Elo sucks — better multiplayer rating systems for smaller games

I have designed a new rating system called the Aco rating system, which is similar to Elo except it fixes a few key problems.

  1. Actual win rates: The win probability is calculated from the actual data of the past 100000 games. This means it does not need to fit an exponential curve. For example, the system could lookup its database and see that a matchup of a 1800 rating player vs a 1300 rating player results in the higher-level player winning 76.3% of the time. If the high-level player is outperforming this, they gain points over time, and that is a fair system, based in actual data.

  2. Newbie suppression: The points able to be gained/lost is reduced when you are playing someone substantially lower in rating than you. This makes people happier because they can’t lose as many points to newbies. Instead, they can only lose the most points to people similar to their skill level, which feels much more fair.

  3. Small increments: In general, you will gain or lose about 1 point per game. This means each game is insignificant in the scheme of things, and it doesn’t hurt to play every game in ranked mode. Competing systems like TrueSkill or Glicko say their advantage is that you can converge on your rating a lot faster, sometimes adding 50 or 100 points from a single game. I actually found this was a disadvantage. The slow rating increase of Aco means that if you reach the top of the leaderboard, you know for sure you have really earned it and that it is not just caused by uncertainty error in the rating system.

  4. Daily decay: Every day, a person’s rating decays by 5 points. This ensures everyone is encouraged to keep playing ranked and defend their title. To preserve the true rating unchanged, the decay is stored separately from the rating, caps at 100 and each game cancels out 1 point of decay. Previously, people would camp at the top leaderboard, simply not playing to maintain their position, and that was not any fun.

Kiinnostava tiivistelmä erilaisista kilpapelien pisteytysjärjestelmistä ja niiden sopivuudesta eri kokoisiin peleihin. Lopussa on tiivistelmä muutoksista, joilla kirjoittajan peli saatiin pisteyttämään pelaajia reilummin.

Kohdan 4 jatkuva pisteiden laskeminen on sekä hyvä että huono asia. Oma etenemiseni Trackmania 2:n Canyon-moduulin listoilla pysähtyi, koska pisteiden voittamiseen tarvittavia korkeamman tason pelaajia ei löytynyt enää hetken päästä pelin palvelimilta. Pisteiden laskeminen olisi korjannut tämän, mutta aiheuttaa ongelmia mikäli pelaaja ei pääse jostain syystä pelaamaan hetkeen.

Toinen ongelma on Trackmania 2:n esport-skene, jossa virtuaaliset turnaustapahtumat tuo korkeatasoiset pelaajat yhteen. Näihin ei itsellä ollut kiinnostusta osallistua aikataulujen takia.

Ratkaisuna voisi ehkä olla useampi eri pistelista, joissa pisteiden putoaminen olisi aggressiivisempaa, mutta tämä ei auta jos peli keskittyy pääasiassa turnaustapahtumiin.