How we rate CS2 teams

Alongside the official ranking, we keep our own strength rating for every team, built on the Elo system first designed for chess. Elo is well suited to head-to-head sports: it gives each team a single number, and the gap between two numbers translates directly into a predicted win probability. That makes it ideal for answering the question fans actually ask — who is more likely to win this match, and by how much?

After every result, the winner takes points from the loser. The size of the swing depends on the surprise: beating a much stronger team is worth a big jump, while a favourite beating an underdog barely moves the needle. Over many matches this settles each team at a rating that reflects how well it actually performs against the field, updating automatically as new results come in.

Because Elo is built around expected outcomes, two teams' ratings give a clean win probability for a match between them. We show this on our head-to-head pages, and where a second-opinion model is available we display it alongside so you can see whether independent methods agree. When two models agree you can lean on the number a little harder; when they diverge, the matchup is usually closer than either rating alone suggests.

Our Elo rating and the official Regional Standings usually tell a similar story, but they are not the same thing. The official ranking is tuned around event stature and decides Major invites; our Elo is a pure predictive model tuned to forecast individual matches. The two are most interesting when they disagree — that gap often points to a team that is over- or under-rated by one system relative to the other.

No rating is perfect. Elo sees results, not context: it doesn't know a star player was ill, that a roster changed yesterday, or that a match was a dead rubber. Treat the numbers as a strong, honest baseline to reason from — not the final word.