Goals and BTTS: what the markets mean
Total goals and both teams to score are related to scoring, but they ask different questions and should not be treated as interchangeable.
Total-goals lines
A total-goals selection compares the combined goals scored by both teams with a stated line. “Over 2.5” requires at least three total goals; “Over 3.5” requires at least four. The half-goal line prevents a tie at the threshold.
The winning team does not determine this market. A 3–0 and a 2–1 score both finish over 2.5, even though only one of those scores has goals from both teams.
Both teams to score
BTTS “Yes” asks whether each team scores at least once. BTTS “No” covers any result where one or both teams fail to score. A 1–1 result is BTTS Yes while remaining under 2.5 total goals; a 4–0 result is over 2.5 while remaining BTTS No.
Those examples show why one estimate cannot be used as a substitute for the other. Each market has its own published probability only when that market is available in the fixture snapshot.
Full match and first half are separate
A first-half goals line applies only to goals before halftime. It is not a shortened version of a full-match prediction and cannot be inferred safely from the full-time total. DeePicks exposes first-half estimates only where the corresponding model section and line are available.
Worked example: the same score answers two questions
Hypothetical teaching example. These invented scorelines illustrate market definitions. They are not DeePicks forecasts, observed results or recommended bets. Assume the stated score is the final score for the period being assessed.
| Score | Total goals | Over 2.5 | BTTS Yes |
|---|---|---|---|
| 0–0 | 0 | No | No |
| 1–1 | 2 | No | Yes |
| 3–0 | 3 | Yes | No |
| 2–1 | 3 | Yes | Yes |
Work through 3–0: adding the scores gives three, which exceeds 2.5, but the away side scored zero. Work through 1–1: both sides scored, yet the total is only two. Neither market logically implies the other.
Worked example: two percentages do not give a joint probability
Hypothetical teaching example. Imagine a model assigns Over 2.5 a 60% estimate and BTTS Yes a 45% estimate. These are invented inputs, not current DeePicks outputs. Multiplying them gives 27%, but that calculation assumes independence, which the two percentages do not establish.
Even if both illustrative inputs were correct, the probability of both happening could lie between 5% and 45% without more information about their overlap: the lower bound is 60% + 45% − 100%, and the upper bound is the smaller input. That wide range is why separate percentages cannot justify a precise combined selection. When BTTS is unavailable, there is not even a BTTS input to combine.
Definitions, periods and interpretation errors
IFAB Law 10 defines when a goal is scored and separates the match result from tie-breaking procedures. It does not set a bookmaker's settlement terms. Check the named market's period, added-time treatment, extra-time exclusions, abandonment rules and correction policy before interpreting a wager.
Do not mix a halftime score with a full-match line, count penalty-shootout attempts as ordinary match goals, or read a goals estimate as a BTTS estimate. A high total-goals probability can coexist with BTTS No. Model probability is also distinct from an expected goal count and from a bookmaker price.
The DeePicks methodology explains why full-time goals and BTTS have different history gates. The engine may publish a goals estimate from a small sample while withholding BTTS. That is an availability distinction, not proof that the goals estimate is reliable.
Limits of the estimate
Scoring models depend on the historical and fixture inputs available when the snapshot is created. Sparse history, missing source fields, lineup changes, red cards, and other match events can make the realised result differ sharply from the estimate. Related markets can also be statistically dependent, so combining percentages as though they were independent can misstate uncertainty.
A goals or BTTS percentage is not a guarantee and does not establish that a bookmaker price is favourable.