Hi everybody.
I'm not Shure if this IS the correct place but I found no place on Internet where I could reach the father of Bayeselo.
First thing to know, I have fistfull of years in studying by myself only some méthods which are the best known : I had to manipulate matrices withthé Massey méthod , Poisson distributions had me to make my own Poisson Distribution on m'y casio graphic calculator... waiting thé next train while mine had machine troubles ...In 2 jours thé basic program began to born...Later I found that m'y graphic tableur Excel -like worked faster in sheet mode than his basic but easy high-level basic casio... Now all m'y formulas s are on sheet (liké open office or Excel) .Programs mode are slow, but I typed basic programs when I was 10 years , Always drawned nice things in graphic polar mode..
Before m'y casio graphic (even if I have Excel sheets Aldo on PC), I used DC MOTO, a Thomson french computer of the 80's 6bits processor6809 which allows more than thé MO/TO which were under-exploited.
WHY DCMOTO ? It's one of the best , if not thé best,THOMSON emulator of the Basic Microsoft which I was sofamiliar with...some Line command on Microsoft II kiné of c:/ list/ help..
And , for chose who remember thé game Numero10 , which was good animated foot game, althought thé IA of the computer was perfectible...Traditionnally little program in basic launched a waiting page before thé game itself in binary low level interprétation 'assembleur', thé more easy for a computer, thé harder to understand for an human...Hère it's not this football game itself but thé awaiting page which was in binary, but Numéro 10 had a little thing more : a .GTI image format de f Michel Platini (hé was mister n°10 ), caution for thegame (thé best on Thomson and even some Amstrad fans admitted " yeah not Bad"...
I Come back to my basic program on 6809 : I had some formulas which , I layer knew that a part was used, but now way I was on thé rightt path, and even use no more since an estimator radier and implémented with régression gives beter results...three quarters of the saison thé stats are quote enough pertinent for thé one who keeps linearrity.
For m'y basic program on MO/TO, I obsiously choose th . GTI format image of Michel PLATINI to make wait (it was no use) and for thé pleasure to have this Graet football payer while I typed datas when thé programme asked me datas.
That was m'y first coding for making oddsandtips. M'y Casio was smart in a pocket everywhere sh'd go I still use her but non more programs all on sheet excel. Estimator of nombre of goals, linear régression (m'y home own cooking) or vector/Matrix, it could even, If Ihad Time enough go to thé Skellam distribution, via Bessel they can be on thé machine.
Oh m'y apologies I write a bit too much, more for inteoducing myself than other.
I used a lot Elo (hockey sur glace, tennis ( mixed with a ratio of Victory by match played and on particuliar surfaces )and of course thé Platini's sport, football (elorating for nationals sélections, clubelo for club -only-european -)..
Elo IS simple (if tpu don't have a Big list to computer everyday) to understand and to use. Thé only trouble that it don't cares of draw match, which are 1/3 of issues.
On a Chess site (Abrobecker.free.fr ,with memoryI will give tout link layer if asked, a Chess fan) I dig a formula that , fonction of the difference of Elo coefs, gives thé probability ofa tiède , draw match, for Chess but works for football too...
I had to modify some coefficients of the formula of draw with Elo becuse if in Chess game 25% of a rien IS the maximum (and 20% thé maximum checked), in football WE need a probability of draw higher than 30%, 1/3 of gales observed IS normal.
.I Always leak m'y sources if I had no autorisation I give the name the link, anyway the coefficients for football have ssized by me according résultats of football.
And thé I Heard of Bayeselo....I knew but not feeling strong enough to use it. One dayI came back to it, downloader it, triés to understand maximisation etc...
First I downloader another thing on thé site in links, which could have helped me, and, althought I'm not a Chess fan ( play yes but watch no more than this) I downloader thé Portable Game Notation .PNG on a site which looked serions ( FIDE maybe) o fb thé game Fischer/Kasparov, one o the most inglorious.
And then , nothing Bayelso can't read...I triés to create one with few football matches and nothing more
It was last year, severals years to try it, not mypriority bit when I have thé Time I open thé case
So my questions are, I Would be glad ro h e a constructive answers:
-Can WE really simulate a football ligue with 20teams given their Elos ( m'y goal IS to compare thé beginning of thé Season to the actual current Elo)...
-Do we must typing on a .pgn all 380 matchs of the Season or today thé 300 matchs played before that day?
-WE CAN add players or team and attribuate them an Elo Coef . in thise case, iss there a model of .pgn which allows input 20 teams and their Elos, tgis Would be easie than make data of 300 matchs adding thé 80 which left...
I Would be, if this s possible, glad if a model of .pgn makes BayesElo working and computing thé end of Season of Ligue2 french (yes Hope you'll excuse m'y english I'm french born in Caen whose professional club Plays L2)...
To sum on up how IS itpossible to simulate a championship with 20 clubs (I have notes the Elo coefs at thé biginning and current day IS given by clubelo.com)
My .pgn of Fisher Kasparov game IS maybe corrupt or have Bad nomenclature I had , according to thèse notations laws triés everytging m'y download of Bayeselo don't recognize the .pgn..
Sorry for this long post, I i'troduced myself, as usual on forums for a first post, Hope somebody Would give me more précision in the case of round Robin tournament with 20 players. Chess or football it's thé same méthod..
Further one of m'y goals Would be to , starting with thé Elo coefs, giving BayesElo thé possibilité, if not yet, to take thé tièd draw matchs, giving 1 point to the teams, when thé probability of a match are P(1)=50 P(N)=30 P(2)=20...Hère if 1 Would win with40% WE have 30% la f draw IS non negligeable...and more with 30/40/30 which leads to no winner and with m'y coef of thé Abrobecker formula thé matchs which reaches 45%oftied draw have often point-shared isses...
In Chess where thé Abrobecker formula fibres 20% to 2 5% maxiIS quite really accurate but 'not adapted to soccer. I did a little change , by empirism...
Making thé Elo computing with such a formula could help BayesElo to bé more compétitive in forevasting football xhampionships. But its not m'y first step.
Any tips ? All m'y .pgn are not recognize by BayesElo.
If Only I could compute like exemples on the site I download it , I'd have make a Big step. I'm not in hurry 1 ou 2 rimes père year I try but since this Time I mist admit that I have to Ask spécialiste.
Have a nice day, or a good night.
Thank tout for attention.
Bayesian Elo rating
-
Rémi Coulom
- Posts: 219
- Joined: Tue Feb 12, 2008 8:31 pm
- Contact:
Re: Bayesian Elo rating
Hi,
Thanks for your interest in bayeselo.
https://www.remi-coulom.fr/Bayesian-Elo/
See the "usage documentation" section.
https://ia902908.us.archive.org/26/item ... -03-12.txt
Thanks for your interest in bayeselo.
bayeselo can simulate a tournament, but not with elo ratings that are input by hand. Ratings must have been computed by bayeselo from data. An example is given on the web page of bayeselo:Can WE really simulate a football ligue with 20teams given their Elos ( m'y goal IS to compare thé beginning of thé Season to the actual current Elo)...
https://www.remi-coulom.fr/Bayesian-Elo/
See the "usage documentation" section.
If you have this data in a computer-readable format, you could write a script to convert it automatically.-Do we must typing on a .pgn all 380 matchs of the Season or today thé 300 matchs played before that day?
You cannot set elos manually in bayeselo. They have to be computed from data.-WE CAN add players or team and attribuate them an Elo Coef . in thise case, iss there a model of .pgn which allows input 20 teams and their Elos, tgis Would be easie than make data of 300 matchs adding thé 80 which left...
Please include a sample of your pgn data, how you read it with bayeselo, and the error message you get. You can also check the PGN specs there:Any tips ? All m'y .pgn are not recognize by BayesElo.
https://ia902908.us.archive.org/26/item ... -03-12.txt
Re: Bayesian Elo rating
In first, I apologize for my late answser which is not at the height of the quick response that followed my post.
I dont work on this file every day, to be true it's when I have time and enough concentration.
First, my PGN that didn't work correctly well, I didn't find it...IN First I maybe throw in the deletin corbeille...I could have made a bad command line to make it work...cause I had this for triing with the best that seems to be fine...BY memory I talked about Fisher Kasparov : in fact I didn't find any kind of PGN of a game between these 2 kings of chess...Fischer Spassky is easy to find and it's maybe the exemple I tried to find and make working on BayesElo... Although I remember a non-successfull PGN on BayesElo reading...As goodest can be my memory I though maker read a well-known PGN was a good try to begin. I sometime read English papers on ACADEMIA, on Bayesian Inference to familiarise myself with it and I see the light in a long tunnel.
I give some time to my work, a few times in a month, so I apologize for answer later than the quick response you had the kindness to give to this post...
Having just a BAC D ( science graduate before university in France ), I understand basis, This Winter I worked a lot on Matrix that were not to teach at this level, but understant enough to practice the Massey method, for exemple...But the linearity -or not-of matrix are a bit understanded, I work on this winter.
So a silly thing I wrote in my first post is that a PGN of Fisher Kasparov existed,I just confused the one of the most known game between Fischer and Spassky....
The PGN I had is hiden on a hard drive or my adaptation attempt was unsuccessfull and I threw it.
If I didn't answer as quickly as your response is that I work on my project when I have time ans I wanted to come back with interessant things to expose.
First I downloaded Little Blitzer, Arena and Little though...Chess Motors, after inquiries -and long searchs- to have the possibilities to arrange a round robin tournament -like Ligue2 freanch-, Areana is to long and re-set up the Elos at 2000 although Idis a file with teams names and Elos...Here, my goal was to create a PGN with all teams, with their own Elos...
Before that, I could, with the copy paste fonction on the BayesElo interface , finally read a PGN file and make some litlles command line. Unfortunatly, if I tried the way to put succesfull in the BayesElo all the teams (20 or soon 19 or 18 but it's another story), the superiority matrix ( los) ajust all the team the same Elo and so they have an uniform probability to win or loose...Arena, even if I copy paste names and Elos, set the Elo at 2000. So find equality of win the chamionship.
This summer my not so relentless work was principaly to tried to copy paste teams names and to familiarise myself with all the command lines ( ? helped me ).
But I have another idea : to see the source code. In first I'd like to understand a little bit more the MM ( maximisation minimisation algo ) and I have to confess that if I see a liitle bit more, I'm not very familliar with this is to my point of view my next work...I understand a lot on matrix this winter...next stage will be MM.
I already talk about my little level in programating...My best idea of this summer was to have a look at the source code...ESpecially the CCP file CEloRatingCUI that give the HELP for some of fonctions that seems to set players and elo (I don't want to type on my Dell the almost 380 of a championship, ...SOme of them have already been played it's the 8th journey, the Elo evoluate and it can be intersesting for me to comparate the first days of the championship Elo's and tthe current Elos ).
To resume this summer I pratcice the command line with a PGN I did with the names of the Elos to be more ease ith the command line...But the elos are the same for all, so they have the same probability to win or have any other place in the championship.
