CS2 CT Side Win Rate Betting: Using Map Side Stats Without Fooling Yourself
Updated · 
Written by Lena Ortiz ·
Every CS2 fan has heard it: "Nuke is CT-sided", "Dust2 is a T map", "you need to win your CT half on Anubis". Some of that is true, some of it is out of date, and very little of it gets turned into actual numbers. If you bet on CS2, side win rates can help, but only in specific markets, bro.
This guide covers what CT side win rate means, what the main stats sources say about the current map pool and why they disagree, how a side edge changes first-half and half-time odds, and why it barely matters for the map winner bet. The side rates come from hltv, Abios and Skinrave, and we name the source each time. The probability calculations are our own, from a simple model, and every price is an illustrative example.
Reminder: betting is entertainment money, not income. Only bet where it's legal for you and if you're of legal age.
What CT side win rate means
In CS2 each map is split into two halves of 12 rounds. One team starts as Counter-Terrorists (CT), defending the bomb sites, and the other as Terrorists (T), trying to plant the bomb. They swap after round 12.
A map's CT round win rate is the share of rounds won by the CT side across a large number of pro matches. If CT wins 54.5% of rounds on a map, that map is "CT-sided". At 49% it's close to balanced or slightly T-sided. Some sites also publish CT win rates for pistol rounds or by half, which tell you slightly different things.
Side balance comes from map design: how fast T can reach the sites, how easy the sites are to retake, how many angles CT must hold. That's why it changes when Valve edits a map. The basics of each map's layout are in our callouts guides.
The current map pool
As of October 2026 the active duty pool, which pro events use, has seven maps: Ancient, Anubis, Cache, Dust2, Inferno, Mirage and Nuke. Two recent changes matter for side stats:
- Anubis came back on 21 January 2026, reworked. Valve changed several areas, including mid and the A site, so side stats from before 2026 describe a different version of the map.
- Cache joined on 8 July 2026, replacing Overpass. Pro teams have only a few months of Cache data, so any Cache side rate is based on a small sample.
Data for maps no longer in the pool, like Overpass, Vertigo or Train, doesn't help you unless they come back.
What the stats sources say
The main sources disagree, and that's the most important thing to know. Here's what each reports.
hltv map stats (as shown in October 2026). hltv's Nuke page shows CT winning 54.5% of rounds. Its Dust2 page shows CT at 49.4%, so Dust2 leans very slightly T. hltv's map pages also give pistol and round 2 numbers by side, plus a win probability matrix for any score.
Abios (2023-24 sample). Abios analysed 16,539 pro matches from January 2023 to January 2024 and published CT/T win rates for CS2 maps: Mirage 51/49, Nuke 46/54, Inferno 50/50, Vertigo 53/47, Ancient 50/50, Overpass 46/54 and Anubis 56/44. That's a big sample, but it's early CS2, and it describes the old Anubis.
Skinrave (continuously updated). Skinrave's map page shows Nuke at 52.9% CT, Mirage at 52% CT, Dust2 at 50.9% T, Inferno at 51% T and Overpass at 54.8% CT. For Anubis it shows the opposite of Abios, 44.5% CT and 55.5% T.
Look at Nuke: 46% CT in the Abios sample, 52.9% on Skinrave and 54.5% on hltv. Anubis flips from 56% CT to 44.5% CT. These differences come from:
- Time period. Early CS2 had different smokes, molotovs and economy rules than CS2 today. Meta and patches shift side balance.
- Map versions. Anubis before and after the 2026 rework are effectively different maps.
- Sample. Some sources include lower-tier matches, some only top events. Some count rounds, others count halves.
- Method. "CT win rate" can mean rounds, pistols, halves or map wins for the team starting CT, depending on the site.
So the honest summary is: Nuke currently looks CT-leaning on the most-used source, Dust2 and Mirage look close to balanced, and everything else needs checking against current data. Don't trust any single number from a blog, including this one, without looking at the date and the sample.
Where side stats matter, and where they don't
Map winner: hardly at all
This is the part many people get wrong. Each team plays 12 rounds on each side, so a side edge mostly cancels out. In our simple model, two equal teams on a map where CT wins 54.5% of rounds each win the map exactly 50% of the time, whether they start CT or T.
Side stats only help for the map winner when one team is much better on one side than the other, compared with how teams usually play the map. That's a team-specific effect, and the samples are small (see below).
First-half markets: quite a lot
First-half markets settle after round 12, so the team starting CT keeps the edge the whole time. Here's what a 12-round half looks like in our model for equal teams, depending on CT round win rate:
| CT round win rate | CT starter wins half | 6-6 | T starter wins half |
|---|---|---|---|
| 49.4% (hltv Dust2) | 37.1% | 22.5% | 40.4% |
| 50% | 38.7% | 22.6% | 38.7% |
| 54.5% (hltv Nuke) | 51.3% | 21.5% | 27.2% |
| 56% | 55.5% | 20.7% | 23.8% |
On a 54.5% CT map, the CT starter wins the first half almost twice as often as the T starter in our model, with a 6-6 in about a fifth of maps. If a book prices a "first-half winner" market without accounting for side, that's where a side edge can matter. Check whether the market is two-way (6-6 voids or pushes) or three-way (6-6 is its own outcome) before you bet.
