The CS2 rating explainedGlicko-2(ELO)
The UpFrag Counter-Strike 2 ranking is built on Glicko-2, a public algorithm created by statistician Mark Glickman and used in chess rating. No secret formula: the calculation is documented, and every point gained or lost is broken down after the match, component by component.
Four components, a fair ranking.
Round-by-round score
The result is evaluated round by round, not just on the win or the loss. A 12-13 loss and a 2-13 loss do not have the same impact on the ranking.
Individual performance
Personal contribution modulates the gain: eliminations, damage, entry frags, clutches, utility. A strong performance in a loss reduces the point loss; a weak performance in a win reduces the gain.
Demo analysis
After every match the demo is replayed in full: real accuracy, flashes that actually blinded, utility damage. It feeds UpData XP and the anticheat review. It does not change your rating, which is final when the match ends.
Placements and certainty
Every player starts at 500 with maximum uncertainty. Early matches therefore move the rating a lot, and the swings settle as your level is confirmed. The 10 placement matches are a progress marker: the maths there are identical to any other match.
Why my gains vary so much.
The biggest factor is not your performance, it is the gap with your opponent. Beating someone stronger is worth a lot, beating someone far below you is worth almost nothing: the result tells the system nothing it did not already know.
Two wins, same effort, six times fewer points. That is expected behaviour, not a penalty.
The second factor is certainty. At the start the system knows nothing about your level, so your first matches move it a lot. As results pile up that uncertainty narrows and your swings get smaller. It is mechanical, and it settles.
After 14 days without playing, uncertainty grows back: your next matches count for more. Your rating itself is never changed by inactivity.
The progression.
0-500
501-750
751-900
901-1050
1051-1200
1201-1350
1351-1530
1531-1750
1751-2000
2001+
The full breakdown is available after each match: base delta, gap with your opponent, performance modifier and round margin.
A transparent ranking.
10 matches to find your level, every point broken down after the game.
Get started →What players ask
What is the difference between Glicko-2 and classic ELO?
ELO keeps a single number. Glicko-2 adds a measure of uncertainty: while the system knows little about your level your results move the rating a lot; once it has placed you, the swings tighten. That is what makes it possible to place a player in about ten games without freezing them afterwards.
How many placement matches are required?
Ten. Every player starts at 500 with maximum uncertainty, and those first games serve to locate the level. The rating keeps moving afterwards, but more finely.
Why do some wins give so few points?
Because the dominant factor is the gap with your opponent. Beating a clearly lower-rated player pays little: the system expected that win. Beating someone above you pays a lot. The full calculation is shown after each match.
Does individual performance count towards the rating?
Yes, but it modulates the result without ever reversing it. A strong performance in a loss reduces the drop, a win carried by the team pays slightly less. The direction of the change stays dictated by the match result.