Approved by vapicuno
What is Elo?
Elo is a rating system originally developed for chess by Arpad Elo. Everyone's rating starts at 1000 and your rating increase or decreases depending on if you lose or win. How much rating you win or lose depends on the difference between your rating and your opponent's rating with the winner of the match gaining more Elo the lower they are rated relative to their opponent and vice versa. The maximum amount of rating that can be lost or gained in a single game is determined by the K value, for example if the K value is 20 then a game between two equal opponents will result in a gain and loss of 10 rating for the winner and loser respectively. This sheet uses a K value of 40 for the first thirty games a person plays and then a K value of 20 afterwards. ELO is also a prog rock band.
Why are the ratings so low?
In chess top end ratings reach over 2800, here they only get over 1400. I used the same K values as FIDE so a lot of this just comes down to sample size, only two players have logged over 500 ADV OU tournament games compared to the thousands played by chess professionals. In addition, draws are significantly less common in Pokemon and by extension, top players lose way more games. But also, yea, RNG puts more of limit on how well someone can do over a long period of time compared to chess.
What games are used in this sheet?
In short, every game from an Official Smogon tournament, Circuit tournament, RoA tournament, Invitational tournament, and the main Revival tournaments. This adds up to over 35,000 games from the finals of OST I in 2005 to today.
Is this even accurate?
Honestly, I'm not sure. Due to most win posts over the years not including specific details multiple assumptions were made for Best of 3 sets if no information could be determined.
1. If the numbers of games played was not reported, assume a 2-0 victory for the winner
2. If all three games were played but game order cannot be determined, assume winner won games 2 and 3 (this includes 4-game sets with ties)
3. If the set was only partially completed then those games are counted but any activity or gifted wins are not
The underlying reasoning for these assumptions was to always favor the winner when in doubt. Overall though this leads to a significant proportion of the total amount of games having their results assumed on my part or possibly missed altogether. I can only work with what I found which has possibly caused to some players' rating to butterfly away from where they should be. If you can prove the results of a series I have entered incorrectly please let me know.
Also this rating system does have a recency bias. This is largely due to the massive increase in the number of ADV tournament games in the past few years (as in a tournament today can have more games played than entire years in the first half of the 2010s).
Other Stuff
In addition to just the exciting Elo ratings pages there are also other pages dedicated to a multitude (3) of fun things for you to look at. First there is Peak Rating sheet which displays the timeline of the highest rating achieved until today as well as when certain rating milestones where reached. The Top 100 individual ratings achieved are also displayed but you can find every peak rating (and inactive current rating) in the Full Rankings tab.
Also to not have every person's play be boiled down to current and peak rating, there are two tabs dedicated to showing the overall progression of ratings and titles to display player's long term achievement. First, there is the Biannual Rankings which dhows a snapshot of the top 20 active player ratings every January 1st and July 1st. The amount of times someone has appeared in the top 20, 10, 5, 3, and 1 are all tracked. And finally there are the Titles. Similar to chess with its Master titles, this sheet includes four levels of sufficiently corny "Trainer" titles with four levels of achievement based on obtaining and maintaining a rating of a certain level. These titles are displayed elsewhere in the sheet by their corresponding Pokeball next to their names.
So that's all. If there is anything that you think could be changed, improved, or anything please let me know either here or on discord. Thanks to Dave and Ticken for their support and thanks to anyone who reads this.
What is Elo?
Elo is a rating system originally developed for chess by Arpad Elo. Everyone's rating starts at 1000 and your rating increase or decreases depending on if you lose or win. How much rating you win or lose depends on the difference between your rating and your opponent's rating with the winner of the match gaining more Elo the lower they are rated relative to their opponent and vice versa. The maximum amount of rating that can be lost or gained in a single game is determined by the K value, for example if the K value is 20 then a game between two equal opponents will result in a gain and loss of 10 rating for the winner and loser respectively. This sheet uses a K value of 40 for the first thirty games a person plays and then a K value of 20 afterwards. ELO is also a prog rock band.
