Identify the result contract
DraftKings' soccer guide gives current operator examples of soccer result, total and handicap markets. The accepted ticket should still be read literally.
| Product | Winning state for Team A | Draw treatment | Extra time or shoot-out |
|---|---|---|---|
| Three-way 1X2 | Team A wins in the stated period | Loses | Usually excluded from regulation-time result |
| Two-way winner | Team A wins under stated scope | Product-specific | May include later stages if stated |
| To advance | Team A qualifies from the tie | Intermediate result only | Included as required by competition |
| Draw no bet | Team A wins in regulation | Stake refunded | Excluded |
| Double chance | One of two named 1X2 outcomes | Can be included | Excluded unless stated otherwise |
DraftKings' soccer rules illustrate why normal time, result scope and tournament settlement must be checked separately.
Preserve the full outcome vector
For a three-way match, record probabilities for home, draw and away so they sum to one under one model version. A selected team's probability alone cannot reveal whether draw risk or the opponent has been omitted.
| Outcome | Illustrative probability |
|---|---|
| Home | 0.52 |
| Draw | 0.27 |
| Away | 0.21 |
| Total | 1.00 |
The table is hypothetical. OpenStax defines the sample space and probabilities of outcomes; real estimates need data, validation and uncertainty.
Worked three-match parlay
Suppose three cross-match result legs are accepted at decimal 1.80, 1.95 and 2.10.
Combined price = 1.80 * 1.95 * 2.10 = 7.371
Gross return on $10 = 10 * 7.371 = $73.71
Net profit if all three win = 73.71 - 10 = $63.71
If a model assigns illustrative independent probabilities of 0.58, 0.53 and 0.48:
Joint probability = 0.58 * 0.53 * 0.48 = 0.1476, or 14.76%
Raw break-even probability = 1 / 7.371 = 0.1357, or 13.57%
OpenStax's multiplication rule applies only if independence is justified. The illustrative difference is not evidence of value because model error, margin, common causes and settlement differences are omitted.
Dependence across matches
Separate fixtures are not automatically independent. They can share a team-state, tournament incentive, weather system, schedule disruption, data provider or model bias. Record each shared cause and either estimate conditional scenarios or widen uncertainty.
Validate probabilities, not picks
A result model should be tested on later matches using the same inclusion, cutoff and market definition rules. scikit-learn's calibration guide explains reliability curves for probability forecasts. Report all eligible forecasts, not only selected parlay legs.
Draw-risk check
Before accepting a result leg:
- Confirm three-way, two-way, draw-no-bet or advancement scope.
- Record the draw probability even when the selected product removes it.
- Verify normal time, extra time and shoot-out treatment.
- Capture the accepted signed American price and full-precision quote.
- Compare the selected product with the same match under alternative result contracts.
Expected-value boundary
Use the complete joint probability and net payoff states under OpenStax's expected-value framework. A favourite label, unbeaten run or low American price does not establish expected value.
Next step
Use 1x2 Betting Explained for the next part of this topic.
Continue learning
- Next guide: Soccer Parlay Bankroll
- Related guide: Soccer Parlay Mistakes
Assumptions and limitations
The worked match probabilities and prices are hypothetical. The method does not publish current teams or picks. Result naming, draw treatment, limits, rules and competition formats vary by sportsbook, state and date.
GambleAware
GamCare
Gambling Therapy
GAMSTOP