Over/under markets let you bet on how long a CS2 match lasts instead of who wins it. That makes them useful when you have a view on how close a matchup is but no strong opinion on the winner. Both main versions, total maps and total rounds, can be priced from a couple of simple formulas, as long as you know where those formulas break.
Total maps in a best-of-three
A best-of-three ends 2-0 or 2-1. So the standard line is 2.5 maps: over wins if a third map is played, under wins on a sweep by either team.
If the favorite wins each map with chance p and maps are independent, a third map happens when the first two are split. The favorite can win map one and lose map two, or the other way round, so:
P(3 maps) = 2 × p × (1 - p)
Fair odds for each side, before bookmaker margin, look like this. The left column is the favorite's chance of winning a single map.
| Map win | 3 maps | Over 2.5 | Under 2.5 |
|---|---|---|---|
| 50% | 50.0% | 2.00 | 2.00 |
| 55% | 49.5% | 2.02 | 1.98 |
| 60% | 48.0% | 2.08 | 1.92 |
| 65% | 45.5% | 2.20 | 1.83 |
| 70% | 42.0% | 2.38 | 1.72 |
| 75% | 37.5% | 2.67 | 1.60 |
| 80% | 32.0% | 3.13 | 1.47 |
Under this model the over can never be more likely than 50%, and it only reaches 50% when the teams are dead even. The chance falls slowly at first (49.5% at 55%) and faster as the gap widens. Under 2.5 maps is the same bet as "either team -1.5 maps", which our handicap guide covers.
Best-of-five finals use lines like 3.5 and 4.5 maps. With the same model, a sweep (3 maps) has chance p³ + (1 - p)³ and a full five maps has chance 6 × p² × (1 - p)². For a 65% favorite that is 31.7% for three maps and 31.1% for five.
Why real series go to three maps more often
The flat model gives every map the same p. Real best-of-threes do not work that way. In a standard veto, each team picks one map, usually its strongest, and the third is a decider (the order is in our map veto guide). So the favorite has one chance a on its own pick and a lower chance b on the underdog's pick. The series reaches a third map when those two are split:
P(3 maps) = a × (1 - b) + (1 - a) × b
For any fixed average of a and b, this is lowest when the two are equal. Any gap between the picks pushes it up. Keeping the favorite's average at 65%:
- 65% on both picks: 45.5% chance of three maps (the flat model).
- 75% on its pick, 55% on the underdog's: 47.5%.
- 80% on its pick, 50% on the underdog's: 50.0%.
- 85% on its pick, 45% on the underdog's: 53.5%, which is above the flat model's ceiling of 50%.
That last case happens in practice: a team that is clearly better overall can still be the underdog on a map its opponent has played far more and prepared in depth.
So when a book prices over 2.5 maps from overall team ratings, the over is the side to check first, especially when both teams have clear, different comfort picks.
Total rounds under MR12
Total rounds markets are usually set per map. CS2 plays MR12: two halves of 12 rounds, first to 13 wins. A regulation map therefore lasts between 13 rounds (13-0) and 24 rounds (13-11).
At 12-12 the map goes to overtime. Under current tournament rules, overtime is MR3: six rounds, first to four, with teams switching sides after three. If an overtime ends 3-3, another one starts. Each overtime starts with both teams on the same fixed bank ($10,000 under the ESL Pro Tour rules), so there are no pistol rounds in overtime.
That gives these possible totals:
- Regulation: 13 + the loser's rounds. 13-9 is 22 rounds, 13-11 is 24.
- First overtime: 28 (16-12), 29 (16-13) or 30 (16-14).
- Second overtime: 34 to 36, and so on.
Take a line such as 21.5. The over wins on 13-9, 13-10, 13-11 or any overtime. The under wins on 13-8 or anything more one-sided. There is no total between 25 and 27, so every overtime map clears any line up to 27.5.