To have abetter understanding of what BayesElo and his many CMD can bring to y project ( I don't want to enter manually all the matches on a PGN, too long) the source code open lot of commands that could be interesting...
inparticularity :
elo [p] [elo] ... get[set] Elo of player number p ; if a player has a number ( 1 to 20 in a football championship ) , I understand that his Elo can be set. Is it true ?
"offset [elo [pl]] here,can we offset a particular Elo for a given player (with his number ) ?
p <w> <b> <r> ... probability of result <r> with Elos <w> and <b> here I don't understand what are <w> <b> et <r> wich variable do they belongs?
adavantage of playing first can be an Home Field Advantage in football...I can make a mstake...Feel free to tell me if I'm wrong
drawelo [x] ..... get[set] draw Elo here, is it the points wins with a draw or the standars number of draw expected ? or is it :
prior [x] ....... get[set] prior (= number of virtual draws...
I have , with the bloc note, make a PGN textfile with all the teams and their elos but my first problem is that the Elos are, finally the same foreverybody (I know the Ligue2 is quite balanced);..even in indicate Elos on PGN ( I had read the long PGN Bible, before write here, that you had the kindness to link and it is the real bible for PGN, I wish to thank you for that.
Now I can open a PGN in Bayelo, not have the patience that Arena make a round robin with almost 20 teams, but can make a PGN with teams names or position (p1, P2, pN,..), my question that make me more quik working is that is it possible to attribute an Elo for a given team...At this moment, I would like to compare the beginning of the championship with the current Elos, which have evoluate...
Is there another code source file that ould help me? The EloRating seems to be a good path, finaly my question is can I attribute an Elo to A team ( I can make it on a Bloc Note and paste it on a elorating command) and the others teams of the divisions? Will I have a superiority matrix and and final classement after each team played both advantage home away, with all teams ?
I'm sorry for my english, and above, if my question have no or too complex answer. I'll try to read all the source code, to have more info on the BayesElo helpe list.
Anyway, THank you very much for the time you agreed for me. I just tried to make BayesElo working for football, especially the Ligue2 in France, where Stade Malherbe Caen (were I was born) plays. I bet that in Elite leagues, the process could have much success, but it's not my goal.
I dont work on this file every day, to be true it's when I have time and enough concentration.
First, my PGN that didn't work correctly well, I didn't find it...IN First I maybe throw in the deletin corbeille...I could have made a bad command line to make it work...cause I had this for triing with the best that seems to be fine...BY memory I talked about Fisher Kasparov : in fact I didn't find any kind of PGN of a game between these 2 kings of chess...Fischer Spassky is easy to find and it's maybe the exemple I tried to find and make working on BayesElo... Although I remember a non-successfull PGN on BayesElo reading...As goodest can be my memory I though maker read a well-known PGN was a good try to begin. I sometime read English papers on ACADEMIA, on Bayesian Inference to familiarise myself with it and I see the light in a long tunnel.
I give some time to my work, a few times in a month, so I apologize for answer later than the quick response you had the kindness to give to this post...
Having just a BAC D ( science graduate before university in France ), I understand basis, This Winter I worked a lot on Matrix that were not to teach at this level, but understant enough to practice the Massey method, for exemple...But the linearity -or not-of matrix are a bit understanded, I work on this winter.
So a silly thing I wrote in my first post is that a PGN of Fisher Kasparov existed,I just confused the one of the most known game between Fischer and Spassky....
The PGN I had is hiden on a hard drive or my adaptation attempt was unsuccessfull and I threw it.
If I didn't answer as quickly as your response is that I work on my project when I have time ans I wanted to come back with interessant things to expose.
First I downloaded Little Blitzer, Arena and Little though...Chess Motors, after inquiries -and long searchs- to have the possibilities to arrange a round robin tournament -like Ligue2 freanch-, Areana is to long and re-set up the Elos at 2000 although Idis a file with teams names and Elos...Here, my goal was to create a PGN with all teams, with their own Elos...
Before that, I could, with the copy paste fonction on the BayesElo interface , finally read a PGN file and make some litlles command line. Unfortunatly, if I tried the way to put succesfull in the BayesElo all the teams (20 or soon 19 or 18 but it's another story), the superiority matrix ( los) ajust all the team the same Elo and so they have an uniform probability to win or loose...Arena, even if I copy paste names and Elos, set the Elo at 2000. So find equality of win the chamionship.
This summer my not so relentless work was principaly to tried to copy paste teams names and to familiarise myself with all the command lines ( ? helped me ).
But I have another idea : to see the source code. In first I'd like to understand a little bit more the MM ( maximisation minimisation algo ) and I have to confess that if I see a liitle bit more, I'm not very familliar with this is to my point of view my next work...I understand a lot on matrix this winter...next stage will be MM.
I already talk about my little level in programating...My best idea of this summer was to have a look at the source code...ESpecially the CCP file CEloRatingCUI that give the HELP for some of fonctions that seems to set players and elo (I don't want to type on my Dell the almost 380 of a championship, ...SOme of them have already been played it's the 8th journey, the Elo evoluate and it can be intersesting for me to comparate the first days of the championship Elo's and tthe current Elos ).
To resume this summer I pratcice the command line with a PGN I did with the names of the Elos to be more ease ith the command line...But the elos are the same for all, so they have the same probability to win or have any other place in the championship.
To have abetter understanding of what BayesElo and his many CMD can bring to y project ( I don't want to enter manually all the matches on a PGN, too long) the source code open lot of commands that could be interesting...
inparticularity :
elo [p] [elo] ... get[set] Elo of player number p ; if a player has a number ( 1 to 20 in a football championship ) , I understand that his Elo can be set. Is it true ?
"offset [elo [pl]] here,can we offset a particular Elo for a given player (with his number ) ?
p <w> <b> <r> ... probability of result <r> with Elos <w> and <b> here I don't understand what are <w> <b> et <r> wich variable do they belongs?
adavantage of playing first can be an Home Field Advantage in football...I can make a mstake...Feel free to tell me if I'm wrong
drawelo [x] ..... get[set] draw Elo here, is it the points wins with a draw or the standars number of draw expected ? or is it :
prior [x] ....... get[set] prior (= number of virtual draws...
I have , with the bloc note, make a PGN textfile with all the teams and their elos but my first problem is that the Elos are, finally the same foreverybody (I know the Ligue2 is quite balanced);..even in indicate Elos on PGN ( I had read the long PGN Bible, before write here, that you had the kindness to link and it is the real bible for PGN, I wish to thank you for that.
Now I can open a PGN in Bayelo, not have the patience that Arena make a round robin with almost 20 teams, but can make a PGN with teams names or position (p1, P2, pN,..), my question that make me more quik working is that is it possible to attribute an Elo for a given team...At this moment, I would like to compare the beginning of the championship with the current Elos, which have evoluate...
Is there another code source file that ould help me? The EloRating seems to be a good path, finaly my question is can I attribute an Elo to A team ( I can make it on a Bloc Note and paste it on a elorating command) and the others teams of the divisions? Will I have a superiority matrix and and final classement after each team played both advantage home away, with all teams ?
I'm sorry for my english, and above, if my question have no or too complex answer. I'll try to read all the source code, to have more info on the BayesElo helpe list.
Anyway, THank you very much for the time you agreed for me. I just tried to make BayesElo working for football, especially the Ligue2 in France, where Stade Malherbe Caen (were I was born) plays. I bet that in Elite leagues, the process could have much success, but it's not my goal.
-
Rémi Coulom
- Posts: 219
- Joined: Tue Feb 12, 2008 8:31 pm
- Contact:
Re: Bayesian Elo rating
Your problem does not seem very clear to me. If you entered Ligue 2 match results into a PGN file, and bayeselo produced the same Elo for every team, then there must be something wrong. Can you share your PGN? If it is not very big, you can attach it to a message in this forum.
Re: Bayesian Elo rating
EXcuse for my english <Engine-Match jeudi 17 août 2023 20:27:16
Tournament Game in 5 Minutes
Pairing 2/190 : Amiens {Human} - Bordeaux {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 B 2 1,0:0,0 0,0:1,0 20:27:22 0.00 0 0:0 0-1-0
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:27:24
Tournament Game in 5 Minutes
Pairing 3/190 : Angers {Human} - Dunkerque {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 = = 0,5:0,5 0,5:0,5 20:27:29 0.00 0 0:0 0-0-1
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:27:31
Tournament Game in 5 Minutes
Pairing 4/190 : Concarneau {Human} - Saint Etienne {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 B 2 1,0:0,0 0,0:1,0 20:29:12 0.00 0 0:0 0-1-0
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:29:13
Tournament Game in 5 Minutes
Pairing 5/190 : Troyes {Human} - Pau {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 = = 0,5:0,5 0,5:0,5 20:30:11 0.00 0 0:0 0-0-1
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:30:13
Tournament Game in 5 Minutes
Pairing 6/190 : Valenciennes {Human} - Paris {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 = = 0,5:0,5 0,5:0,5 20:30:19 0.00 0 0:0 0-0-1
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:30:21
Tournament Game in 5 Minutes
Pairing 7/190 : Guingamp {Human} - Quevilly {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 * * 0,0:0,0 0,0:0,0 20:30:30 0.00 0 0:0 0-0-0
'm frenhe.
Byreading your post I inderstant litle more.
I fear that I had to write a pgn for full championsip.
THE PGN with follox show that when the championnat begins each team have equal ElO....in another round-robin tournament, it is possiblebut very long aand th stop the match it's seem to choose the winner, wiche is not thegoal.
Engine Score AuBoAnSaTrPaGuBaAjSoCaRoGrQuVaPaCoDuAmLa S-B
01: Auxerre {Human} 0,0/0 · ? 0,00
01: Bordeaux {Human} 0,0/0 · 0,00
01: Angers {Human} 0,0/0 · 0,00
01: Saint Etienne {Human} 0,0/0 · 0,00
01: Troyes {Human} 0,0/0 · 0,00
01: Paris {Human} 0,0/0 · 0,00
01: Guingamp {Human} 0,0/0 · 0,00
01: Bastia {Human} 0,0/0 · 0,00
01: Ajaccio {Human} 0,0/0 · 0,00
01: Sochaux {Human} 0,0/0 · 0,00
01: Caen {Human} 0,0/0 · 0,00
01: Rodez {Human} 0,0/0 · 0,00
01: Grenoble {Human} 0,0/0 · 0,00
01: Quevilly {Human} 0,0/0 · 0,00
01: Valenciennes {Human} 0,0/0 · 0,00
01: Pau {Human} 0,0/0 · 0,00
01: Concarneau {Human} 0,0/0 · 0,00
01: Dunkerque {Human} 0,0/0 · 0,00
01: Amiens {Human} 0,0/0 · 0,00
01: Laval {Human} 0,0/0 ? · 0,00
0 of 380 games played
Name of the tournament: L22324 TournamentELO
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Tournaments\L22324 TournamentELO.pgn
Website:
E-Mail Address:
After an try, whixh I choose toa inner because the tme to matchcan reach several hours ; maybe a parameter that a don't know.
