Why Over 2.5 and Over 3.5 in One Game Aren't Two Bets
Bets on the same match usually win or lose together. How that inflated our own backtest's bet count and luck band, and how to count honestly.
Last reviewed November 3, 2026
Key takeaways
- >Over 2.5 and Over 3.5 only split when a match has exactly 3 goals
- >Our Eredivisie backtest: 122 bets on 95 matches; in 23 of the 27 matches with several bets, all had the same result
- >Counted one bet per match the luck band widens from ±17.5% to ±20.9%
- >Correlated bets make streaks and drawdowns bigger than the bet count suggests
- >Always state the number of matches next to the number of bets
If you bet Over 2.5 goals and Over 3.5 goals on the same match, you have two bet slips but not two bets. Both depend on the same thing, the number of goals, and in most matches they win or lose together. That sounds obvious, but it quietly inflates almost every backtest and betting record you will ever see, including ours.
The problem in one match
Say a match ends 1-0. Over 2.5 loses, Over 3.5 loses. It ends 4-1: both win. Only a match with exactly 3 goals splits them (Over 2.5 wins, Over 3.5 loses). Two slips, but most of the time one outcome.
What it did to our own backtest
Our Eredivisie overs strategy backs every Over line priced 1.50–2.50 at Toto. From 7 August to 31 October 2026 it placed 122 bets on 95 matches. In 27 of those matches it placed two or more bets (54 bets in total). In 23 of those 27 matches, every bet on the match had the same result; only 4 were split.
So the "122 bets" behave much more like 95-something independent outcomes. That matters for three numbers:
Measure | Every qualifying line ("all") | One bet per match ("best") |
|---|---|---|
Bets | 122 | 95 |
Profit | +€225.60 | +€213.70 |
ROI | +18.5% | +22.5% |
Luck band | ±17.5% | ±20.9% |
Longest losing streak | 5 | 5 |
Deepest drop | €98.00 | €75.80 |
- The luck band is too narrow when you count correlated bets as separate. With 122 "bets" the band says ±17.5%, but those bets are not 122 independent coin flips. Counting one bet per match, the band is ±20.9%, the honest one.
- Wins and losses arrive in clumps. A bad match costs two or three stakes at once. Drawdowns and streaks are bigger, or arrive faster, than the bet count suggests.
- The ROI barely moves, because the extra bets have similar odds. The bet count is what flatters.
Where you will meet this
- Several lines of one market: Over 2.5 + Over 3.5, Under 1.5 + Under 2.5, handicaps −1 and −1.5.
- Related markets in one match: a home win and over 2.5 goals go together more often than chance (in 758 matches we measured, 14% more often than if they were independent); a draw and under 2.5 goals much more (67% more often).
- Several matches with one common cause: every match on a stormy weekend, or every game of a team whose striker is injured.
This is also why a bet builder (several legs in one match combined into one price) is not the same as multiplying the legs' odds: the book knows the legs are related, and prices the combination accordingly.
How we handle it
- Every backtest on this site says whether it counted every qualifying line or one bet per match.
- When several lines qualify, the honest headline is the one-bet-per-match version, with the full version shown next to it.
- The bet count is always stated together with the number of matches: "122 bets on 95 matches" tells you more than "122 bets".
For your own betting
If you put several bets on one match, think of them as one bigger bet. Your record of "100 bets" might really be 60 independent events, and every conclusion you draw from it, including how lucky or unlucky you have been, should use the smaller number.
Source: Edge Terminal backtester, rule leagues=eredivisie&class=totals&outcomes=over&min=1.5&max=2.5&stake=10, per=all and per=best, Toto prices before kickoff, re-run 3 November 2026 on Eredivisie matches 7 August – 31 October 2026. Same-match correlation: Edge Terminal match results, 758 matches, measured 3 November 2026. Past results only; nothing here is a tip.