Pricing a total rounds line
The cleanest starting point is a round-by-round model: each round is independent, and one team wins each round with the same chance. The table below runs that model through MR12 and MR3 overtime. The over 21.5 column shows the probability with fair odds in brackets.
| Round win | Map win | Over 21.5 | Overtime |
|---|---|---|---|
| 50% | 50.0% | 61.7% (1.62) | 16.1% |
| 52% | 58.5% | 60.9% (1.64) | 15.8% |
| 54% | 66.7% | 58.5% (1.71) | 14.9% |
| 56% | 74.2% | 54.7% (1.83) | 13.5% |
| 58% | 80.7% | 49.8% (2.01) | 11.8% |
| 60% | 86.1% | 44.1% (2.27) | 9.9% |
What the model says:
- For two even teams, the most likely single result is a 13-11 (16.1% across both teams), and overtime happens just as often. The expected length is about 23.1 rounds, and the line that splits outcomes closest to 50/50 is 22.5 (47.6% over).
- The gap between teams drives the total. A team that wins 60% of rounds wins 86% of maps, and the expected length drops to about 21.5 rounds.
- Small round edges matter. Moving from 50% to 54% of rounds only takes the over 21.5 from 61.7% to 58.5%, but it takes the map win chance from 50% to 66.7%.
The model has one big blind spot: CS2 rounds are not independent. Money carries over, so a team that wins a round is often better armed for the next one. That makes rounds come in streaks, and streaks produce more one-sided maps than the model expects. In practice, treat the model's over probability as an upper bound and shade toward the under, more so for teams with a strong economy game.
Overtime and how books settle totals
Settlement rules differ, so check the market rules on your book. The published rules we link below say:
- bet365: for CS2, overtime is included in settlement unless a draw option is quoted for the market, in which case regulation rounds are used.
- SX Bet: if no draw option is offered, extra time counts if played. Bets on a map are cancelled if a team retires or is disqualified before all scheduled rounds are played.
- Pinnacle: overtime and other tiebreakers are valid for deciding results.
- Sky Bet: map round handicaps include overtime unless stated otherwise, and results left undetermined because not all maps were played (for example a third map after a 2-0) are void.
Because overtime jumps straight to 28 rounds or more, it decides a lot of over bets. In the model, even teams go to overtime about one map in six. If you think a map is closer than the price implies, the over is usually a better way to express that view than either moneyline.
Series total rounds (rounds added up across a best-of-three) are a different market. They depend mostly on whether a third map is played, so price them by combining the total maps estimate with a per-map rounds estimate.
What moves round totals
- Matchup gap: the biggest factor, as the table shows. Use round win rates per map against similar opposition rather than an overall ranking gap. Our stats sites guide lists where to find them.
- The map itself: some maps favor CT or T. A side bias on its own does not change the total much for even teams, but it does mean the half-time score can mislead: a team up 8-4 after starting on the strong side has not done as well as it looks.
- Map pick: a team on its own pick tends to win more rounds, which widens the gap and favors the under. Deciders, where neither team has its comfort pick, can be closer.
- Pistol rounds: each half starts with one. The pistol plus the two rounds that usually follow against a weaker buy are 3 of the 12 rounds in a half, a quarter of it. A team that wins both pistols and converts them has 6 of its 13 rounds before the other team gets a full buy, which is a common route to an under.
- Team history: look at how each team's recent maps on this map finished. A record full of 13-10 and 13-11 scores says more about the total than a reputation for being "streaky" or "solid".
A quick checklist before you bet a total
- Check the format and the market: per map or per series, and whether a draw option changes overtime rules.
- Wait for the veto if you can. Total rounds depend heavily on which map is played.
- Estimate each team's round win rate on that map, then read the over probability off the table.
- Shade toward the under for streaky, economy-driven matchups.
- Convert the book's prices with the no-vig calculator and only bet when your number clears the fair price. The odds guide shows how.
Our own picks are posted with their odds before each match, come from SkinBetHub's proprietary signal model, built on prediction-market data, and are graded at a flat 1 unit. Every result stays on the results page.