I have developed a system to automatically rate players using information from the score sheets only. I have been using it since October, 2008, for all the US tournaments that I attend, or that I can obtain the score sheets, and it keeps all the players’ ratings current without making any rating mistakes. The APA doesn’t use it for their official ratings, but I’ve been tracking them to see how they compare to the ratings given by raters that the APA uses for their official ratings.
This is how my automatic system works. I calculate the ratings of the two sections playing and compare those ratings with the goals those sections scored. I end up with a number, either positive or negative showing that the section under-performed or over-performed their rating in that game. I do those calculations for all the games in the tournament. Then I add all the individual game performance scores for each section. I end up with a performance score showing how the section performed over all their games. Then I see which section’s performance is farthest away from their rating, either most over-performed, or most under-performed. I change the total rating of that section to move them 1 point closer to their actual performance. When I change that 1 rating, it also changes the performance scores of all the teams they played against, and there is a completely different performance score for all the teams in their division. With those new performance ratings, I see which section is now farthest from their performance, change them one point and continue changing the section that is farthest from their performance until I get the performance scores as close to zero as possible. Once the performance scores are as close to zero as possible each section’s rating will be exactly how they performed in that tournament. Explaining this makes it sound complicated, but I have developed a macro that does all the calculating semi-automatically.
If a section performed close to their rating, I leave them alone. If a section performed 3 points or more above their rating, I raise the rating of each player on the section 1 point. If a section performed 3 points or more below their rating, I lower the rating of each player on the section 1 point.
Zero is used as the standard, so if the performance of a team is negative, all the teams in that division are raised so that the lowest rated team is not lower than zero.
Then those are the ratings that are used for the next tournament. I only move a player no more than 1 point at each tournament.
The APA only uses ratings to seed teams in the correct division in tournaments. When a change is made with this automatic system, all the players on the section are changed. The section’s entire rating ends up being exactly accurate, but the individuals on the section could be off. One player’s rating may be too high if another player’s rating is too low. If those same 3 players always play on the same section, the individual’s ratings won’t be corrected, but since we only use the team’s rating for seeding teams, the team’s rating will still be correct. If any of the 3 players plays with different players, then the individual ratings get corrected also.
We have always had people complaining about the ratings, and this automatic system solves a lot of the problems people complain about.
People would complain that their rating is too low. This system automatically raises their rating if they sufficiently over-perform their previous rating.
People complain that the raters are biased in their ratings. This system automatically adjusts the ratings, and there is no subjectivity. I have seen instances where the raters raise a player’s rating when the player’s section under-performed their previous rating. These kinds of “mistakes” never happen with the automatic system. No one is raised unless they sufficiently over-perform their rating; no one is lowered unless they sufficiently under-perform their rating.
People that are assigned to do the ratings complain that they have to watch all the players, and that they have to meet to discuss the ratings. This system adjusts the ratings automatically without anyone having to watch, or meet. Someone just needs to copy information from the score sheets into a spreadsheet program and the new ratings will be calculated automatically.
The automatic system can be used with a 10 point system, a 20 point system, or any other numbering system that anyone wants to use. You start with the ratings that you already have and the adjustments are made to those ratings. The US was on a theoretical 20 points system when I started using the automatic system and the highest US rated player at that time was a 16.
There are a couple ways to deal with a visiting player who doesn’t have a US rating. The method I have been using is that I ask the visiting player’s hosts to give a best guess as to what they think his rating should be, and I use that. The other method would be to just assume that the other two players on his section are correct and assign him the rating that would make his section’s rating add up to the rating to which they performed minus 2. I subtract the 2 to make sure they earn their rating, and they wouldn’t have been assigned the rating unless their section performed 3 points over their rating.
Another benefit of this automatic system is that the ratings are universal. Women or juniors don’t have to have a different rating when they play in a women’s or junior division. Players from any country using this system (assuming all the countries use the same number of rating points in their system) would be rated uniformly no matter what country they played in.
If anyone is interested in using this system, further developing it, or just experimenting with it, I have an Excel file that is fairly easy to use. You just list the sections with the players and their ratings, then list the scores each section scored from the score sheets. When all the games are entered you click on the macro button until it calculates all the sections’ actual performance, then look in another area of the spreadsheet and see what the new ratings are. E-mail me at pjpxmcd@earthlink.net if you would like for me to send you the spreadsheet.
Paul Johnson
Cell Phone 512-698-6827
Click here to Check out our horses
Here’s an article about my automatic rating system.
I have developed a system to automatically rate players using information from the score sheets only. I have been using it since October, 2008, for all the US tournaments that I attend, or that I can obtain the score sheets, and it keeps all the players’ ratings current without making any rating mistakes. The APA doesn’t use it for their official ratings, but I’ve been tracking them to see how they compare to the ratings given by raters that the APA uses for their official ratings.
