On Kalshi, winning the game is not winning the match

Short answer: on Kalshi, the Counter-Strike 2 market called KXCS2GAME is not a game. It settles on the whole match, best of one or best of three. The market that settles on a single game is the Counter-Strike 2 map market, KXCS2MAP. Every esport on the exchange names them this way, and if you are trading off anything happening inside a live game, you are probably holding the wrong one.

This lab found out by doing it. The details are at the bottom, along with the ledger.

What the two tickers actually resolve on

The titles say it plainly once you line them up.

The word doing the damage is game. In the ticker it means the match. In the sport it means one map.

The same inversion holds across every esport Kalshi lists. In each pair the first is the whole match and the second is a single map.

Winning map one wins the match 82% of the time

Across 793 settled CS2 events this lab recorded, a team won map one on 770 occasions and went on to take the match 630 times. That is 81.8%.

So the moment a team has map one sewn up, the map contract is worth 100 cents and the match contract is worth about 82. Those are different instruments with different values, and the difference is 18 cents.

The market is not wrong about this

The obvious next thought is that the match should therefore trade below map one, and that a trader could farm the gap. It does not, and they cannot.

Across 310,557 moments where both markets were quoted for the same team in the same minute, the match traded a median of +0.00 cents against map one, and a mean of -0.06. There is no systematic discount in either direction.

The two are not welded together, though. The middle half of those observations runs from -5.0 to +5.0 cents, so the pair wanders several cents apart in both directions minute to minute. It wanders symmetrically, which is the part that matters here: the match is not being sold cheap relative to the map. And where a team is a strong map one favourite, at 65 cents or better, the match actually trades dearer, 79.3 against 75.2 across 53,758 observations.

That is correct pricing, not a mistake. The same thing that makes a team likely to win map one, being the better team, makes them likely to win map two. Before the match starts there is no discount to collect.

Where the money actually leaks

The trap is not in the prices. It is in what your information is worth.

A live game clock, a scoreboard, an economy read, anything happening inside the current map, tells you about one game. That information is worth up to the full hundred cents on the map contract. On the match contract it is worth about 82, because winning the current map only banks part of the job.

But the two contracts cost the same. So a trader acting on in game information through the match ticker pays full freight for roughly 82% of the edge, and never sees the missing 18 cents leave, because nothing about the fill looks wrong.

It is worse in the specific case that feels most compelling. A team dominating the current map while trailing nil one in a best of three is not close to winning the match; they are close to levelling it. The clock is screaming and the match contract is barely moving, which reads like an opportunity and is arithmetic.

And the volume is on the wrong side

This would be a curiosity if the two contracts were traded equally. They are not.

There are fewer match markets than map markets, and they carry 4.7 times the volume, about 6 times as much per market. The exchange's esports liquidity sits almost entirely in the contract that in game information is worth least on.

That is not proof that anybody is making this mistake. Plenty of good reasons exist to trade the match: it is the natural bet, it lasts longer, and it is what a spectator thinks they are betting on. But if you are watching a live game and reaching for the busier ticker because it is the busier ticker, the arithmetic above is working against you.

How this lab got it wrong

Honesty section, because this one was ours.

On 14 August this lab started a paper lane that watched Valve's official live game feed and backed heavy favourites once a game reached its late stage. It matched markets on team names. Team names appear in both ticker families, and the lane took the first thing it found, which was the match contract.

It ran 16 fills. Every single one of them, 16 of 16, was on the Dota 2 match contract KXDOTA2GAME, placed off a single game's clock. It won 14 of them and averaged +6.63 cents a fill, which looked like a working strategy and was measuring something nobody had designed.

The record is kept and has been relabelled, because it is a real result for a strategy this lab did not intend to run and did not describe correctly. It proves nothing about the match clock idea, which now has to earn its number again from zero on the right instrument.

What follows from that is arithmetic, and it is the same for anybody. Information from inside a game belongs on the map contract for the map being played. It belongs on the match contract only when winning that map ends the match, which is best of one always, best of three at one nil up, and best of five at two nil or two one. Every other time, the clock is telling you about a map while the match contract is a different bet.

Check it yourself

Every claim here is recomputable. GET /trade-api/v2/markets?series_ticker=KXCS2GAME and the same for KXCS2MAP, the Counter-Strike 2 match and map markets, return the titles quoted above. Settlement comes from Kalshi's own resolution. The conditional and the price comparison are from this lab's recordings of the book, swept 22 August 2026, and the lane ledger is 16 rows of the lab's own paper fills.

The lab's full board publishes every Sunday. The Seats lists what each strategy does.