Engine Score AuBoSaAnTrPaVaPaDuSoCaRoGrQuGuBaCoAjAmLa S-B
01: Auxerre {Human} 1,0/1 · 1 0,00
01: Bordeaux {Human} 1,0/1 · 1 0,00
01: Saint Etienne {Human} 1,0/1 · 1 0,00
04: Angers {Human} 0,5/1 · = 0,25
04: Troyes {Human} 0,5/1 · = 0,25
04: Paris {Human} 0,5/1 · = 0,25
04: Valenciennes {Human} 0,5/1 = · 0,25
04: Pau {Human} 0,5/1 = · 0,25
04: Dunkerque {Human} 0,5/1 = · 0,25
10: Sochaux {Human} 0,0/0 · 0,00
10: Caen {Human} 0,0/0 · 0,00
10: Rodez {Human} 0,0/0 · 0,00
10: Grenoble {Human} 0,0/0 · 0,00
10: Quevilly {Human} 0,0/0 · ? 0,00
10: Guingamp {Human} 0,0/0 ? · 0,00
10: Bastia {Human} 0,0/0 · 0,00
10: Concarneau {Human} 0,0/1 0 · 0,00
10: Ajaccio {Human} 0,0/0 · 0,00
10: Amiens {Human} 0,0/1 0 · 0,00
10: Laval {Human} 0,0/1 0 · 0,00
6 of 380 games played
Name of the tournament: L22324 Tournament
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
Website:
E-Mail Address:
enfin mes PGN,basés sur la bible que vous m'avez donné en lien et qu j'avais déjà repéré. J'avoue que le site bayeselo m'a appris , dans un permier temps car plus abordable et accecible..
le PGN de L2 , je pense avoir fait<2 façon de faire, peut être est-cela qui bloque...mais quand je fait une ssimulation à partir d'n pgn. les parties sont longues, voire très longues, alors qu'avec un pgn un bayeslo c'est plus lent, c'est pas soa réputation.
voilà à quoi ressemble une partie sur bayelso
Engine Score AuBoSaAnTrPaVaPaDuSoCaRoGrQuGuBaCoAjAmLa S-B
01: Auxerre {Human} 1,0/1 · 1 0,00
01: Bordeaux {Human} 1,0/1 · 1 0,00
01: Saint Etienne {Human} 1,0/1 · 1 0,00
04: Angers {Human} 0,5/1 · = 0,25
04: Troyes {Human} 0,5/1 · = 0,25
04: Paris {Human} 0,5/1 · = 0,25
04: Valenciennes {Human} 0,5/1 = · 0,25
04: Pau {Human} 0,5/1 = · 0,25
04: Dunkerque {Human} 0,5/1 = · 0,25
10: Sochaux {Human} 0,0/0 · 0,00
10: Caen {Human} 0,0/0 · 0,00
10: Rodez {Human} 0,0/0 · 0,00
10: Grenoble {Human} 0,0/0 · 0,00
10: Quevilly {Human} 0,0/0 · ? 0,00
10: Guingamp {Human} 0,0/0 ? · 0,00
10: Bastia {Human} 0,0/0 · 0,00
10: Concarneau {Human} 0,0/1 0 · 0,00
10: Ajaccio {Human} 0,0/0 · 0,00
10: Amiens {Human} 0,0/1 0 · 0,00
10: Laval {Human} 0,0/1 0 · 0,00
6 of 380 games played
Name of the tournament: L22324 Tournament
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
Website:
E-Mail Address:
Je vais perseverer pour observerr le deébut du championat et la fin du chamionnat...
La Mtéthode massez done peut de réseutat au foot et i faut travailler dur de malrtices de 20 sur 20 et 1 sur 2
Je m'iteresse z lz distribution de poisson , en couplage avec Elo? Skellam sera pour plus tard quand j'aurais compris la formule de <bessel.
je met ici un extrait du tutoriel de <bayeelo Engine Score AuBoSaAnTrPaVaPaDuSoCaRoGrQuGuBaCoAjAmLa S-B
01: Auxerre {Human} 1,0/1 · 1 0,00
01: Bordeaux {Human} 1,0/1 · 1 0,00
01: Saint Etienne {Human} 1,0/1 · 1 0,00
04: Angers {Human} 0,5/1 · = 0,25
04: Troyes {Human} 0,5/1 · = 0,25
04: Paris {Human} 0,5/1 · = 0,25
04: Valenciennes {Human} 0,5/1 = · 0,25
04: Pau {Human} 0,5/1 = · 0,25
04: Dunkerque {Human} 0,5/1 = · 0,25
10: Sochaux {Human} 0,0/0 · 0,00
10: Caen {Human} 0,0/0 · 0,00
10: Rodez {Human} 0,0/0 · 0,00
10: Grenoble {Human} 0,0/0 · 0,00
10: Quevilly {Human} 0,0/0 · ? 0,00
10: Guingamp {Human} 0,0/0 ? · 0,00
10: Bastia {Human} 0,0/0 · 0,00
10: Concarneau {Human} 0,0/1 0 · 0,00
10: Ajaccio {Human} 0,0/0 · 0,00
10: Amiens {Human} 0,0/1 0 · 0,00
10: Laval {Human} 0,0/1 0 · 0,00
6 of 380 games played
Name of the tournament: L22324 Tournament
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
Website:
E-Mail Address:
un tut pour bayeselo tiré du site ResultSet>readpgn wbec.pgn
37 game(s) loaded, 0 game(s) with unknown result ignored.
ResultSet>elo
ResultSet-EloRating>mm
Iteration 100: 1.60455e-005
00:00:00,00
ResultSet-EloRating>exactdist
00:00:00,04
ResultSet-EloRating>ratings
Rank Name Elo + - games score oppo. draws
1 Dragon 4.7.5 120 167 148 8 75% 5 50%
2 Cerebro 2.05a 60 229 211 4 63% 1 25%
3 Movei 0.08.336 9 138 138 12 50% 6 17%
4 Zarkov 4.67 2 146 150 9 44% 18 44%
...
readpgn wbec1to9.pgn
elo
mm
prediction
rounds 2 ;This indicates that each player plays 4 games, 2 with each color.
results
addplayer Esc 1.16
addplayer Pharaon 2.62
addplayer Gandalf 4.32h
addplayer TheCrazyBishop 0045
addplayer Toto
x
players ;Displays the list of players
simulate ;This runs 100000 random simulations of the tournament
x
x
x
ResultSet commands
~~~~~~~~~~~~~~~~~~
players ......... list players
names ........... alphabetical list of player names
results ......... list game results
pgn ............. write results in PGN format
reset ........... reset results and players
addplayer <pn> .. add one player with name <pn>
addresult w b r . add result (w = white, b = black, r = result)
r = 0 (b wins), 1 (draw), or 2 (w wins)
removeresult n .. remove result number
removeplayer n .. remove games of player n
removerare n .... remove games of players with less than n games
pack ............ pack players (remove players with 0 games)
readpgn <file>... read PGN file
connect [p] [fr] remove players not connected to p [fr=forbidden result]
elo ............. open Elo-estimation interface
;EXEMPLE SIMPLE;
[White "p1"][Black "p2"][Result "1-0"] 1. c4 Nf6
[White "p1"][Black "p2"][Result "0-1"] 1. c4 Nf6
;EXEMPLE SCRIPT
[White "p1"][Black "p2"][Result "1-0"] 1. c4 Nf6
[White "p1"][Black "p2"][Result "0-1"] 1. c4 Nf6
addplayer test
addplayer baseline
addwld 0 1 25 10 123
elo
J'ai 2 type de PGN; un par ariana ou littleblitzer(qui organisent de round robinet peuvent attribuer des ELos à des joueurs mais c'est très long).
de mémoire le premier est
[ "Auxerre" ]
[ Auxerre Elo "1506" ]
[ "Bordeaux" ]
[Bordeaux Elo "1494" ]
[ "Saint Etienne" ]
[ Saint Etienne Elo "1462 ]
[ "Troyes" ]
[ Troyes Elo "1457" ]
[ "Paris" ]
[ Paris Elo "1435" ]
[ "Angers" ]
[ Angers Elo "1429" ]
[ "Guingamp" ]
[ Guingamp Elo "1428" ]
[ "Bastia" ]
[ Bastia Elo "1425" ]
[ "Ajaccio" ]
[ Ajaccio Elo "1425" ]
[ "Caen" ]
[ Caen Elo "1425" ]
[ "Sochaux" ]
[ Sohaux Elo "1401" ]
[ "Rodez" ]
[ Rodez Elo "1374" ]
[ "Grenoble" ]
[ Grenoble Elo "1363" ]
[ "Quevilly" ]
[ Quevilly Elo "1359" ]
[ "Valenciennes" ]
[ Valenciennes Elo "1358" ]
[ "Pau" ]
[ Pau Elo "1357" ]
[ "Amiens" ]
[ Amiens Elo "1337" ]
[ "Laval" ]
[ Laval Elo "1332" ]
[ "Concarneau" ]
[ Concarneau Elo "1283" ]
[ "Dunkerque" ]
[ Dunkerque Elo "1283" ]
Le 2ème est
[ "Auxerre" ]
[ Auxerre Elo "1506" ]
[ "Bordeaux" ]
[Bordeaux Elo "1494" ]
[ "Saint Etienne" ]
[ Saint Etienne Elo "1462 ]
[ "Troyes" ]
[ Troyes Elo "1457" ]
[ "Paris" ]
[ Paris Elo "1435" ]
[ "Angers" ]
[ Angers Elo "1429" ]
[ "Guingamp" ]
[ Guingamp Elo "1428" ]
[ "Bastia" ]
[ Bastia Elo "1425" ]
[ "Ajaccio" ]
[ Ajaccio Elo "1425" ]
[ "Caen" ]
[ Caen Elo "1425" ]
[ "Sochaux" ]
[ Sohaux Elo "1401" ]
[ "Rodez" ]
[ Rodez Elo "1374" ]
[ "Grenoble" ]
[ Grenoble Elo "1363" ]
[ "Quevilly" ]
[ Quevilly Elo "1359" ]
[ "Valenciennes" ]
[ Valenciennes Elo "1358" ]
[ "Pau" ]
[ Pau Elo "1357" ]
[ "Amiens" ]
[ Amiens Elo "1337" ]
[ "Laval" ]
[ Laval Elo "1332" ]
[ "Concarneau" ]
[ Concarneau Elo "1283" ]
[ "Dunkerque" ]
[ Dunkerque Elo "1283" ]
Tournament Game in 5 Minutes
Pairing 2/190 : Amiens {Human} - Bordeaux {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 B 2 1,0:0,0 0,0:1,0 20:27:22 0.00 0 0:0 0-1-0
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:27:24
Tournament Game in 5 Minutes
Pairing 3/190 : Angers {Human} - Dunkerque {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 = = 0,5:0,5 0,5:0,5 20:27:29 0.00 0 0:0 0-0-1
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:27:31
Tournament Game in 5 Minutes
Pairing 4/190 : Concarneau {Human} - Saint Etienne {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 B 2 1,0:0,0 0,0:1,0 20:29:12 0.00 0 0:0 0-1-0
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:29:13
Tournament Game in 5 Minutes
Pairing 5/190 : Troyes {Human} - Pau {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 = = 0,5:0,5 0,5:0,5 20:30:11 0.00 0 0:0 0-0-1
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:30:13
Tournament Game in 5 Minutes
Pairing 6/190 : Valenciennes {Human} - Paris {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 = = 0,5:0,5 0,5:0,5 20:30:19 0.00 0 0:0 0-0-1
--------------------------------------------------------------
Engine-Match jeudi 17 août 2023 20:30:21
Tournament Game in 5 Minutes
Pairing 7/190 : Guingamp {Human} - Quevilly {Human}
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
No: B/W 1/2 W:B 1:2 Time Value Moves Dep 1:2 1-2-Draws
1 * * 0,0:0,0 0,0:0,0 20:30:30 0.00 0 0:0 0-0-0
'm frenhe.