First-half handicaps work the same way. In the same model, the CT starter on a 54.5% map wins the half by 4 rounds or more (8-4 or better) 29.2% of the time, against 19.4% on a perfectly balanced map.
Claim free cases on Clash.gg, no code needed
Half-time live odds: the biggest effect
At half-time the teams swap, and the score doesn't tell the whole story. Say Team A started T on Nuke and is 5-7 down at half. Now they switch to CT. In our model with a 54.5% CT rate, here's their chance of winning the map:
| Score at half (Team A first) | A switching to CT (54.5%) | Balanced map (50%) | A switching to T (45.5%) |
|---|---|---|---|
| 7-5 | 80.5% | 71.0% | 59.8% |
| 6-6 | 62.0% | 50.0% | 38.0% |
| 5-7 | 40.2% | 29.1% | 19.6% |
| 4-8 | 21.0% | 13.3% | 7.9% |
A team 5-7 down but about to play CT on Nuke has about a 40% chance in this model, compared with 29% on a balanced map. If the live price treats 5-7 like it would on any map, the team switching to CT could be underpriced. Books are aware of this, and many use models with side built in, so don't assume there's always an edge. But it's the right question to ask at half-time.
hltv's map pages include a win probability matrix based on past results at each score, which already includes the side effect. It's a useful cross-check against our simple model. Our CS2 live betting guide covers stream delay and other in-play practicalities.
Round handicaps and totals
Side balance affects the shape of a map's score, not just the result. On a strongly sided map, each team tends to bank rounds on its strong side, which pulls scores closer together. That nudges 12-12 and overtime slightly upward and makes big round handicaps a little harder to cover. In our model, the chance of a 12-12 between equal teams rises from 16.1% on a balanced map to 16.2% at 54.5% CT and 16.5% at 60% CT, so the effect is small unless a map is very one-sided. For handicap basics, see our CS2 handicap betting guide.
The sample-size problem
Map-level side stats use thousands of rounds, so they're fairly stable. Team-level side stats are a different story. Say you want to know if a team is genuinely strong on CT on Nuke. Here's how much uncertainty a CT rate of 54.5% carries depending on how many rounds you have:
| CT rounds in the sample | Approx. maps on that side | Margin of error (95%) |
|---|---|---|
| 480 | 40 | plus or minus 4.5 points |
| 1,200 | 100 | plus or minus 2.8 points |
| 4,800 | 400 | plus or minus 1.4 points |
Even 40 maps of CT rounds leaves an error of about 4.5 points, almost the whole gap between a balanced map and a CT-sided one. And rounds within a map aren't independent (a team on a good economy run wins several in a row), so the real uncertainty is higher than this. Most teams don't play 40 maps of Nuke in a few months, and rosters change in between.
So use map-level stats as your baseline and treat team-level side stats as a weak signal, not a reason to bet.
A practical checklist
Before using side stats for a bet:
- Check the date and sample of the stat. Pre-2026 Anubis data and early CS2 data may not apply.
- Know the market. Side matters for first-half and half-time markets, and very little for the map winner.
- Find out who starts on which side. Usually it's decided by a knife round or a side pick in the veto, so you might only know just before the map.
- Compare with the book's price. Convert odds to implied probability with our esports odds calculator and compare with your estimate.
- Don't stack weak signals. A small side edge plus a small form edge plus a "feeling" isn't a big edge.
CSGOEmpire match betting + free case with FREECASES4U
A worked example
Here's a made-up example with illustrative prices. Two evenly matched teams play Nuke, and you take hltv's 54.5% CT rate as your baseline. Team A wins the knife round and picks CT.
A book offers a two-way first-half winner market (6-6 voids) at 1.90 for each team. In our model, Team A (starting CT) wins the half 51.3% of the time and Team B 27.2%, with 6-6 in 21.5%. Remove the 6-6 and Team A wins the half about 65% of the decided cases, so 1.90 (52.6% implied) looks generous. In this illustration, the value would be about 0.653 × 1.90 − 1 = +24% per decided bet.
In real life, a book pricing this market at 1.90/1.90 on Nuke would be unusual. Most books account for side, and the two teams may not be equal. But it shows why side stats belong in first-half markets, not in the map winner.
Legal and responsible betting
Side stats can make betting feel more like analysis, and that can lead to bigger stakes than you planned. Even a sound model loses plenty of bets. Set a limit before the match and keep stakes small.
Only bet with licensed books that accept players from your country, and only if you're 18+ (or the legal age where you live). If betting stops being fun, take a break and use the tools and links on our responsible gambling page.
Bottom line
CT side win rate tells you which way a map leans. On hltv's current numbers, Nuke leans CT at 54.5% while Dust2 is close to even. But sources disagree, the map pool changed twice in 2026, and team-level samples are small. Side stats barely affect the map winner because both teams play both sides. They matter most in first-half markets and at half-time, where a team about to play its strong side is better placed than the score suggests. For other markets, see our CS2 betting guide and odds explained page.