What a matchup number says
A matchup is a pairing of two things that can face each other: two characters, two team compositions, or two drafts. Matchup win chances are an attempt to say how often one side wins that pairing, usually expressed as a split such as 55-45 or 60-40.
The number is a summary of many games. It does not describe any single game, and it does not say what will happen when two particular players meet. It is best read as the average tilt the pairing gives, with everything else, including skill, treated as equal.
How the figures are usually estimated
There are a few common sources.
- Observed results. Collect many games where the pairing occurred and compute the share each side won. This is the most direct approach, and it improves as the sample size grows.
- Player judgement. In smaller communities, experienced players assign matchup ratings by consensus. These can be informative, but they reflect opinion, and opinion can lag behind changes to the game.
- Models. Some analysts adjust results for the strength of the players involved, so that a character is not credited for wins that belong to a very strong player.
Each method has weaknesses, which is why matchup charts from different sources can disagree.
Why raw numbers can mislead
Skill is mixed in. If better players tend to choose one option, its results look strong even if the option itself is average. This is a form of selection effect, and it is a common reason that tier lists, which summarise results, can be misread.
Small samples. A matchup seen in a handful of games has a very uncertain figure. A split of 8-2 over ten games is not reliable evidence of an 80% matchup.
Context changes. Patches alter the underlying balance. A matchup number from before a change may no longer apply, which is one reason balance patches are followed so closely by competitive players.
Averages hide variation. A pairing might be even on average but lopsided on certain maps, or depend on which side starts with an advantage.
What a counter really changes
A counter is an option that does well against a specific opponent option. The word suggests a decisive answer, but in most competitive games it is a modest tilt rather than a lock.
Consider a matchup that favours one side at 55% per game. Over a best-of-three series, the probability of winning is 0.55² + 2 × 0.55² × 0.45 = 0.3025 + 0.27225 = 0.57475, which is about 57.5%. In a best-of-five it is a little higher, but still well short of certain. The favoured side wins slightly more than half the time, and the other side wins nearly as often.
For a more lopsided 65-35 matchup, a best-of-three is 0.65² + 2 × 0.65² × 0.35 = 0.4225 + 0.29575 = 0.71825, about 72%. That is a meaningful edge, yet the underdog still wins more than one series in four.
This is why counters are usually discussed as part of a wider plan. In fighting games, picking a character to answer the opponent’s choice is a legitimate part of a set, and character counterpicks work within this kind of modest-edge logic. In team games with a draft, the same idea applies at larger scale, with the draft phase serving as the stage where each side tries to build the better matchup.
Team and draft matchups
When the matchup is between whole compositions, there are more moving parts. A team’s draft is a bundle of strengths and weaknesses, and the opposing draft may exploit some of them. Analysts look at questions such as whether one composition tends to be stronger early and the other later, which affects how games unfold.
An important distinction is between a drafting advantage and a playing advantage. A draft that looks better on paper only helps if the team can execute it. A team practised on an unusual composition may outperform what the generic matchup figure predicts, and the reverse is true when a team is handed something unfamiliar. Matchup figures from large public samples describe average play and cannot capture that.
Cyclical relationships also appear, in which A beats B, B beats C and C beats A. These are the basis of an archetype triangle, and in such systems no option is best in general. The correct pick depends on what the opponent is likely to do, which brings in uncertainty about hidden choices. Where picks are simultaneous, it becomes a guessing problem, related to mixed strategies, because predictable choices can be exploited.
Reading matchups in a tournament context
A matchup percentage is useful for a single game, but series involve repeated play, adaptation and map or character choices over several games. The side behind in a matchup can adapt, choose differently the next game, or aim to avoid the pairing. For this reason, a matchup edge that looks large in isolation often shrinks in a series with choices on both sides.
It is also worth noting that the opposite direction applies. An even matchup does not guarantee an even series if one side has better preparation.
A practical approach
When assessing a matchup claim:
- Ask how many games support the figure, and from what range of players.
- Check whether it predates a balance change.
- Think of the figure as a tilt, then convert it into a series chance rather than treating it as a result.
- Combine it with other evidence, such as the players’ own form and records, rather than leaning on the chart alone.
The conservative reading of any matchup figure is that it moves the odds a little, not a lot, and that each game still contains a large measure of uncertainty.