Byreading your post I inderstant litle more.
I fear that I had to write a pgn for full championsip.
THE PGN with follox show that when the championnat begins each team have equal ElO....in another round-robin tournament, it is possiblebut very long aand th stop the match it's seem to choose the winner, wiche is not thegoal.
Engine Score AuBoAnSaTrPaGuBaAjSoCaRoGrQuVaPaCoDuAmLa S-B
01: Auxerre {Human} 0,0/0 · ? 0,00
01: Bordeaux {Human} 0,0/0 · 0,00
01: Angers {Human} 0,0/0 · 0,00
01: Saint Etienne {Human} 0,0/0 · 0,00
01: Troyes {Human} 0,0/0 · 0,00
01: Paris {Human} 0,0/0 · 0,00
01: Guingamp {Human} 0,0/0 · 0,00
01: Bastia {Human} 0,0/0 · 0,00
01: Ajaccio {Human} 0,0/0 · 0,00
01: Sochaux {Human} 0,0/0 · 0,00
01: Caen {Human} 0,0/0 · 0,00
01: Rodez {Human} 0,0/0 · 0,00
01: Grenoble {Human} 0,0/0 · 0,00
01: Quevilly {Human} 0,0/0 · 0,00
01: Valenciennes {Human} 0,0/0 · 0,00
01: Pau {Human} 0,0/0 · 0,00
01: Concarneau {Human} 0,0/0 · 0,00
01: Dunkerque {Human} 0,0/0 · 0,00
01: Amiens {Human} 0,0/0 · 0,00
01: Laval {Human} 0,0/0 ? · 0,00
0 of 380 games played
Name of the tournament: L22324 TournamentELO
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Tournaments\L22324 TournamentELO.pgn
Website:
E-Mail Address:
After an try, whixh I choose toa inner because the tme to matchcan reach several hours ; maybe a parameter that a don't know.
Engine Score AuBoSaAnTrPaVaPaDuSoCaRoGrQuGuBaCoAjAmLa S-B
01: Auxerre {Human} 1,0/1 · 1 0,00
01: Bordeaux {Human} 1,0/1 · 1 0,00
01: Saint Etienne {Human} 1,0/1 · 1 0,00
04: Angers {Human} 0,5/1 · = 0,25
04: Troyes {Human} 0,5/1 · = 0,25
04: Paris {Human} 0,5/1 · = 0,25
04: Valenciennes {Human} 0,5/1 = · 0,25
04: Pau {Human} 0,5/1 = · 0,25
04: Dunkerque {Human} 0,5/1 = · 0,25
10: Sochaux {Human} 0,0/0 · 0,00
10: Caen {Human} 0,0/0 · 0,00
10: Rodez {Human} 0,0/0 · 0,00
10: Grenoble {Human} 0,0/0 · 0,00
10: Quevilly {Human} 0,0/0 · ? 0,00
10: Guingamp {Human} 0,0/0 ? · 0,00
10: Bastia {Human} 0,0/0 · 0,00
10: Concarneau {Human} 0,0/1 0 · 0,00
10: Ajaccio {Human} 0,0/0 · 0,00
10: Amiens {Human} 0,0/1 0 · 0,00
10: Laval {Human} 0,0/1 0 · 0,00
6 of 380 games played
Name of the tournament: L22324 Tournament
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
Website:
E-Mail Address:
enfin mes PGN,basés sur la bible que vous m'avez donné en lien et qu j'avais déjà repéré. J'avoue que le site bayeselo m'a appris , dans un permier temps car plus abordable et accecible..
le PGN de L2 , je pense avoir fait<2 façon de faire, peut être est-cela qui bloque...mais quand je fait une ssimulation à partir d'n pgn. les parties sont longues, voire très longues, alors qu'avec un pgn un bayeslo c'est plus lent, c'est pas soa réputation.
voilà à quoi ressemble une partie sur bayelso
Engine Score AuBoSaAnTrPaVaPaDuSoCaRoGrQuGuBaCoAjAmLa S-B
01: Auxerre {Human} 1,0/1 · 1 0,00
01: Bordeaux {Human} 1,0/1 · 1 0,00
01: Saint Etienne {Human} 1,0/1 · 1 0,00
04: Angers {Human} 0,5/1 · = 0,25
04: Troyes {Human} 0,5/1 · = 0,25
04: Paris {Human} 0,5/1 · = 0,25
04: Valenciennes {Human} 0,5/1 = · 0,25
04: Pau {Human} 0,5/1 = · 0,25
04: Dunkerque {Human} 0,5/1 = · 0,25
10: Sochaux {Human} 0,0/0 · 0,00
10: Caen {Human} 0,0/0 · 0,00
10: Rodez {Human} 0,0/0 · 0,00
10: Grenoble {Human} 0,0/0 · 0,00
10: Quevilly {Human} 0,0/0 · ? 0,00
10: Guingamp {Human} 0,0/0 ? · 0,00
10: Bastia {Human} 0,0/0 · 0,00
10: Concarneau {Human} 0,0/1 0 · 0,00
10: Ajaccio {Human} 0,0/0 · 0,00
10: Amiens {Human} 0,0/1 0 · 0,00
10: Laval {Human} 0,0/1 0 · 0,00
6 of 380 games played
Name of the tournament: L22324 Tournament
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
Website:
E-Mail Address:
Je vais perseverer pour observerr le deébut du championat et la fin du chamionnat...
La Mtéthode massez done peut de réseutat au foot et i faut travailler dur de malrtices de 20 sur 20 et 1 sur 2
Je m'iteresse z lz distribution de poisson , en couplage avec Elo? Skellam sera pour plus tard quand j'aurais compris la formule de <bessel.
je met ici un extrait du tutoriel de <bayeelo Engine Score AuBoSaAnTrPaVaPaDuSoCaRoGrQuGuBaCoAjAmLa S-B
01: Auxerre {Human} 1,0/1 · 1 0,00
01: Bordeaux {Human} 1,0/1 · 1 0,00
01: Saint Etienne {Human} 1,0/1 · 1 0,00
04: Angers {Human} 0,5/1 · = 0,25
04: Troyes {Human} 0,5/1 · = 0,25
04: Paris {Human} 0,5/1 · = 0,25
04: Valenciennes {Human} 0,5/1 = · 0,25
04: Pau {Human} 0,5/1 = · 0,25
04: Dunkerque {Human} 0,5/1 = · 0,25
10: Sochaux {Human} 0,0/0 · 0,00
10: Caen {Human} 0,0/0 · 0,00
10: Rodez {Human} 0,0/0 · 0,00
10: Grenoble {Human} 0,0/0 · 0,00
10: Quevilly {Human} 0,0/0 · ? 0,00
10: Guingamp {Human} 0,0/0 ? · 0,00
10: Bastia {Human} 0,0/0 · 0,00
10: Concarneau {Human} 0,0/1 0 · 0,00
10: Ajaccio {Human} 0,0/0 · 0,00
10: Amiens {Human} 0,0/1 0 · 0,00
10: Laval {Human} 0,0/1 0 · 0,00
6 of 380 games played
Name of the tournament: L22324 Tournament
Site/ Country: DESKTOP-FEITOJB, France
Level: Tournament Game in 5 Minutes
Hardware: Intel(R) Core(TM) i5-7300U CPU @ 2.60GHz with 7,9 GB Memory
Operating system: Windows 10 Enterprise Professional (Build 9200) 64 bit
PGN-File: C:\Users\Dell\Documents\LittleBlitzer-2.74\arena_3.5.1\Arena.pgn
Website:
E-Mail Address:
un tut pour bayeselo tiré du site ResultSet>readpgn wbec.pgn
37 game(s) loaded, 0 game(s) with unknown result ignored.
ResultSet>elo
ResultSet-EloRating>mm
Iteration 100: 1.60455e-005
00:00:00,00
ResultSet-EloRating>exactdist
00:00:00,04
ResultSet-EloRating>ratings
Rank Name Elo + - games score oppo. draws
1 Dragon 4.7.5 120 167 148 8 75% 5 50%
2 Cerebro 2.05a 60 229 211 4 63% 1 25%
3 Movei 0.08.336 9 138 138 12 50% 6 17%
4 Zarkov 4.67 2 146 150 9 44% 18 44%
...
readpgn wbec1to9.pgn
elo
mm
prediction
rounds 2 ;This indicates that each player plays 4 games, 2 with each color.
results
addplayer Esc 1.16
addplayer Pharaon 2.62
addplayer Gandalf 4.32h
addplayer TheCrazyBishop 0045
addplayer Toto
x
players ;Displays the list of players
simulate ;This runs 100000 random simulations of the tournament
x
x
x
ResultSet commands
~~~~~~~~~~~~~~~~~~
players ......... list players
names ........... alphabetical list of player names
results ......... list game results
pgn ............. write results in PGN format
reset ........... reset results and players
addplayer <pn> .. add one player with name <pn>
addresult w b r . add result (w = white, b = black, r = result)
r = 0 (b wins), 1 (draw), or 2 (w wins)
removeresult n .. remove result number
removeplayer n .. remove games of player n
removerare n .... remove games of players with less than n games
pack ............ pack players (remove players with 0 games)
readpgn <file>... read PGN file
connect [p] [fr] remove players not connected to p [fr=forbidden result]
elo ............. open Elo-estimation interface
;EXEMPLE SIMPLE;
[White "p1"][Black "p2"][Result "1-0"] 1. c4 Nf6
[White "p1"][Black "p2"][Result "0-1"] 1. c4 Nf6
;EXEMPLE SCRIPT
[White "p1"][Black "p2"][Result "1-0"] 1. c4 Nf6
[White "p1"][Black "p2"][Result "0-1"] 1. c4 Nf6
addplayer test
addplayer baseline
addwld 0 1 25 10 123
elo
J'ai 2 type de PGN; un par ariana ou littleblitzer(qui organisent de round robinet peuvent attribuer des ELos à des joueurs mais c'est très long).
