2014.06.06 Stockfish 5 with 4pc Syzygy bases added. The engine gained 6 Elo (in the one on one TOP 16 comparision) and just passed H4. With this the IPON-RRRL has a new No.1! It is that close that the next new entry might change this! New anchor for the lists is Stockfish 5 Syzygy with 3113 Elo. For the "experts": I added Bayes, Ordo and Elostat results in the IPON Archive. 2014.06.01 Stockfish 5 entered the complete list with a very impressiv plus of 40 Elo to Stockfish DD and ended just 5 Elo below H4. As a little remark I have to say that ELOSTAT would have Stockfish in front in the Complete List and in the RRRL. It is very close. 2014.05.31 A new No. 12 entered the IPON-RRRL. Texel 1.04 with a rating of 2835 Elo. Congratulations! As Texel pushed out my reference engine of the TOP 16 I will try a new method of "reference". It is always the No. 1 which is giving the reference point. Today it is Houdini 4 with 3111 Elo. As soon as a an engine passes the No.1 it initial rating became the new reference point. With this I always have a reference and the ratings hopefully doesn't change to much. My "theoretical" gut feeling tell me that over time (which means 2, 3, 4 new No1 engines) the overall rating will drop a bit e.g. the old reference, Shredder 12, will drop below 2800 Elo. Most important is the difference of the engines and that is not different than with any other reference point. 2014.05.26 Komodo 7a now with all games in the Main List. A new No.2 with a very impressive 32 Elo jump! I got requests about statistics for individual results. Please check the Archive section. It is all available there. 2014.05.25 Komodo 7a included in the Complete List. 31 Elo increase over K-TCECr Statistics and Main List will be updated after the last games. 2014.05.04 Protector 1.6.0 added to all lists. 33 Elo better than the last release. Rank 11 in the IPON-RRRL! 2014.04.20 Finished the missing games of Gull 3 vs DF14 and made a new main list. No. 2, 3 and 4 are all within one SD basically they are that close that their playing strength is hard to distinguish. 2014.04.19 Gull 3 added to the complete list. Nice increase of 40 Elo and a new No. 3. The IPON-RRRL will follow as soon as the missing games are finished. 2014.03.20 Senpai 1.0 included. An entry of 2841 Elo for the main list. Impressive first release! As Senpai is pushing a number 16 to 17 the last engine will be excluded out of the TOP 16 enignes. It was very close but it is Depp Sjeng c't 2010, an engine which I really liked to have in my list. This would be a TOP 17 List of the IPON: 1 Houdini 4 3119 9 9 3520 77% 2906 29% 2 Stockfish DD 3072 8 8 3520 72% 2909 41% 3 Komodo TCECr 3057 8 8 3520 70% 2909 38% 4 Gull 2.8 3023 8 8 3520 65% 2912 40% 5 Critter 1.4a 2982 8 8 3520 59% 2914 46% 6 Equinox 2.02 2978 8 8 3520 59% 2914 46% 7 Deep Rybka 4.1 2968 8 8 3520 57% 2915 47% 8 Deep Fritz 14 2901 8 8 3520 47% 2919 45% 9 Chiron 2 2893 8 8 3520 46% 2920 45% 10 Hannibal 1.4b 2875 8 8 3520 44% 2921 43% 11 Senpai 1.0 2843 8 8 3520 39% 2923 41% 12 Naum 4.2 2838 8 8 3520 39% 2923 41% 13 Protector 1.5.0 2836 8 8 3520 38% 2923 43% 14 HIARCS 14 WCSC 32b 2822 8 8 3520 36% 2924 40% 15 Jonny 6.00 2805 8 8 3520 34% 2925 37% 16 Deep Shredder 12 2800 8 8 3520 33% 2926 38% 17 Deep Sjeng c't 2010 32b 2798 8 8 3520 33% 2926 39% 2014.03.17 Gull 2.8 included. 3023 Elo in the main list. Ranked 4th! 2014.03.14 All counters removed - No vanity, no liabilities! 2014.01.05 1. Equinox 2.02 included. 2978 Elo in the MAIN lst. 2. For historical reasons the 4 year old Robbolito 0.085g is included. It is possible to compare it with the alleged source but keep in mind that Robbo did not play the same opponents. It's average opponent elo is much higher. As "the source" is playing with a contempt there is the possibility that it would lose more points when playing the same strong engines. 3. The IPON is fixed to 2800 elo for Shredder 12. This is an offset of 2783 over ALL engines. I applied this offset to the three ECO based lists. This is more accurate than having a fixed engine there. I will do that principle with all ECO lists in the future. 