APCR (ASL Player Comparative Ratings)

[Mister T]

How do I opt out so no game I play is recorded? -- jim

I don't think you can opt out as it would impact your opponent(s). However it should be feasible to opt out from publication of your name, using something else (like sparafucil ).

 

[Sparafucil3]

I don't think you can opt out as it would impact your opponent(s). However it should be feasible to opt out from publication of your name, using something else (like sparafucil ).

I should be able to opt out. I didn't want to be in the old AREA. I didn't want to be in Aaron's APCR, nor do I want to be in this one. I don't see why I have to be. I should be able to opt out.

Absent that, I should be allowed to make my personal rating private. Then, you can have the information, but I decide who get's to see it and use it. -- jim

 

[DougRim]

How do I opt out so no game I play is recorded? -- jim

1. Don’t play in tournaments.

2. If you do play in tournaments, ask the TD not to forward your info to the ASL PLayer RAting system.

3. One the site is developed remind us that you want to opt out.

 

[DougRim]

We can have more discussion about the mechanics of this as we go forward but I fully agree that people should not be forced to be part of this.

 

[Stewart]

All

I am in the process (well advanced) of building a very generic ASL player ratings engine based on the GLICKO Methodology
and covering both friendly and tournament play.

The GLICKO methodology is an improvement over the original ELO (Chess Ratings system)

http://glicko.net/glicko/glicko.pdf

It is designed to take account not only of the ratings difference of players when computing Rating changes but also the deviation (Volatility) of a players rating.
It is designed for computation of rating over a period rather than game by game (I have chosen 2 month periods as per the papers recommendations).

So, if you play 10 games in a weekend and win them all vs 10 games over 3 months and win them all....

How does it differ and WHY should it differ?

 

[Stewart]

At the very least, no matter what, AREA ratings must be updated to include the last two years of results, so we have a benchmark to compare those results with this new method based on chess, which, by the way, has no random element, so I'm not convinced that it is better by any means.

I am sure everyone would like to see what the new ratings would be after being transmogrified by this new method. Wouldn't it be interesting if Steve Pleva was all of a sudden ranked lower from #1?

No, we absolutely need AREA as a benchmark.

No reason NOT to use the "current" AREA ratings, Pleva not being #1 is meaningless as he'd just move accordingly.

 

[Stewart]

You’re incorrect, there always is a player one and a player two. Player one is always the first firer/mover so player one is always the attacker.

Been that way since ‘85

This is not true.
U20 Germans move first.....Germans are defending.

 

[Stewart]

In the Glicko note Aaron disseminated, it is written: "the Glicko system works best when the number of games in a rating period is moderate, say an average of 5-10 games per player in a rating period".

As the rating period selected was two months, it would mean that it would suit players with around 50 tournament games per year, which is clearly too high for an overwhelming majority of players.

For robustness check purposes, i would suggest to extent the rating period to 6 and 12 months and see whether it makes any difference. One needs to test the algo under different specifications before calling it "open".

LOL so what this list does is exclude regional players that dont have access to tournaments during the year. West coast tournaments are few and far between.

What about just using the numbers from ROAR and continuing that trend.

Otherwise, its a Have and have not listing.

 

[Sparafucil3]

1. Don’t play in tournaments.

2. If you do play in tournaments, ask the TD not to forward your info to the ASL PLayer RAting system.

3. One the site is developed remind us that you want to opt out.

1) Why should I NOT play tournaments? I only know of one tournament that uses this system to rank players (Bitter Ender). I think the one in Oslo may as well. If I want to attend a tournament, I shouldn't have to decide not to because my name may appear in your data system. You should build a way to either opt out (might be hard), or better yet, a way for me to log in and opt to keep my records private. Allowing me to keep my playing's private would allow you to still collect the data but allow me to keep things out of the public eye. It's well known that I am a shite player, I have nothing to hide or run from.

2) I have asked to be opted out, yet there I am. I think the issue is that if I opt out, then EVERY player who plays me has to annotate their record that I opted out too or the information gets put in and aggregated anyway.

3) Will do. I also sent a note to Sam Tyson asking for the same thing.