de mémoire le premier est
[ "Auxerre" ]
[ Auxerre Elo "1506" ]
[ "Bordeaux" ]
[Bordeaux Elo "1494" ]
[ "Saint Etienne" ]
[ Saint Etienne Elo "1462 ]
[ "Troyes" ]
[ Troyes Elo "1457" ]
[ "Paris" ]
[ Paris Elo "1435" ]
[ "Angers" ]
[ Angers Elo "1429" ]
[ "Guingamp" ]
[ Guingamp Elo "1428" ]
[ "Bastia" ]
[ Bastia Elo "1425" ]
[ "Ajaccio" ]
[ Ajaccio Elo "1425" ]
[ "Caen" ]
[ Caen Elo "1425" ]
[ "Sochaux" ]
[ Sohaux Elo "1401" ]
[ "Rodez" ]
[ Rodez Elo "1374" ]
[ "Grenoble" ]
[ Grenoble Elo "1363" ]
[ "Quevilly" ]
[ Quevilly Elo "1359" ]
[ "Valenciennes" ]
[ Valenciennes Elo "1358" ]
[ "Pau" ]
[ Pau Elo "1357" ]
[ "Amiens" ]
[ Amiens Elo "1337" ]
[ "Laval" ]
[ Laval Elo "1332" ]
[ "Concarneau" ]
[ Concarneau Elo "1283" ]
[ "Dunkerque" ]
[ Dunkerque Elo "1283" ]
Le 2ème est
[ "Auxerre" ]
[ Auxerre Elo "1506" ]
[ "Bordeaux" ]
[Bordeaux Elo "1494" ]
[ "Saint Etienne" ]
[ Saint Etienne Elo "1462 ]
[ "Troyes" ]
[ Troyes Elo "1457" ]
[ "Paris" ]
[ Paris Elo "1435" ]
[ "Angers" ]
[ Angers Elo "1429" ]
[ "Guingamp" ]
[ Guingamp Elo "1428" ]
[ "Bastia" ]
[ Bastia Elo "1425" ]
[ "Ajaccio" ]
[ Ajaccio Elo "1425" ]
[ "Caen" ]
[ Caen Elo "1425" ]
[ "Sochaux" ]
[ Sohaux Elo "1401" ]
[ "Rodez" ]
[ Rodez Elo "1374" ]
[ "Grenoble" ]
[ Grenoble Elo "1363" ]
[ "Quevilly" ]
[ Quevilly Elo "1359" ]
[ "Valenciennes" ]
[ Valenciennes Elo "1358" ]
[ "Pau" ]
[ Pau Elo "1357" ]
[ "Amiens" ]
[ Amiens Elo "1337" ]
[ "Laval" ]
[ Laval Elo "1332" ]
[ "Concarneau" ]
[ Concarneau Elo "1283" ]
[ "Dunkerque" ]
[ Dunkerque Elo "1283" ]
Re: Bayesian Elo rating
Avecc ebout de source trouvé, où j'ai à mon niveau compris pas mal de chose dans ce quz je souhaitait, le elo rating et donc peut-être des confrontations multiplies.
le bout de la source code estle fichier CCCP -30Ko u des plus grand sinon grand de tous les fichiers ode source...Le nom m'a intringué aussi par pas apect alléchant : CEloRatingCUI.
il est assz long et e maitrisant que le basic et le microsoft 2, un PC bien sur et ma fidèle calucatrice Casio Graph75 (grace au basic et aux tableurs j'étais aussi doué et réputé pour dessiner sur une graphique....Bref passons le copié collé que certains ont peut-être eu la uriosité d'éplucher ( comme moi, j'ai retenu le plus interessant pour moi, mon but étant de jouer avec les buts...)...
on ouvre ici la console elorating command.Si j'ai bien compris, l'avantage de joueur en premier peut etre assimilé au Home Field Advantage ( en ELO c'est l'avantage du terrain, il est en moyenne de 50)...Si je ne me trompe on peut attribuer 1 point en cas denul.ON peut ausi, arretez moi si je dis des betisses setting le % de nuls ( au foot c'est 30% environ, environ plus qu'aux échecs)
elo of player number me laisse comprendre que l'onn peut, si une equipe a un numero, on peut lui attribuer un Elo..Si c'est le cas, c'est le graal que je cherche. Je me renseignerait sut le differents types de calculs, mais je suis loin de comprendre. j'ai épluché quelques codes sources que je comprend un peu mais celui là se rapproche de mon but ;se servir de BayesElo car si je maitrise la méthode Massey( à base de matrice ma calculatrice n'en a que trop peu pour 2à équipes il faut 20x20 et 1x20, La distribution de Poisson ( avec mon astuce regresion linéaire°), BayesElo m'a demandé u couple d'années pour comprendre.J'aurais été plus vite enlisant les fichier codes source car je maitrise non pas l'asembleur maisles langges logiques -ne dit-on pas prolog-ic ?)...Je travaille près de 20 ans après avoir passé mon bac C et les maths me passionnent et quand que joue, je perd pas
.J'spère ne pas avoir été trop lourd mais si je peux arriver à faire fonctionner pour un championnat de l2 (je suis né à Caen et il y a moins d'équipe qu'en ligu21) la machine qui, après tâtonement semble marcher pour des joueurs multiples,je serai , même si j'en retirerais un peu de fierté, éternellement reconnaissant pour son créateur. Car il m'aura aidé sur un point sur lequel je bloque, mais avec quelques éclaircissement sur en programmation ( plus je lirai le code source plus je comprandrait vite) mais n'ayant pas ses aptitudes et compétence il se pourrait que je pose une question parfois.J'avance déjç plus en 1 semaine 10 jours en lisant le code source que la page de <bayeslo>.manque de confiance en moi...Je vais épelucher lorsque j'a'urais le temps les fonctions help qui m'ont ait avancer...pour le langage machine ou C je pense pas avoir besoin pour faire tourner la machine.Je pense que BaYesElo pourrait intéresser des amateur de sport qui s'interesent aux maths et aux sciences. je vais à mon rythme mais j y arriverai.Merci 
out << "EloRating commands\n";
out << "~~~~~~~~~~~~~~~~~~\n";
out << "advantage [x] ... get[set] advantage of playing first\n";
out << "drawelo [x] ..... get[set] draw Elo\n";
out << "prior [x] ....... get[set] prior (= number of virtual draws)\n";
out << "elo [p] [elo] ... get[set] Elo of player number p\n";
out << "mm [a] [d] ...... compute maximum-likelihood Elos:\n";
out << " a: flag to compute advantage (default = 0)\n";
out << " d: flag to compute elodraw (default = 0)\n";
out << "elostat ......... compute ratings with ELOstat algorithm\n";
out << '\n';
out << "ratings [min [f [F]]] list players and their ratings:\n";
out << " min: minimum number of games\n";
out << " f: file name that contains the list of names\n";
out << " F: Full-rank flag (0 or 1, default 0)\n";
out << "details ......... detailed list\n";
out << "offset [elo [pl]] get[set] Elo offset, or player (pl) elo\n";
out << "scale ........... get[set] Elo scale\n";
out << '\n';
out << "p <w> <b> <r> ... probability of result <r> with Elos <w> and <b>\n";
out << "likelihood ...... log-likelihood of the current Elo list\n";
out << "prediction ...... open prediction interface\n";
out << '\n';
out << "plotres [M] . plot result likelihoods as a function of rating diff:\n";
out << " i: number of intervals (default = 21)\n";
out << " M: maximum rating difference (default = 500)\n";
out << "plotdraw [n] [d] plot draw frequency as a function of average rating\n";
out << " n: number of games per group (default = 100)\n";
out << " d: maximum rating difference (default = 100)\n";
out << '\n';
out << "confidence ...... get[set] level of confidence intervals\n";
out << "variance ........ compute intervals with the diagonal of the Hessian\n";
out << "covariance ...... compute intervals with the full Hessian\n";
out << "los [f] [p] [w] . likelihood of superiority (f=first,p=players,w=width)\n";
out << '\n';
out << "minelo [x] ...... get[set] minimum Elo\n";
out << "maxelo [x] ...... get[set] maximum Elo\n";
out << "resolution [n] .. get[set] resolution\n";
out << "jointdist [p] ... compute intervals from joint distribution\n";
out << "exactdist [p] ... compute intervals assuming exact opponent Elos\n";
out << "advdist ......... likelihood distribution of advantage\n";
out << "drawdist ........ likelihood distribution of drawelo\n";
out << '\n';
out << "pairstats i j ... get stats between players i and j\n";
out << '\n';
le bout de la source code estle fichier CCCP -30Ko u des plus grand sinon grand de tous les fichiers ode source...Le nom m'a intringué aussi par pas apect alléchant : CEloRatingCUI.
il est assz long et e maitrisant que le basic et le microsoft 2, un PC bien sur et ma fidèle calucatrice Casio Graph75 (grace au basic et aux tableurs j'étais aussi doué et réputé pour dessiner sur une graphique....Bref passons le copié collé que certains ont peut-être eu la uriosité d'éplucher ( comme moi, j'ai retenu le plus interessant pour moi, mon but étant de jouer avec les buts...)...
