There has never really been a complex way of ranking pool players outside of averages and points. I know pool leagues and regional tours try their best to figure basic standings; however, I've learned that a lot more can be done to figure our not only who is the best, but who is "predicted" to win.
There are currently three main types of ranking systems:
• Spot-based, where rank of the player is his position in the list of participants, e.g. 15th. After each game, winner and loser swap their ranks.
• Point-based, where the rank of a player is based on the number of points he/she earned participating in games/events, e.g. 15. After each game/event, both parties get a number of points depending on the outcome; this number does not depend on current players’ ranks.
• Elo-based, where rank of the player is based on the number of points he has, e.g. 1345. After each game players are added and subtracted a number of points depending on who is the winner and their current ranks. In case of victory stronger player gets fewer points then the weaker one and reverse. This is frequently used in chess and scrabble (for those who play).
Every system has its weaknesses and the main ones are that ranks depend on game order and don’t always display real skill of players (which makes them unusable when it comes to figuring the probability of victory and bookmaking).
Tarasov’s Ranking Index:
The Tarasov’s Ranking Index (TRI) is based on complex math and provides much stronger prediction capabilities. The main merits of the TRI system are:
1. The possibility to create accurate rankings with few games that is particularly useful for large rankings with tens of thousands of participants.
2. The TRI takes all the games that have even been played into account, as opposed to the other ranking systems that take only current opponents’ ranks into account.
3. The TRI builds an "interconnection system" among participants, in contrast to the other systems that merely change the number of rank points by the outcome of the game.
4. The game score is taken into account.
5. The game age is taken into account, i.e. the older the match, the less its result affects the rankings.
6. The TRI system allows calculation of win chances of teams and could be used in calculation of the sizes of bets on game results.
7. If a cheating is detected the system allows to remove any number of fake games while completely correcting the ranking as though these games had not been played at all.
Applications in Bookmaking
TRI can be easily applied in bookmaking and do the following tasks:
• Determine relative skill of each player comparing to each other.
• Determine probability or victory for each player.
• Determine probable score in case of victory.
Every sport (including pool, obviously) has its specifics, and sports greatly vary in the degree in which randomness influences game outcomes (lucked-in 9balls, lucking someone behind the 8 with only 2 balls on the table, etc). The TRI system can be easily tuned to take these factors into account.
--------
I know this is boring stuff for the standard pool player, but cutting-edge ranking technology like this can do a lot for pool. It basically provides an artificial intelligence factor in determining ranks, probability of wins and predictions of outcomes. Not like any of you guys are running booking operations for pool, but technology like this could have a huge affect on calcuttas as well as the "true" rankings of players.
For example, if one person wins the winners bracket and each match goes 11-10 his favor, and the guy he beats in the winners bracket finals won 11-0 in each of his other matches - and comes back to the final and loses 11-10 again, the true-ranking between the two players isn't as far away as a points ranking would typically be (hope that makes sense).
If any tournament director, league owner, room-owner, etc, would like to learn more --- feel free to PM me with any questions. I'm considering bringing this technology to America and would like to hear your input.
Sorry for the long post
Dave
There are currently three main types of ranking systems:
• Spot-based, where rank of the player is his position in the list of participants, e.g. 15th. After each game, winner and loser swap their ranks.
• Point-based, where the rank of a player is based on the number of points he/she earned participating in games/events, e.g. 15. After each game/event, both parties get a number of points depending on the outcome; this number does not depend on current players’ ranks.
• Elo-based, where rank of the player is based on the number of points he has, e.g. 1345. After each game players are added and subtracted a number of points depending on who is the winner and their current ranks. In case of victory stronger player gets fewer points then the weaker one and reverse. This is frequently used in chess and scrabble (for those who play).
Every system has its weaknesses and the main ones are that ranks depend on game order and don’t always display real skill of players (which makes them unusable when it comes to figuring the probability of victory and bookmaking).
Tarasov’s Ranking Index:
The Tarasov’s Ranking Index (TRI) is based on complex math and provides much stronger prediction capabilities. The main merits of the TRI system are:
1. The possibility to create accurate rankings with few games that is particularly useful for large rankings with tens of thousands of participants.
2. The TRI takes all the games that have even been played into account, as opposed to the other ranking systems that take only current opponents’ ranks into account.
3. The TRI builds an "interconnection system" among participants, in contrast to the other systems that merely change the number of rank points by the outcome of the game.
4. The game score is taken into account.
5. The game age is taken into account, i.e. the older the match, the less its result affects the rankings.
6. The TRI system allows calculation of win chances of teams and could be used in calculation of the sizes of bets on game results.
7. If a cheating is detected the system allows to remove any number of fake games while completely correcting the ranking as though these games had not been played at all.
Applications in Bookmaking
TRI can be easily applied in bookmaking and do the following tasks:
• Determine relative skill of each player comparing to each other.
• Determine probability or victory for each player.
• Determine probable score in case of victory.
Every sport (including pool, obviously) has its specifics, and sports greatly vary in the degree in which randomness influences game outcomes (lucked-in 9balls, lucking someone behind the 8 with only 2 balls on the table, etc). The TRI system can be easily tuned to take these factors into account.
--------
I know this is boring stuff for the standard pool player, but cutting-edge ranking technology like this can do a lot for pool. It basically provides an artificial intelligence factor in determining ranks, probability of wins and predictions of outcomes. Not like any of you guys are running booking operations for pool, but technology like this could have a huge affect on calcuttas as well as the "true" rankings of players.
For example, if one person wins the winners bracket and each match goes 11-10 his favor, and the guy he beats in the winners bracket finals won 11-0 in each of his other matches - and comes back to the final and loses 11-10 again, the true-ranking between the two players isn't as far away as a points ranking would typically be (hope that makes sense).
If any tournament director, league owner, room-owner, etc, would like to learn more --- feel free to PM me with any questions. I'm considering bringing this technology to America and would like to hear your input.
Sorry for the long post
Dave