.
Build team-specific live inputs
Timestamp the clock, score, player count, line-ups, substitutions, recent events and provider status. Sportmonks' livescore documentation illustrates that live fixtures, events, statistics, line-ups and odds are separate includeable entities whose update status must be handled explicitly.
A recent attack is not automatically a raised probability. The forecast method should be specified before the outcome, trained without future events, and evaluated on later matches for each clock and score-state slice.
Check probability coherence
Every listed outcome should have a non-negative probability, and the complete mutually exclusive set should sum to one before rounding under the OpenStax probability framework. Compare team-level next-goal probabilities with the model's total remaining-goals distribution: the probability of either team scoring next must equal the probability of at least one more qualifying goal under the same assumptions.
Product wording controls settlement
Betfair's football rules and DraftKings' soccer rules provide current examples of next-team and goalscorer settlement. Check qualifying period, own goals, abandoned matches, player participation, no-scorer states, official source and later corrections for the accepted product.
Execution checklist
- Save the exact outcome list, including no-more-goals where offered.
- Record the live state and source timestamp.
- Stop while a goal, penalty or VAR incident is unresolved.
- Compare the complete model distribution with a complete market.
- Capture requested and accepted price and size.
- Reconcile event credit and settlement from the saved rules.
The Gambling Commission's in-play standard warns that live information may be delayed. A broadcast goal and market suspension cannot safely be ordered without timestamps.
Next step
Use In Play Over Under for the next part of this topic.
Continue learning
- Next guide: VAR and In-Play Betting
- Related guide: In-Play Corners and Cards
Assumptions and limitations
The worked model assumes independent constant scoring rates for the remaining period and a complete three-state team market. Real rates change with time and match state, and player products contain more states. Rounding, margin, limits and execution are excluded.