My biggest complaint is ASL is not chess. There is no way to account for all the vagaries that make up ASL. What if the scenario being played is a dog? What about the luck factor? Even within a tournament, not every pairing playing in the first round is playing the same scenario. Even if they are playing the same scenario, the setups vary, their understanding of the objectives vary, etc. Even if their understanding is consistent across every pairing, the dice will not be consistent. These sorts of things work well in chess because EVERY player playing white is playing the EXACT same scenario as EVERY other player playing white. Every player playing BLACK is playing the same EXACT scenario as EVERY other player playing black. There is little luck in chess beyond selecting sides, there is only good play and better play. In ASL, you can play a great game and be fucked by dice. You can also play a bad game and be carried by dice. In the long run, dice even out but in the course of a single tournament-sized scenario, luck can play a non-zero sized factor. IMO, trying to apply these types of rating systems to ASL is a fools errand and I want little part of it. I hope it works out for the sake of the effort you're putting into it and wish you luck. -- jim

 

[STAVKA]

All, after a mountain of work all AREA results are in (including those with no winning Side or scenario Code)
along with a smattering of other tournie results, 27252 games of which ~ 27000 are from AREA

Tournie Results from Jan 1 2020 will require these, it seems unfair to penalize tournies from the first part of the year If the requirement wasn't known.

Recording Friendly games will always require winning side and scenario.

I will be pounding through some results send to me by various people over the next few days and dreaming up some reports.

I have added a blended rating to the outputs here.

Rating Deviations range from 350 to 30 depending on frequency of play and consistency of results 190 is the mid point of 350 an 30, so 190
will be added to people rating and their RD subtracted, So a low RD acts as a bonus, most of the top 20 players this has very little net effect,
again at this point the blended rating is just an idea (though I think a good one from George T).

So the 298 qualified players (RD < 300 and Games played >49 are as below (Ranked by Blended rating), A supplemental report at the end is the top 100 players with 20-49 Games.

P.S. Melvin owes me some raw spirit ?

Rank	Name	Rating	RDeviation	Games Played	Blended Rating
1​	Melvin Falk	2156​	193​	287​	2153​
2​	Bret Hildebran	2055​	188​	152​	2057​
3​	Brian Wiersma	1982​	142​	221​	2030​
4​	Gary Fortenberry	2052​	227​	153​	2015​
5​	Kenneth Young	2016​	223​	67​	1983​
6​	Toby Pilling	1987​	202​	127​	1975​
7​	Phil Palmer	1968​	188​	80​	1970​
8​	François Boudrenghien	1966​	200​	199​	1956​
9​	Michael Hastrup-Leth	1956​	191​	177​	1955​
10​	Jim Taylor	1972​	212​	134​	1950​
11​	Steve Pleva	1982​	223​	183​	1949​
12​	Aaron Cleavin	1905​	150​	198​	1945​
13​	Bill Cirillo	1943​	193​	100​	1940​
14​	Paul Sidhu	1923​	175​	242​	1938​
15​	Gary Mei	1929​	190​	118​	1929​
16​	Miguel Guerrero	1977​	240​	76​	1927​
17​	Darren Kovacs	1932​	200​	65​	1922​
18​	Rich Domovic	1867​	143​	400​	1914​
19​	Lionel Colin	1926​	212​	50​	1904​
20​	Mattias Rönnblom	1907​	196​	235​	1901​
21​	Dave Ginnard	1868​	157​	155​	1901​
22​	Ken Mioduski	1917​	217​	123​	1890​
23​	Paolo Cariolato	1907​	207​	70​	1890​
24​	Ron Duenskie	1884​	188​	119​	1886​
25​	Richard Spilky	1857​	168​	126​	1879​
26​	G. Tournemire	1866​	177​	256​	1879​
27​	Peter Struijf	1867​	183​	247​	1874​
28​	Bo Siemsen	1879​	196​	111​	1873​
29​	Hennie van der Salm	1845​	168​	350​	1867​
30​	Wes Vaughn	1838​	163​	355​	1865​
31​	Sean Deller	1869​	197​	67​	1862​
32​	Will Fleming	1708​	39​	238​	1859​
33​	JR Tracy	1896​	227​	137​	1859​
34​	William Stoppel	1863​	199​	70​	1854​
35​	Bob Bendis	1838​	189​	250​	1839​
36​	Randy Yeates	1879​	235​	51​	1834​
37​	Jim Bishop	1839​	198​	102​	1831​
38​	Arnaud Sanchis	1853​	214​	59​	1829​
39​	Jim (jr.) burris	1804​	175​	162​	1819​
40​	Neil Stanhagen	1857​	245​	92​	1802​
41​	Michel Bongiovanni	1826​	214​	128​	1802​
42​	Fabrizio Da pra	1882​	271​	56​	1801​
43​	Chris Chapman	1807​	197​	50​	1800​
44​	Sebastian Hummel	1821​	213​	86​	1798​
45​	Andrea Pagni	1802​	197​	126​	1795​
46​	Andy Hershey	1842​	239​	64​	1793​
47​	Carl Nogueira	1778​	180​	196​	1788​
48​	Scott Martin	1766​	169​	50​	1787​
49​	Chris Mazzei	1793​	197​	201​	1786​
50​	John Knowles	1694​	104​	66​	1780​
51​	Alan Smee	1739​	152​	73​	1777​
52​	Olav Heie	1823​	238​	62​	1775​
53​	John McDiarmid	1775​	191​	72​	1774​
54​	Martin Svärd	1881​	297​	135​	1774​
55​	Christian Herde	1804​	225​	76​	1769​
56​	Mike Klautky	1802​	224​	68​	1768​
57​	Guillaume Mailhan	1841​	264​	87​	1767​
58​	David Longworth	1770​	193​	61​	1767​
59​	Dave Lamb	1827​	251​	90​	1766​
60​	Dan Stanhagen	1849​	276​	52​	1763​
61​	Brian Routh	1744​	174​	62​	1760​
62​	Steven Frum	1805​	237​	55​	1758​
63​	Patrik Manlig	1780​	216​	89​	1754​
64​	Jes Touvdal	1764​	200​	102​	1754​
65​	Robert Feinstein	1787​	224​	103​	1753​
66​	Eric Bongiovanni	1792​	232​	171​	1750​
67​	Randy Shurtz	1722​	165​	91​	1747​
68​	Enrico Catanzaro	1748​	191​	127​	1747​
69​	Paul Works	1723​	169​	84​	1744​
70​	Paul Lauger	1740​	197​	98​	1733​
71​	Mattias Bergwall	1702​	166​	123​	1726​
72​	Curtis Brooks	1735​	199​	83​	1726​
73​	Paul Haseler	1733​	197​	71​	1726​
74​	Andreas Carlsson	1721​	186​	220​	1725​
75​	Eoin Corrigan	1738​	207​	72​	1721​
76​	Boudewijn Van schalkwijk	1790​	262​	71​	1718​
77​	Sam Henze	1739​	212​	51​	1717​
78​	Rick White	1754​	227​	53​	1717​
79​	Bill Hayward	1693​	169​	194​	1714​
80​	Philippe Briaux	1720​	198​	347​	1712​