on ouvre ici la console elorating command.Si j'ai bien compris, l'avantage de joueur en premier peut etre assimilé au Home Field Advantage ( en ELO c'est l'avantage du terrain, il est en moyenne de 50)...Si je ne me trompe on peut attribuer 1 point en cas denul.ON peut ausi, arretez moi si je dis des betisses setting le % de nuls ( au foot c'est 30% environ, environ plus qu'aux échecs)
elo of player number me laisse comprendre que l'onn peut, si une equipe a un numero, on peut lui attribuer un Elo..Si c'est le cas, c'est le graal que je cherche. Je me renseignerait sut le differents types de calculs, mais je suis loin de comprendre. j'ai épluché quelques codes sources que je comprend un peu mais celui là se rapproche de mon but ;se servir de BayesElo car si je maitrise la méthode Massey( à base de matrice ma calculatrice n'en a que trop peu pour 2à équipes il faut 20x20 et 1x20, La distribution de Poisson ( avec mon astuce regresion linéaire°), BayesElo m'a demandé u couple d'années pour comprendre.J'aurais été plus vite enlisant les fichier codes source car je maitrise non pas l'asembleur maisles langges logiques -ne dit-on pas prolog-ic ?)...Je travaille près de 20 ans après avoir passé mon bac C et les maths me passionnent et quand que joue, je perd pas
.J'spère ne pas avoir été trop lourd mais si je peux arriver à faire fonctionner pour un championnat de l2 (je suis né à Caen et il y a moins d'équipe qu'en ligu21) la machine qui, après tâtonement semble marcher pour des joueurs multiples,je serai , même si j'en retirerais un peu de fierté, éternellement reconnaissant pour son créateur. Car il m'aura aidé sur un point sur lequel je bloque, mais avec quelques éclaircissement sur en programmation ( plus je lirai le code source plus je comprandrait vite) mais n'ayant pas ses aptitudes et compétence il se pourrait que je pose une question parfois.J'avance déjç plus en 1 semaine 10 jours en lisant le code source que la page de <bayeslo>.manque de confiance en moi...Je vais épelucher lorsque j'a'urais le temps les fonctions help qui m'ont ait avancer...pour le langage machine ou C je pense pas avoir besoin pour faire tourner la machine.Je pense que BaYesElo pourrait intéresser des amateur de sport qui s'interesent aux maths et aux sciences. je vais à mon rythme mais j y arriverai.Merci out << "EloRating commands\n";
out << "~~~~~~~~~~~~~~~~~~\n";
out << "advantage [x] ... get[set] advantage of playing first\n";
out << "drawelo [x] ..... get[set] draw Elo\n";
out << "prior [x] ....... get[set] prior (= number of virtual draws)\n";
out << "elo [p] [elo] ... get[set] Elo of player number p\n";
out << "mm [a] [d] ...... compute maximum-likelihood Elos:\n";
out << " a: flag to compute advantage (default = 0)\n";
out << " d: flag to compute elodraw (default = 0)\n";
out << "elostat ......... compute ratings with ELOstat algorithm\n";
out << '\n';
out << "ratings [min [f [F]]] list players and their ratings:\n";
out << " min: minimum number of games\n";
out << " f: file name that contains the list of names\n";
out << " F: Full-rank flag (0 or 1, default 0)\n";
out << "details ......... detailed list\n";
out << "offset [elo [pl]] get[set] Elo offset, or player (pl) elo\n";
out << "scale ........... get[set] Elo scale\n";
out << '\n';
out << "p <w> <b> <r> ... probability of result <r> with Elos <w> and <b>\n";
out << "likelihood ...... log-likelihood of the current Elo list\n";
out << "prediction ...... open prediction interface\n";
out << '\n';
out << "plotres [M] . plot result likelihoods as a function of rating diff:\n";
out << " i: number of intervals (default = 21)\n";
out << " M: maximum rating difference (default = 500)\n";
out << "plotdraw [n] [d] plot draw frequency as a function of average rating\n";
out << " n: number of games per group (default = 100)\n";
out << " d: maximum rating difference (default = 100)\n";
out << '\n';
out << "confidence ...... get[set] level of confidence intervals\n";
out << "variance ........ compute intervals with the diagonal of the Hessian\n";
out << "covariance ...... compute intervals with the full Hessian\n";
out << "los [f] [p] [w] . likelihood of superiority (f=first,p=players,w=width)\n";
out << '\n';
out << "minelo [x] ...... get[set] minimum Elo\n";
out << "maxelo [x] ...... get[set] maximum Elo\n";
out << "resolution [n] .. get[set] resolution\n";
out << "jointdist [p] ... compute intervals from joint distribution\n";
out << "exactdist [p] ... compute intervals assuming exact opponent Elos\n";
out << "advdist ......... likelihood distribution of advantage\n";
out << "drawdist ........ likelihood distribution of drawelo\n";
out << '\n';
out << "pairstats i j ... get stats between players i and j\n";
out << '\n';
-
Rémi Coulom
- Posts: 219
- Joined: Tue Feb 12, 2008 8:31 pm
- Contact:
Re: Bayesian Elo rating
Merci pour les informations supplémentaires, mais je ne comprends toujours pas bien quel est le PGN utilisé. Pour que bayeselo fonctionne, il faut que chaque partie ait un "Result". Les champs "Elo" du PGN ne sont pas utilisés par bayeselo.
Si bayeselo produit des résultats étonnants, alors il me faut les données et les commandes exactes utilisées par bayeselo, de manière à ce que je puisse reproduire exactement les choses sur ma machine. Sinon, c'est difficile de donner des explications.
Si je comprends bien, ton objectif est de faire des prédictions pour le foot. Cela peut se faire en 2 temps avec bayeselo:
Rémi
Si bayeselo produit des résultats étonnants, alors il me faut les données et les commandes exactes utilisées par bayeselo, de manière à ce que je puisse reproduire exactement les choses sur ma machine. Sinon, c'est difficile de donner des explications.
Si je comprends bien, ton objectif est de faire des prédictions pour le foot. Cela peut se faire en 2 temps avec bayeselo:
- Calculer le Elo des équipes avec les résultats passés (au format PGN)
- Faire des pronostics avec ces Elos pour les résultats futurs
Rémi
Re: Bayesian Elo rating
LittleBlitzer is a Windows tournament manager program for UCI based chess engines. It was created to enable
playing thousands of very fast games using multiple threads for chess engine testing purposes. It creates a
PGN file designed to be used by Bayeselo or Elostat to give a final rating for each engine.
Download LittleBlitzer from http://www.kimiensoftware.com
Contact me at nathanthom@gmail.com
Instructions
------------
1. Create a Engines.lbe file with a list of UCI engines to use. Do not put spaces either side of the =.
Blank lines are fine, UCI parameters should appear after the relevant Engine line, so in the following
example "Time Buffer" is only set for t20090922.exe and "Hash=128" is only set for Hamsters.exe. You can
use the special variable LB_Name to override the default engine name used in the results file.
Engine=C:\Projects\LittleThought\Release\LittleThought 1.06.54.exe
NumThreads=2
TraceLevel=1
LB_Name=LT 1.06.54
Engine=C:\Downloads\Engines\Twisted\twisted20090922\t20090922.exe
Time Buffer=100
Engine=C:\Downloads\Engines\AnMon\AnMon_5.75.exe
Engine=C:\Downloads\Engines\Hamsters_0.0.6_Win32\Hamsters.exe
Hash=128
2. Clck the Load Engines button
3. Click the Load Tournament Settings button. This will load settings from the specified Tournament.lbt
file if it exists, otherwise uses defaults. Make whatever changes are required and OK to save.
Num Total Rounds: Number of games to play (split across Num Parallel Tournaments). Note that because of the
multithreaded nature of LittleBlitzer, the actual number may be slightly higher.
Num Parallel Tournaments: Number of tournaments to run in parallel. E.g. if using 4cores, use no more than
4 threads for optimal performance.
Hash: Size of Hash tables for all engines (MB) (UCI parameter = Hash)
Ponder: TBD
Own Book: Tell each engine to use their own book if they have one (UCI parameter = OwnBook).
Type:
Gauntlet: 1st engine vs all others repeatedly. 1st engine plays many more games than other engines.
Round-Robin: 1vs2, 1vs3, 2vs1, 2vs3, 3vs1, 3vs2 etc. All engines play the same number of games.
Variant:
Standard
Chess960 (FRC)
Adjudications:
Mate Score/Mate Moves: If one engine thinks it is winning by at least this score (+ve, centi-pawns) and
the other engine thinks its losing by at least this score, and this situation has
held for the specified number of moves, LittleBlitzer will terminate the game as
a win/loss.
Draw Moves: If a game goes for this many moves it is declared a draw, regardless of score/time.
Time Control:
Fixed Time/Move: This time (in ms) will be used for every move - useful if the engines get lots of time
losses on other time controls.
Blitz: Uses a total base time and an increment which is added to the base time after making each move.
Tournament: Specify a time in which the specified number of moves must be made. When the moves have been
made, the clock/moves left are reset to the starting values, repeatedly.
Starting Positions:
Opening: Always use a standard opening (or randomised opening for Chess960 variant).
FEN: Specify a single position via a FEN string.
EPD: Load starting positions from an EPD file. Positions are used sequentially and fairly - i.e. each
engine plays both white and black against each opponent.
PGN: Load starting positions from a PGN file. If it sees [FEN] tags it will load that position. If not,
it will play through all moves shown in the game and the resulting position will form the starting
position.
Randomize: Randomly selects a game from the loaded opening positions. Note that this means engines A and B
will not necessarily play the same opening against each other as different colours. However, over
enough games this potential advantage should not cause a problem.
4. Click the Start New button. If the results.pgn file specified already exists, you will be asked to append
or overwrite.
5. Use the Pause/Resume/+/- buttons to control the number of threads while its running. Try to minimise how
often you do this, as it can affect the total number of games played. Specifically, the first engine
will tend to get more games each time additional threads are added.
6. When done, set the threads to zero and wait for it to finish. Then run the Results.pgn file
through Bayeselo, using commands something like:
readpgn results1.pgn
readpgn results2.pgn
readpgn results3.pgn
elo
mm
covariance
ratings
x
Notes
-----
- Enabling FRC via Variant=1 will automatically set each engine's UCI_Chess960 parameter to true.
- Parameters specified within the engines file will override any tournament settings, e.g. Hash size.
- For EPD opening suites, each position will get played 2*(numengines-1) times, so best to have number
of Rounds a multiple of this but also evenly divisible by the number of parallel tournaments.
- For mate adjudication, both engines must agree on the score e.g. engine1 > 900 and engine2 < -900
Result Columns
--------------
After the engine names, there will be 5 groups of columns with data.
For example:
1. LittleThought v1.06.72 11989.0/23644 11023-10689-1932 (L: m=5083 t=198 i=0 a=5408) (D: r=1018 i=511 f=225 s=87 a=91) (tpm=112.4 d=10.2 nps=827951)
1. Score (points/games played)
e.g. score of 11989 over 23644 games
2. Score Breakdown (wins-losses-draws)
e.g. 11023 wins, 10689 losses, 1932 draws
3. Loss reasons (m = mated, t = timed out, i = illegal moves, a = adjudication)
e.g. 5083 mated, 198 timeouts, 0 illegal moves, 5408 adjudicated losses
4. Draw reasons (r = repetition, i = insufficient material, f = 50 moves, s = stalemate, a = adjudication)
e.g. 1018 repetitions, 511 insufficient material, 225 50-moves, 87 stalemates, 91 adjudicated draws
5. Engine stats (tpm = avg time per move, d = avg depth, nps = avg nps)
e.g. 112.4 ms/move, 10.2 ply, 827951 nps
Version History
---------------
v2.74 06/06/2012
- Added extra digit of precision to the avg depth per engine output.
v2.73 21/05/2012
- Forced all moves output to lower case to handle engines that output upper case moves and engines that
cannot handle them as input.
v2.72 20/04/2011
- Fixed output of some empty PGN games.
v2.71 10/04/2011
- Fixed calculation of adjudication by score loss.
- PGN and EPD files used for opening positions may now have spaces in their paths.
- Fixed ambiguity in algebraic notation used (e.g. Rdd4 instead of R3d4).
- Added number of opening positions to results display.
- Fixed output of first move as white.
v2.7 25/03/2011
- Added option to output full PGN of games. Testing shows no measurable slowdown.
- Can read starting positions from PGN files.
v2.6 22/02/2011
- Fixed parsing of FEN starting postitions
- Fixed bug with identifying mate and stalemate positions
v2.5 10/11/2010
- Fixed number of games played in total and per engine. Will now always play the exact number asked.
- Fixed reporting of illegal move when it should have been a repetition draw.
- Each illegal move now generates a new file with a random suffix.
- Some GUI tweaking.
- Improved logging output.