2014.01.03 The three rating lists sorted by opening system have to be recalculated. It makes no seance to fix this with a certain engine as this might cover an in- or decrease of another engine in case the fixed engines is weaker or stronger in one of the lists. It is better to give the lists the same average elo value as all list have the same engines. This way it is just the distribution which counts, not the relative distance to a specific engine. Unfortunately I deleted the individual PGN data for the lists. I will remodel the list with the next engine I include. 2014.01.01 I restarted the IPON 2014 with some changes. 1. The old 75 opening positions set is increased by 35 new position to 110 opening positions. The ECO distribution is 21% open games (ECO C20-C99), 30% half open games (ECO B + C00-C19) and 49% closed games (ECO A,D, E). This is less 'GM tourney' practice but more suited to the average chess player and his analytical needs. 2. Only 16 top engines are tested. The reasons are a smaller width of opponent elo and less games to play. Additionally the error bar is below 10 Elo now for all top 16 engines in the IPON-RRRL. These 10 elo are the "border of irrelevance" for me, as no one can feel or distinguish a + or - of 10 elo. (Having smaller error margins in a testing environment is a different thing of course). 3. I will provide live games from time to time but not necessarily the day an engine is released ... With 110 opening positions and a O20/H30/C50 distribution the individual comparison of 220 games becomes interesting to a certain extend. It should not be stretched to much, but 220 games can be decisive in some cases. In the rating section I offer three new lists sorted by closed, half open and open openings. This is just an experiment to show some interesting trends. It should not be taken too serious as the number of games as well as the number of used openings is limited. Nonetheless, in some cases it might give a hint what can happen if a certain opening distribution is used (here and somewhere else). It shows too, that there is no 'right' or 'wrong' set of openings. Everything is correct and the observer has to ask if a particular set-up suits his needs! This is a free service. If you don't like it please have a look at some of the other excellent lists. There are plenty available to satisfy anyone! Bye Ingo
The chess engine decides what moves to make, but typically does not interact directly with the user. Most chess engines do not have their own graphical user interface (GUI) but are rather console applications that communicate with a GUI such as XBoard, WinBoardor glChess via a standard protocol. This allows the user to play against multiple engines without learning a new user interface for each, and allows different engines to play against each other.
The command-line interface of GNU Chess became the initial de facto standard, called the Chess Engine Communication Protocol and was first supported by XBoard. When XBoard was ported to the Windows operating system as WinBoard, this protocol was popularly renamed to 'WinBoard Protocol'. The WinBoard Protocol was itself upgraded and the two versions of the protocols are referred to as: 'WinBoard Protocol 1' (original version) and 'WinBoard Protocol 2' (newer version). There is another protocol, the Universal Chess Interface. Some engines support both major protocols, and each protocol has its supporters. The Winboard Protocol is more popular but many chess engine developers feel that the Universal Chess Interface is easier to implement. Some interfaces support both protocols, while others, such as WinBoard, support only one and depend on subsidiary interpreters to translate.
Chess engines increase in playing strength each year. This is partly due to the increase in processing power that enables calculations to be made to ever greater depths in a given time. In addition, programming techniques have improved, enabling the engines to be more selective in the lines that they analyse and to acquire a better positional understanding.