Thanks Aaron, I am sure you got the numbers right, it was expected ?

 

[The Magnus]

1) Why should I NOT play tournaments? I only know of one tournament that uses this system to rank players (Bitter Ender). I think the one in Oslo may as well. If I want to attend a tournament, I shouldn't have to decide not to because my name may appear in your data system.

The tournament in Oslo was held in 2004-2008. It has not been held since, and all records from that tournament have been lost so I cannot tell if they used AREA or not (I know because I gathered Meta Data for Aroon's APCR and the TD told me he did not have any records any more).

All major Scandinavian Tournaments (Copenhagen, Borås, Friendly Fire) use AREA for the seeding and/or reporting results of players.

DougRim says:
2. If you do play in tournaments, ask the TD not to forward your info to the ASL PLayer RAting system.

I would be OK with this if it is made in a such a way that it does not brake the integrity of the data in AREA/APCR. All games have to be reported though, or the integrity of the data would be compromised. If it is enough to call you "Player X", I am fine with that. But I believe AREA (as well as APCR) uses the differential between the rating of the two players to calculate the incremental change of your individual ratings, so a "Player X" entry will not work as his AREA rating would not be known.

I do not believe that this is a major issue, though. When Aaron started to develop APCR he used ALL of my gaming data (tournament AND friendly) to debug his system. When he published the very first outputs from that system, ONE player contacted me and asked me rather rudely why I had reported our two friendly games to Aaron. I proceeded writing to all players I had played with and had a valid email address to, asking if they had a problem with me reporting friendly games to APCR. No other player objected, and the one player who originally objected was told that I cannot play him anymore as I intended to report all games to APCR, including friendly ones. So you might be relatively alone, Jim (sorry).

 

[STAVKA]

The AREA/APCR is a great tool both for TD to seed players and find scenarios used in tournaments.

Another important aspect is its function as a memory bank for players - of who and what you have played.

 

[The Magnus]

Thanks Aaron, I am sure you got the numbers right, it was expected ?

There were some bugs in the beginning .