- Reduced LB overheads and more accurate measurement of real time taken by engines.
v2.4 23/08/2010
- Added Adjudication parameters to help speed up games
- Added additional time controls (fixed time per move + tournament x moves in y secs)
- Added dialog to edit tournament settings
- Some general polishing to make it more user friendly
v2.3 08/08/2010
- Can now specify starting positions via the Position parameter (fixed or EPD of positions)
- Fixed some illegal move reporting problems
- Modified engines file format, now accepts UCI parameters
v2.2 04/08/2010
- Added Chess960 support (Variant = 1)
- Added some error checking when loading engines and support for path names with spaces
- Enhanced illegal move dump file
- Can now log the first engine's communications (best used with single tournaments)
- Now works on old XP installations without the VC redist
- Increased max threads to 16
v2.1 15/03/2010 (First public release)
- Supports Gauntlet or Round Robin tournaments
- Supports up to 4 parallel tournaments
- Designed to play extremely fast, e.g. all moves in 1 sec
- Pondering is not enabled yet
playing thousands of very fast games using multiple threads for chess engine testing purposes. It creates a
PGN file designed to be used by Bayeselo or Elostat to give a final rating for each engine.
Download LittleBlitzer from http://www.kimiensoftware.com
Contact me at nathanthom@gmail.com
Instructions
------------
1. Create a Engines.lbe file with a list of UCI engines to use. Do not put spaces either side of the =.
Blank lines are fine, UCI parameters should appear after the relevant Engine line, so in the following
example "Time Buffer" is only set for t20090922.exe and "Hash=128" is only set for Hamsters.exe. You can
use the special variable LB_Name to override the default engine name used in the results file.
Engine=C:\Projects\LittleThought\Release\LittleThought 1.06.54.exe
NumThreads=2
TraceLevel=1
LB_Name=LT 1.06.54
Engine=C:\Downloads\Engines\Twisted\twisted20090922\t20090922.exe
Time Buffer=100
Engine=C:\Downloads\Engines\AnMon\AnMon_5.75.exe
Engine=C:\Downloads\Engines\Hamsters_0.0.6_Win32\Hamsters.exe
Hash=128
2. Clck the Load Engines button
3. Click the Load Tournament Settings button. This will load settings from the specified Tournament.lbt
file if it exists, otherwise uses defaults. Make whatever changes are required and OK to save.
Num Total Rounds: Number of games to play (split across Num Parallel Tournaments). Note that because of the
multithreaded nature of LittleBlitzer, the actual number may be slightly higher.
Num Parallel Tournaments: Number of tournaments to run in parallel. E.g. if using 4cores, use no more than
4 threads for optimal performance.
Hash: Size of Hash tables for all engines (MB) (UCI parameter = Hash)
Ponder: TBD
Own Book: Tell each engine to use their own book if they have one (UCI parameter = OwnBook).
Type:
Gauntlet: 1st engine vs all others repeatedly. 1st engine plays many more games than other engines.
Round-Robin: 1vs2, 1vs3, 2vs1, 2vs3, 3vs1, 3vs2 etc. All engines play the same number of games.
Variant:
Standard
Chess960 (FRC)
Adjudications:
Mate Score/Mate Moves: If one engine thinks it is winning by at least this score (+ve, centi-pawns) and
the other engine thinks its losing by at least this score, and this situation has
held for the specified number of moves, LittleBlitzer will terminate the game as
a win/loss.
Draw Moves: If a game goes for this many moves it is declared a draw, regardless of score/time.
Time Control:
Fixed Time/Move: This time (in ms) will be used for every move - useful if the engines get lots of time
losses on other time controls.
Blitz: Uses a total base time and an increment which is added to the base time after making each move.
Tournament: Specify a time in which the specified number of moves must be made. When the moves have been
made, the clock/moves left are reset to the starting values, repeatedly.
Starting Positions:
Opening: Always use a standard opening (or randomised opening for Chess960 variant).
FEN: Specify a single position via a FEN string.
EPD: Load starting positions from an EPD file. Positions are used sequentially and fairly - i.e. each
engine plays both white and black against each opponent.
PGN: Load starting positions from a PGN file. If it sees [FEN] tags it will load that position. If not,
it will play through all moves shown in the game and the resulting position will form the starting
position.
Randomize: Randomly selects a game from the loaded opening positions. Note that this means engines A and B
will not necessarily play the same opening against each other as different colours. However, over
enough games this potential advantage should not cause a problem.
4. Click the Start New button. If the results.pgn file specified already exists, you will be asked to append
or overwrite.
5. Use the Pause/Resume/+/- buttons to control the number of threads while its running. Try to minimise how
often you do this, as it can affect the total number of games played. Specifically, the first engine
will tend to get more games each time additional threads are added.
6. When done, set the threads to zero and wait for it to finish. Then run the Results.pgn file
through Bayeselo, using commands something like:
readpgn results1.pgn
readpgn results2.pgn
readpgn results3.pgn
elo
mm
covariance
ratings
x
Notes
-----
- Enabling FRC via Variant=1 will automatically set each engine's UCI_Chess960 parameter to true.
- Parameters specified within the engines file will override any tournament settings, e.g. Hash size.
- For EPD opening suites, each position will get played 2*(numengines-1) times, so best to have number
of Rounds a multiple of this but also evenly divisible by the number of parallel tournaments.
- For mate adjudication, both engines must agree on the score e.g. engine1 > 900 and engine2 < -900
Result Columns
--------------
After the engine names, there will be 5 groups of columns with data.
For example:
1. LittleThought v1.06.72 11989.0/23644 11023-10689-1932 (L: m=5083 t=198 i=0 a=5408) (D: r=1018 i=511 f=225 s=87 a=91) (tpm=112.4 d=10.2 nps=827951)
1. Score (points/games played)
e.g. score of 11989 over 23644 games
2. Score Breakdown (wins-losses-draws)
e.g. 11023 wins, 10689 losses, 1932 draws
3. Loss reasons (m = mated, t = timed out, i = illegal moves, a = adjudication)
e.g. 5083 mated, 198 timeouts, 0 illegal moves, 5408 adjudicated losses
4. Draw reasons (r = repetition, i = insufficient material, f = 50 moves, s = stalemate, a = adjudication)
e.g. 1018 repetitions, 511 insufficient material, 225 50-moves, 87 stalemates, 91 adjudicated draws
5. Engine stats (tpm = avg time per move, d = avg depth, nps = avg nps)
e.g. 112.4 ms/move, 10.2 ply, 827951 nps
Version History
---------------
v2.74 06/06/2012
- Added extra digit of precision to the avg depth per engine output.
v2.73 21/05/2012
- Forced all moves output to lower case to handle engines that output upper case moves and engines that
cannot handle them as input.
v2.72 20/04/2011
- Fixed output of some empty PGN games.
v2.71 10/04/2011
- Fixed calculation of adjudication by score loss.
- PGN and EPD files used for opening positions may now have spaces in their paths.
- Fixed ambiguity in algebraic notation used (e.g. Rdd4 instead of R3d4).
- Added number of opening positions to results display.
- Fixed output of first move as white.
v2.7 25/03/2011
- Added option to output full PGN of games. Testing shows no measurable slowdown.
- Can read starting positions from PGN files.
v2.6 22/02/2011
- Fixed parsing of FEN starting postitions
- Fixed bug with identifying mate and stalemate positions
v2.5 10/11/2010
- Fixed number of games played in total and per engine. Will now always play the exact number asked.
- Fixed reporting of illegal move when it should have been a repetition draw.
- Each illegal move now generates a new file with a random suffix.
- Some GUI tweaking.
- Improved logging output.
- Reduced LB overheads and more accurate measurement of real time taken by engines.
v2.4 23/08/2010
- Added Adjudication parameters to help speed up games
- Added additional time controls (fixed time per move + tournament x moves in y secs)
- Added dialog to edit tournament settings
- Some general polishing to make it more user friendly
v2.3 08/08/2010
- Can now specify starting positions via the Position parameter (fixed or EPD of positions)
- Fixed some illegal move reporting problems
- Modified engines file format, now accepts UCI parameters
v2.2 04/08/2010
- Added Chess960 support (Variant = 1)
- Added some error checking when loading engines and support for path names with spaces
- Enhanced illegal move dump file
- Can now log the first engine's communications (best used with single tournaments)
- Now works on old XP installations without the VC redist
- Increased max threads to 16
v2.1 15/03/2010 (First public release)
- Supports Gauntlet or Round Robin tournaments
- Supports up to 4 parallel tournaments
- Designed to play extremely fast, e.g. all moves in 1 sec
- Pondering is not enabled yet
Re: Bayesian Elo rating
C'est parceque j'ai appris tout seul avec ma calculatrce (beaucoup) et un peu mon ordinateur à calculer avec diveres méthodes,dont le Elo auquel j'ai adjoint un coefficient de surface. J'ai sur ma calcultrice scientifique (même si les smartphones ont désormais leur propre tableur...ce qui est plus simple)
SI la loi du BayesElo est de ne pas prendere le champs Elo du PGN c'est en fait compliqué. J'ai testé arena il me semble, onpeut généerer avec plusieurs moteurs d'échecs des saisonsq de championat. Le problème est que le match est interminable et on doit choisir qui gagne.Pir, on peut attribuer le Elo de l'équipe mais il à 2000 ès qu'elle joue, un truc doit m'échaper...L'idée que j'avais était de créer un tournoi de hcampionat genre avec arena et l'importer, en PNG dans bayesELO
Le Elo ne donne pas la forme actuelle du joueur (qui ne peut compter que sur lui ) ou de l'équipe (un malade et hop, un remplaçant) ,mais elle est capable, avec l'argent qui est relatif à un niveau année par année, bien qu'il y ait des surprises, garder u niveau chaque année.
En effet jm'ocupe à trouver le gagnant...
Pour le football par exemple je vais surClubElo, qui classe depuis l'historique des matchs des années 1950 ou 1960..le classement Elo est bien affiné et c'est bien dommage car sur son ancien site il expliquait bien et avait quelques idées comme le calcul du Home Field Avantage (l'avantage des blancs un peu..il est en moyenne autours de 50 Pts Elo) Et un formule pour trouver un nombre de buts estimés dans le but d'une distribution de Poisson..l'effort était louable et marche pour les équipes au niveau très similaires...je préfère me tenir aux buts pour but ontre de chaque équpe avec ma sauce pour la distribution..j'ai mis la formule sur un tableur de calculatrice et ajouté , avec le coef de ma trouvaille, lea formule pour trouer la probabilité d'un match nul d'ABROBecker.je fais une régression linéaire comme le Elo, la forme d'une équipe appartient au passé-ou au présent-,et à vrai dire même si j'admire ce Elo, une formeou méforme subite est lente à déceler, mais j'ai une parade qui n'a rien à voir avec Elo donc pas sujet ici, puis j'effectue une distribution de poisson. ClubElo est mis à jour régulièrement, très fiables pour les clubs de nations bien représentées en UEFA...Pour certains pays qui n'envoie qu'un club , certes champion dans son pays, se faire exploser le Elo du club est compté mais certains petits pays européens (exclusivement européen) le Elo des clubs peut ne pas exister...qui pour miser sur Tiraspol ? Il y a unebonne vingtaine de ligues où c'est bien tenu...A peu de chose près celles qui sont autorisées à la mise en France.