Some chess engines use endgame tablebases to increase their playing strength during the endgame. An endgame tablebase is a database of all possible endgame positions with small groups of material. Each position is conclusively determined as a win, loss, or draw for the player whose turn it is to move, and the number of moves to the end with best play by both sides. Endgame tablebases in all cases identify the absolute best move in all positions included (identifying the move that wins fastest against perfect defense, or the move that loses slowest against optimal opposition). Such tablebases are available for all positions containing three to six pieces (counting the kings) and for some seven-piece combinations. When the maneuvering in an ending to achieve an irreversible improvement takes more moves than the horizon of calculation of a chess engine, an engine is not guaranteed to find the best move without the use of an endgame tablebase, and in many cases can fall foul of the fifty-move rule as a result.
Many engines use permanent brain as a method to increase their strength.
The results of computer tournaments give one view of the relative strengths of chess engines. However, tournaments do not play a statistically significant number of games for accurate strength determination. In fact, the number of games that need to be played between fairly evenly matched engines, in order to achieve significance, runs into the thousands and is, therefore, impractical within the framework of a tournament. Most tournaments also allow any types of hardware, so only engine/hardware combinations are being compared.
Historically, commercial programs have been the strongest engines. To some extent, this is a self-fulfilling prophecy; if an amateur engine wins a tournament or otherwise performs well (for example, Zappa in 2005), then it is quickly commercialized. Titles gained in these tournaments garner much prestige for the winning programs, and are thus used for marketing purposes.
Chess engine rating lists aim to provide statistically significant measures of relative engine strength. These lists play multiple games between engines on standard hardware platforms, so that processor differences are factored out. Some also standardize the opening books, in an attempt to measure the strength differences of the engines only. These lists not only provide a ranking, but also margins of error on the given ratings. Also rating lists typically play games continuously, publishing many updates per year, compared to tournaments which only take place annually.
There are a number of factors that vary among the chess engine rating lists:
These differences affect the results, and make direct comparisons between rating lists difficult. All listed engines are 64-bit.
|Rating list||Time control|
|Top three engines||Rating|
|2005||March 1, 2014||1427||500,530||Houdini 4 x64 4CPU|
Stockfish DD x64 4CPU
Komodo TCEC x64 4CPU
|2006||March 2, 2014||1198||723,221||Houdini 4 x64 4CPU|
Stockfish DD x64 4CPU
Komodo TCEC x64 4CPU
|2006||January 5, 2014||134||295,420||Houdini 4|
|2009||January 31, 2014||21||17,500||Houdini 4 x64 1CPU|
Stockfish 4 130512-beta x64 1CPU
Komodo TCEC x64 1CPU
|SSDF||40/120 --> 20/60|
|1984||November 11, 2013||337||132,746||Komodo 5.1 MP x64 2GB Q6600 2.4 GHz|
Stockfish 3 MP x64 2GB Q6600 2.4 GHz
Deep Rybka 4 x64 2GB Q6600 2.4 GHz
These ratings, although calculated by using the Elo system (or similar rating methods), have no direct relation to FIDE Elo ratings or to other chess federation ratings of human players. Except for some man versus machine games which the SSDF had organized many years ago (which were far from today's level), there is no calibration between any of these rating lists and player pools. Hence, the results which matter are the ranks and the differences between the ratings, not the absolute level of the numbers. Also, each list calibrates their Elo via a different method. Therefore no Elo comparisons can be made between the lists. Nevertheless, in view of recent man versus machine matches, it is generally undisputed that top computer chess engines should be rated at least in the range of top human performances, and probably significantly higher.
Missing from many rating lists are IPPOLIT and its derivatives (e.g. Fire). Although very strong and open source, there are allegations from commercial software interests that they were derived from disassembled binary of Rybka. Due to the controversy, all these engines have been blacklisted from many tournaments and rating lists. Rybka in turn was accused of being based on Fruit, and in June 2011, the ICGA formally claimed Rybka was derived from Fruit and Crafty and banned Rybka from the International Computer Games Association World Computer Chess Championship, and revoked its previous victories (2007, 2008, 2009, and 2010) The ICGA was criticized for this decision by Dr. Søren Riis, a longstanding supporter of the Rybka program. Rybka is still included on several society ranking lists.