Why are the ratings so low?
In chess top end ratings reach over 2800, here they only get over 1400. I used the same K values as FIDE so a lot of this just comes down to sample size, only two players have logged over 500 ADV OU tournament games compared to the thousands played by chess professionals. In addition, draws are significantly less common in Pokemon and by extension, top players lose way more games. But also, yea, RNG puts more of limit on how well someone can do over a long period of time compared to chess.
What games are used in this sheet?
In short, every game from an Official Smogon tournament, Circuit tournament, RoA tournament, Invitational tournament, and the main Revival tournaments. This adds up to over 35,000 games from the finals of OST I in 2005 to today.
Official Individual Tournaments
Official Smogon Tournament (1-3)
Smogon Tour (1-5, 11-16)
Smogon Classic & ADV Cup (1-10)
Smogon Frontier (8)
Official Team Tournaments
Smogon Premier League (1-16)
World Cup of Pokemon (2, 9-14)
Other Team Tournaments
RoA Premier League
Retro Cup of Pokemon
ADV Premier League
RoA Team Trial
Invitationals
CALLOUS Invitational
Jimvitational
Circuit Tournaments
ADV Circuit Playoffs
ADV Global Championship
ADV Seasonal
ADV Majors
ADV Swiss
Smogon Championship
Ruins of Alph Individual Tournaments
RoA Tour
RoA Olympics
Victory RoAd
Indigo Plateau
Major League RoA
RoA Ladder Tournament
RoA Forum Championship
World's Longest Tournament
Other Tournaments
ADV Revival
Rising Stars
If there is anything that you think should be added, let me know
Official Team Tournaments
Other Team Tournaments
RoA Team Trial
Invitationals
CALLOUS Invitational
Jimvitational
Circuit Tournaments
Ruins of Alph Individual Tournaments
World's Longest Tournament
Other Tournaments
ADV Revival
Rising Stars
If there is anything that you think should be added, let me know
Is this even accurate?
Honestly, I'm not sure. Due to most win posts over the years not including specific details multiple assumptions were made for Best of 3 sets if no information could be determined.
1. If the numbers of games played was not reported, assume a 2-0 victory for the winner
2. If all three games were played but game order cannot be determined, assume winner won games 2 and 3 (this includes 4-game sets with ties)
3. If the set was only partially completed then those games are counted but any activity or gifted wins are not
The underlying reasoning for these assumptions was to always favor the winner when in doubt. Overall though this leads to a significant proportion of the total amount of games having their results assumed on my part or possibly missed altogether. I can only work with what I found which has possibly caused to some players' rating to butterfly away from where they should be. If you can prove the results of a series I have entered incorrectly please let me know.
Also this rating system does have a recency bias. This is largely due to the massive increase in the number of ADV tournament games in the past few years (as in a tournament today can have more games played than entire years in the first half of the 2010s).
Other Stuff
In addition to just the exciting Elo ratings pages there are also other pages dedicated to a multitude (3) of fun things for you to look at. First there is Peak Rating sheet which displays the timeline of the highest rating achieved until today as well as when certain rating milestones where reached. The Top 100 individual ratings achieved are also displayed but you can find every peak rating (and inactive current rating) in the Full Rankings tab.
Also to not have every person's play be boiled down to current and peak rating, there are two tabs dedicated to showing the overall progression of ratings and titles to display player's long term achievement. First, there is the Biannual Rankings which dhows a snapshot of the top 20 active player ratings every January 1st and July 1st. The amount of times someone has appeared in the top 20, 10, 5, 3, and 1 are all tracked. And finally there are the Titles. Similar to chess with its Master titles, this sheet includes four levels of sufficiently corny "Trainer" titles with four levels of achievement based on obtaining and maintaining a rating of a certain level. These titles are displayed elsewhere in the sheet by their corresponding Pokeball next to their names.
So that's all. If there is anything that you think could be changed, improved, or anything please let me know either here or on discord. Thanks to Dave and Ticken for their support and thanks to anyone who reads this.