Pour l'international EloRatings s'occupe plus des sélections nationales (d'ailleurs le système du foot féminin est très fortement basé sur le Elo avec perte de points si par exemple une équipe devrait sur la différence Elo gagne 1-0 au lieu de 3-0)...EloRatings fait figure sérieuse, c'est ce qu'est aux selections le Club Elo des clubs. On en trouve d'autres 538 et son mélange à base d'ESPN, mais secret...Le Elo club ELO
dévoilait ses formules et techniques.On pouvait trouver le lien caché dans un ici en fait j'éditeil existe là http://clubelo.com/System ..IL me reste des notes prises si quelq'un est interessé : calcul de la moyenne de ligue de l'avantage à domicile ( plus les pays sont petit, plus ce coef est faible)...calcul du nombre de but je m'en sert, par exemple pour un distribution de Poisson car l'écart de la formule d'Abrobrecker,qui donne une proba de nul adapté aux échecs, mais avec avec mes coefs ( adapté au foot le nombre de draw est pas le même, a tendance à soit eê^tre serrée, soit exagérée...la moyenne des deux me donne environ les mêmes probas que les bookmakers, parfois -;) .
je dois encore bosser les maximisation minimisation pour tenterd'ameliorer le Elo, mais je doute que ma casio gragph75 y arrive...Par contre Bayeselo a , je pense, du potentiel et je me ferai d'abord la main pour ensuite aller plis loin 'j'espère.Car BayesElo a un tas de fonctionalités, mais si il faut 6 mois un an de cacalcul d elo - et je compte pas les autres équipes (je fait des regressions linéaire puis distribution de Poisson et ça prend du temps pour les équipes.Mais terriblement plus efficace pour trouver les outsiders le pire c'est que je joue pas ou plus au tabac je m'amuse avec des mises de 0,1 Euros qui se transforment en euros ou en dizaines d'euros...c'est que je veux pas me ruiner,c'est l'exercice intellectuel qui me motive).
J'espère revenir avec des bonnes nouvelle, pour ça il faudra travailler dur ce que j'ignore ! et j'espère en apprendre un peu plus en passant lire , même si mon objectif ici est de simuler un championnat de ligue ( foot ou autre ).ET beaucoup de temps, de patience pour taper au moins un an ou deux de 380 matchs, et calculer les Elos et points gagnés sur ma calculatrice...J'essaierai avec BayesElo en commençant la saison dernière avec 1500 en L1 et 1400 en L2. puis rebelote la semaine suivante- non je suis pas sérieux dès que possible..C'est peut-etre la solution^..Quand on veut queqlquechose faut savoir se retrousser les manches...Si BayesElo ne peut pas ajuster à chaque équipe dans un championnat sur l'année, que Arena met un temps fou, permet de changer manuellement les ELos mais au début du tournoi remet à 2000, et pour finir pour 380 matchs au bout de 5 minutes on choisit le gagnant, ce qui n'et pas mon but.
Voilà la solution : rejoueur (j'ai gardé les elos mais seulement celui de mon club), par exemple cette saison en notant les Elos de chaque équipe...Calculer les points à échanger ou non et calcul du nouvel Elo avec tableur...puis trouver l'astuce pour que BayesElo puisse jouer le championnat d'après les précédentes données..ça représente une tache de forçatsi il faut remonter sur 5 ans et peu fiable sur 2 ou moins, mais BayesElo aurait des données passées sur lequelles travailler. D''abord match par match sur toute la saison toutes les équipes sur u exel (ma graph 75 tiendra pas et ne pondra jamais de PGN. UNe fois la journée de championat faite, j'ai récupéré sur l'ancien site ClubElo le calcul de l'avantage à domicile, le calcul d'échange des points est connu des amateur de Elo et donc simple à mettre en shell...Le problème , mais en étant ingenieux ça pourrait se réaliser, au pire on recopie un PGN sur Bloc Note...et voilà, BayesElo est enfi nourri d'anciens Elos. CE serait la solution qui sortirait de mes pavés dont je m'excuse si ils vous paraissent trop lourd...
SI la loi du BayesElo est de ne pas prendere le champs Elo du PGN c'est en fait compliqué. J'ai testé arena il me semble, onpeut généerer avec plusieurs moteurs d'échecs des saisonsq de championat. Le problème est que le match est interminable et on doit choisir qui gagne.Pir, on peut attribuer le Elo de l'équipe mais il à 2000 ès qu'elle joue, un truc doit m'échaper...L'idée que j'avais était de créer un tournoi de hcampionat genre avec arena et l'importer, en PNG dans bayesELO
Le Elo ne donne pas la forme actuelle du joueur (qui ne peut compter que sur lui ) ou de l'équipe (un malade et hop, un remplaçant) ,mais elle est capable, avec l'argent qui est relatif à un niveau année par année, bien qu'il y ait des surprises, garder u niveau chaque année.
En effet jm'ocupe à trouver le gagnant...
Pour le football par exemple je vais surClubElo, qui classe depuis l'historique des matchs des années 1950 ou 1960..le classement Elo est bien affiné et c'est bien dommage car sur son ancien site il expliquait bien et avait quelques idées comme le calcul du Home Field Avantage (l'avantage des blancs un peu..il est en moyenne autours de 50 Pts Elo) Et un formule pour trouver un nombre de buts estimés dans le but d'une distribution de Poisson..l'effort était louable et marche pour les équipes au niveau très similaires...je préfère me tenir aux buts pour but ontre de chaque équpe avec ma sauce pour la distribution..j'ai mis la formule sur un tableur de calculatrice et ajouté , avec le coef de ma trouvaille, lea formule pour trouer la probabilité d'un match nul d'ABROBecker.je fais une régression linéaire comme le Elo, la forme d'une équipe appartient au passé-ou au présent-,et à vrai dire même si j'admire ce Elo, une formeou méforme subite est lente à déceler, mais j'ai une parade qui n'a rien à voir avec Elo donc pas sujet ici, puis j'effectue une distribution de poisson. ClubElo est mis à jour régulièrement, très fiables pour les clubs de nations bien représentées en UEFA...Pour certains pays qui n'envoie qu'un club , certes champion dans son pays, se faire exploser le Elo du club est compté mais certains petits pays européens (exclusivement européen) le Elo des clubs peut ne pas exister...qui pour miser sur Tiraspol ? Il y a unebonne vingtaine de ligues où c'est bien tenu...A peu de chose près celles qui sont autorisées à la mise en France.
Pour l'international EloRatings s'occupe plus des sélections nationales (d'ailleurs le système du foot féminin est très fortement basé sur le Elo avec perte de points si par exemple une équipe devrait sur la différence Elo gagne 1-0 au lieu de 3-0)...EloRatings fait figure sérieuse, c'est ce qu'est aux selections le Club Elo des clubs. On en trouve d'autres 538 et son mélange à base d'ESPN, mais secret...Le Elo club ELO
dévoilait ses formules et techniques.On pouvait trouver le lien caché dans un ici en fait j'éditeil existe là http://clubelo.com/System ..IL me reste des notes prises si quelq'un est interessé : calcul de la moyenne de ligue de l'avantage à domicile ( plus les pays sont petit, plus ce coef est faible)...calcul du nombre de but je m'en sert, par exemple pour un distribution de Poisson car l'écart de la formule d'Abrobrecker,qui donne une proba de nul adapté aux échecs, mais avec avec mes coefs ( adapté au foot le nombre de draw est pas le même, a tendance à soit eê^tre serrée, soit exagérée...la moyenne des deux me donne environ les mêmes probas que les bookmakers, parfois -;) .
je dois encore bosser les maximisation minimisation pour tenterd'ameliorer le Elo, mais je doute que ma casio gragph75 y arrive...Par contre Bayeselo a , je pense, du potentiel et je me ferai d'abord la main pour ensuite aller plis loin 'j'espère.Car BayesElo a un tas de fonctionalités, mais si il faut 6 mois un an de cacalcul d elo - et je compte pas les autres équipes (je fait des regressions linéaire puis distribution de Poisson et ça prend du temps pour les équipes.Mais terriblement plus efficace pour trouver les outsiders le pire c'est que je joue pas ou plus au tabac je m'amuse avec des mises de 0,1 Euros qui se transforment en euros ou en dizaines d'euros...c'est que je veux pas me ruiner,c'est l'exercice intellectuel qui me motive).
J'espère revenir avec des bonnes nouvelle, pour ça il faudra travailler dur ce que j'ignore ! et j'espère en apprendre un peu plus en passant lire , même si mon objectif ici est de simuler un championnat de ligue ( foot ou autre ).ET beaucoup de temps, de patience pour taper au moins un an ou deux de 380 matchs, et calculer les Elos et points gagnés sur ma calculatrice...J'essaierai avec BayesElo en commençant la saison dernière avec 1500 en L1 et 1400 en L2. puis rebelote la semaine suivante- non je suis pas sérieux dès que possible..C'est peut-etre la solution^..Quand on veut queqlquechose faut savoir se retrousser les manches...Si BayesElo ne peut pas ajuster à chaque équipe dans un championnat sur l'année, que Arena met un temps fou, permet de changer manuellement les ELos mais au début du tournoi remet à 2000, et pour finir pour 380 matchs au bout de 5 minutes on choisit le gagnant, ce qui n'et pas mon but.
Voilà la solution : rejoueur (j'ai gardé les elos mais seulement celui de mon club), par exemple cette saison en notant les Elos de chaque équipe...Calculer les points à échanger ou non et calcul du nouvel Elo avec tableur...puis trouver l'astuce pour que BayesElo puisse jouer le championnat d'après les précédentes données..ça représente une tache de forçatsi il faut remonter sur 5 ans et peu fiable sur 2 ou moins, mais BayesElo aurait des données passées sur lequelles travailler. D''abord match par match sur toute la saison toutes les équipes sur u exel (ma graph 75 tiendra pas et ne pondra jamais de PGN. UNe fois la journée de championat faite, j'ai récupéré sur l'ancien site ClubElo le calcul de l'avantage à domicile, le calcul d'échange des points est connu des amateur de Elo et donc simple à mettre en shell...Le problème , mais en étant ingenieux ça pourrait se réaliser, au pire on recopie un PGN sur Bloc Note...et voilà, BayesElo est enfi nourri d'anciens Elos. CE serait la solution qui sortirait de mes pavés dont je m'excuse si ils vous paraissent trop lourd...
Re: Bayesian Elo rating
La solution serai de créer un PGN de routes les rencontrer de la saison passée pour que Bayeselo anticipe les résultats de rencontres future...un travaille fourmi... à la base je pensais attribuer un Elo à une équipe en début de saison puis plus tard, afin de voir quelles équipes progressent et d'autres régressent.
Je comprends qu'un PGN de la saison passée est la solution la plus simple, même si ça peut être fastidieux.
Merci pour l' éclairage se mes tatonnements.
Je comprends qu'un PGN de la saison passée est la solution la plus simple, même si ça peut être fastidieux.
Merci pour l' éclairage se mes tatonnements.