This is how my automatic system works. I calculate the ratings of the two sections playing and compare those ratings with the goals those sections scored. I end up with a number, either positive or negative showing that the section under-performed or over-performed their rating in that game. I do those calculations for all the games in the tournament. Then I add all the individual game performance scores for each section. I end up with a performance score showing how the section performed over all their games. Then I see which section’s performance is farthest away from their rating, either most over-performed, or most under-performed. I change the total rating of that section to move them 1 point closer to their actual performance. When I change that 1 rating, it also changes the performance scores of all the teams they played against, and there is a completely different performance score for all the teams in their division. With those new performance ratings, I see which section is now farthest from their performance, change them one point and continue changing the section that is farthest from their performance until I get the performance scores as close to zero as possible. Once the performance scores are as close to zero as possible each section’s rating will be exactly how they performed in that tournament. Explaining this makes it sound complicated, but I have developed a macro that does all the calculating semi-automatically.
If a section performed close to their rating, I leave them alone. If a section performed 3 points or more above their rating, I raise the rating of each player on the section 1 point. If a section performed 3 points or more below their rating, I lower the rating of each player on the section 1 point.
Zero is used as the standard, so if the performance of a team is negative, all the teams in that division are raised so that the lowest rated team is not lower than zero.
Then those are the ratings that are used for the next tournament. I only move a player no more than 1 point at each tournament.
The APA only uses ratings to seed teams in the correct division in tournaments. When a change is made with this automatic system, all the players on the section are changed. The section’s entire rating ends up being exactly accurate, but the individuals on the section could be off. One player’s rating may be too high if another player’s rating is too low. If those same 3 players always play on the same section, the individual’s ratings won’t be corrected, but since we only use the team’s rating for seeding teams, the team’s rating will still be correct. If any of the 3 players plays with different players, then the individual ratings get corrected also.
We have always had people complaining about the ratings, and this automatic system solves a lot of the problems people complain about.
People would complain that their rating is too low. This system automatically raises their rating if they sufficiently over-perform their previous rating.
People complain that the raters are biased in their ratings. This system automatically adjusts the ratings, and there is no subjectivity. I have seen instances where the raters raise a player’s rating when the player’s section under-performed their previous rating. These kinds of “mistakes” never happen with the automatic system. No one is raised unless they sufficiently over-perform their rating; no one is lowered unless they sufficiently under-perform their rating.
People that are assigned to do the ratings complain that they have to watch all the players, and that they have to meet to discuss the ratings. This system adjusts the ratings automatically without anyone having to watch, or meet. Someone just needs to copy information from the score sheets into a spreadsheet program and the new ratings will be calculated automatically.
The automatic system can be used with a 10 point system, a 20 point system, or any other numbering system that anyone wants to use. You start with the ratings that you already have and the adjustments are made to those ratings. The US was on a theoretical 20 points system when I started using the automatic system and the highest US rated player at that time was a 16.
There are a couple ways to deal with a visiting player who doesn’t have a US rating. The method I have been using is that I ask the visiting player’s hosts to give a best guess as to what they think his rating should be, and I use that. The other method would be to just assume that the other two players on his section are correct and assign him the rating that would make his section’s rating add up to the rating to which they performed minus 2. I subtract the 2 to make sure they earn their rating, and they wouldn’t have been assigned the rating unless their section performed 3 points over their rating.
Another benefit of this automatic system is that the ratings are universal. Women or juniors don’t have to have a different rating when they play in a women’s or junior division. Players from any country using this system (assuming all the countries use the same number of rating points in their system) would be rated uniformly no matter what country they played in.
If anyone is interested in using this system, further developing it, or just experimenting with it, I have an Excel file that is fairly easy to use. You just list the sections with the players and their ratings, then list the scores each section scored from the score sheets. When all the games are entered you click on the macro button until it calculates all the sections’ actual performance, then look in another area of the spreadsheet and see what the new ratings are. E-mail me at pjpxmcd@earthlink.net if you would like for me to send you the spreadsheet.
Paul Johnson
Cell Phone 512-698-6827
Click here to Check out our horses
From: ahicks@polocrosse-international.org [mailto:ahicks@polocrosse-international.org]
Sent: Wednesday, March 03, 2010 11:23 AM
To: pjpxmcd@earthlink.net
Subject: IPC
Hi Paul
Working on more articles for the IPC web site, wondered if you would like to contribute?
Allen Hicks
RR#1 Site 4 Box 10
Millarville AB T0L-1K0 Canada
Phone: 403-931-3610
Cell: 403-816-5251
E-mail: info@polocrosse-international.org
Web: www. polocrosse-international.org
No virus found in this incoming message.
Checked by AVG – www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2752 – Release Date: 03/19/10 01:33:00