The goal of o!TR is to measure relative performance in osu! tournaments across different rank ranges. osu! is unusual among competitive games in its thriving rank-restricted tournament scene. This encourages "sandbagging" and "deranking," which are terms the community has used to describe the act of manipulating one's osu! rank in order to play in tournaments intended for a lower skill level. o!TR assigns a numerical rating to each player, with higher ratings indicating better performance in matches. This rating value may be used as a replacement for rank restriction, as a screening process, or simply for players to keep track of their relative skill in tournaments over time.
Please note that rating systems do not reflect exact skill! There are subtleties and competing factors that go into any such system. We do not aim to answer the question of "Who is the best osu! tournament player?" and instead aim to help players interpret performance history. Remember that playing in more matches will improve the accuracy of your TR, and think of a higher rating as corresponding to better in-match performance rather than being a better player. If you have any questions not answered by this FAQ, feel free to join the o!TR Discord and a member of the team will be happy to help.
Both Elitebotix and Skill Issue are excellent tools. o!TR actually takes inspiration from both systems and uses the same fundamental algorithm as Skill Issue, OpenSkill, for calculating ratings. Qualitatively, the primary differences of o!TR are the following:
-
Active maintenance and match processing: To ensure validity of competitive data, tournament matches are only counted after match links are manually approved by the o!TR team. Qualifier lobbies, tryouts, and scrimmages are not approved. So long as submissions adhere to our guidelines, anyone can submit requests to add tournaments, remove warmups, and so on. This way, rating is only affected by actual competitive play in a valid tournament, making artificial inflation or deflation infinitely more challenging than deranking a play in solo or setting up a fake tournament lobby to manipulate your rating. Players would have to intentionally influence the outcome of a match to manipulate their TR, and the effects would only appear several weeks later after the tournament has concluded and passed our approval process.
-
Website interactivity: All collected match data and rating history is accessible via a website, and various statistics and graphs about player performance are viewable. In particular, users are able to see how they perform relative to others in each match and how that affects their rating. Other features include leaderboards, tournament and match-specific statistics, information about performance across mods, and so on. The website is also where you can submit tournament data to be counted in the o!TR calculations.
-
Fundamental interpretation: ETX estimates a star rating level that one can play comfortably, and SIP measures relative skill across various skillsets. o!TR is intended specifically to gauge competitive performance, though again, SIP uses a similar algorithm and can be useful for answering questions related to head-to-head matchups. In contrast to ETX and SIP, o!TR intentionally does not take map difficulty into account, which may be particularly useful for observation. Specifically, a player with a significantly higher o!TR ranking compared to ETX or SIP is likely playing in tournaments significantly below their skill level.
-
Open source code: o!TR is open source and communicates all algorithm changes publicly, meaning it can be used for screening in badged tournaments. We aim to have an open, compliant tool that meets the osu! Tournament Committee's high transparency standards.
-
Support for other gamemodes: Our rating formulas and submission processes are just as applicable to osu!taiko, osu!catch, and osu!mania. With that said, osu!mania has its own far more robust skillbanning system in place which is more than likely superior for screening.
Please note that your past rating history may frequently change as the rating algorithm is tweaked, new matches are approved, etc., even if you do not play any new matches. Thus, the rating graph should be viewed as a general trend, not an absolute measurement of your performance at every given time. For the sake of allowing o!TR to be used for screening, though, o!TR will update ratings at regular times. See the ratings page of this wiki for more details.