Seriously, though, I do believe that Aaron's more complex algorithm, with all its various weightings and time dependencies, more accurately tracked a person's ASL abilities than the simple AREA algorithm did. After all, there is a reason it is the standard in the chess community. As Jim correctly states, ASL is not chess, and there will be randomness that does not exist in Chess, but in the long run the APCR ratings will probably be a good approximation of your ASL skill. Mine seems somewhat correct .

 

[Sparafucil3]

The tournament in Oslo was held in 2004-2008. It has not been held since, and all records from that tournament have been lost so I cannot tell if they used AREA or not (I know because I gathered Meta Data for Aroon's APCR and the TD told me he did not have any records any more).

Sorry, I said Oslo but meant Copenhagen. Mea Culpa. -- jim

 

[Sparafucil3]

No other player objected, and the one player who originally objected was told that I cannot play him anymore as I intended to report all games to APCR, including friendly ones. So you might be relatively alone, Jim (sorry).

As I said, one other option would be to allow me to make my standing and playings private. You could report all you want, the system could rate me, but my playings would remain in the dark.

WRT reporting "friendly games": is there something in the system which limits the impact of playing the same person over and over? I would also add that anything where a player can input data himself or herself is subject to abuse. If this system is intended to be remotely accurate, not everyone should be able to enter pairings and playings. The system is only as good as the data put into it. Ask @Nadir_E what happened to the ASL players map when that was open.

Absent that, you not playing me is as much your loss as mine. I will regret it, but c'est la guerre. -- jim

 

[DougRim]

1) Why should I NOT play tournaments? I only know of one tournament that uses this system to rank players (Bitter Ender). I think the one in Oslo may as well. If I want to attend a tournament, I shouldn't have to decide not to because my name may appear in your data system. You should build a way to either opt out (might be hard), or better yet, a way for me to log in and opt to keep my records private. Allowing me to keep my playing's private would allow you to still collect the data but allow me to keep things out of the public eye. It's well known that I am a shite player, I have nothing to hide or run from.

I completely agree with you that you, and anyone else, have the right not to appear in any Rating System. I appreciate your willingness to allow your data to be held in the system but not made visible. That is very doable.

2) I have asked to be opted out, yet there I am. I think the issue is that if I opt out, then EVERY player who plays me has to annotate their record that I opted out too or the information gets put in and aggregated anyway.

You may currently appear in the test data on the new development site. That is an unfortunate oversight in the transition from one tool (AREA) to another. Now that I know that you don't want to be visible, Sam and I can ensure that you are not.

These comments are very helpful in terms of identifying issues that need to be addressed.

 

[Sparafucil3]

I completely agree with you that you, and anyone else, have the right not to appear in any Rating System. I appreciate your willingness to allow your data to be held in the system but not made visible. That is very doable.

Thanks. I think this strikes a balance of making the system as accurate as possible while ensuring people can be outside the system as well. Much appreciated. -- jim

 

[The Magnus]

WRT reporting "friendly games": is there something in the system which limits the impact of playing the same person over and over?

Absent that, you not playing me is as much your loss as mine. -- jim

Well, I cannot speak for the new system developed by Doug, but Aarons idea was:

• Reporting friendly games would (of course) be voluntary.
• There would be two variants of every report, official tournament games only or all games including friendly. TD's would of course use the tournament only7 version for seeding.

It would of course be possible to slant the rating which include friendlies, if for example Melvin played only me his all-games-rating would be even more impressive . But the tournament-only-rating would be unbiased in this regards.

If it is possible to report friendlies in the new system, it of course has to be voluntary (who could force you ). I would report my games in order to have them centrally reported, but that is just my preference.

 

[DougRim]

There were some bugs in the beginning .

Seriously, though, I do believe that Aaron's more complex algorithm, with all its various weightings and time dependencies, more accurately tracked a person's ASL abilities than the simple AREA algorithm did. After all, there is a reason it is the standard in the chess community. As Jim correctly states, ASL is not chess, and there will be randomness that does not exist in Chess, but in the long run the APCR ratings will probably be a good approximation of your ASL skill. Mine seems somewhat correct .

I think Aaron did some great work. Unfortunately, I do not have access to it at this time so this current project is building off of the AREA data and tools. Hopefully as time goes by we can reconnect with Aaron and take advantage of what did, one way or another.

 

[Sparafucil3]

I think Aaron did some great work. Unfortunately, I do not have access to it at this time so this current project is building off of the AREA data and tools. Hopefully as time goes by we can reconnect with Aaron and take advantage of what did, one way or another.

Aaron is on VASL quite a bit. He plays under his own name. If nothing else works, drop into VASL from time to time and see if he is on and playing. -- jim