Charles Roberson – President
John Timmel – First Vice President
Grant Oen – Scholastic Vice President
Wayne H. Spon – Secretary/Treasurer
Dr. Walter High
Charles Roberson – Editor
Web support, edit, and distribution
North Carolina Gambit – April 2022
In this issue, you will find: Walter - The effect of titled players in North Carolina Wayne - Soviet Chess Domination during the Cold War Charles - Monthly Computer Chess article- Training with and Using Chess Computers and an Interview of Yang Yu, first grade National and NC K-1 State Chess Champion John - this month's "To Move and Win" problem
We’re looking for contributors to the Gambit!
Submit games (pgn preferred), articles, etc. to the editor: firstname.lastname@example.org
CHESS NUGGETS – April 2022
A Monthly Blog for North Carolina Chess Players The effect of titled players in North Carolina
by Walter High (email@example.com)
Beginning in 2012 the Carolinas Chess Initiative annually bid to host the United States Masters Chess Championship. The motive behind placing this bid was a plan to gradually increase the rating levels of top players in North Carolina. There were two aspects to this plan: (1) to bring GMs, IMs, and FMs to our state to play, and (2) to raise the ratings of North Carolina players by giving them exposure to competition that had previous not been available within the state borders. The hope was that we could eventually encourage a few FIDE-titled players to move to NC with the idea that they would teach chess to our up-and-coming players. We also wanted to give our local players the opportunity to face much stronger competition without having to travel out of state.
When I became part of the NC chess community in 2006, there were no Grandmasters with a North Carolina affiliation. Mikhail Nepomnyashchiy and Jonathan Schroer were the only International Masters, but Nepomnyashchiy had retired from playing and Schroer only played sporadically in the next few years before retiring. Both were active teachers of chess, however. The state scholastic ranks only showed a few participants who managed to exceed the 2000 US Chess rating level (Expert). There had been a few shooting stars, like Mike Klein in the 1990s and Nelson Lopez in the early 2000s. Klein became an FM and was 2300 in high school and Lopez won the Denker Tournament of State High School Champions, but there had been no consistent production of strong competitors within the state.
Roll forward to 2022. We now have two strong chess teaching centers in the Charlotte Chess Center (CCC) and the Triangle Chess Center. The continued presence of the U.S. Masters Championship in the state and the strength of the Charlotte Chess Center eventually attracted a number of GMs and IMs to move to North Carolina. We now have many more strong teachers for our chess population. GM Elshan Moradiabadi and WGM Sabina Foisor were to first to make NC home. GMs Daniel Naroditsky and Magesh Panchanathan followed. Then IMs Levan Bregadze and Dean Ippolito came also. Every one of these titled players makes a substantial part of their living by teaching chess here.
The results have been spectacular as we now have 46 players with NC affiliations who are at least National Masters (2200 rating – US Chess). Steve Wang (2429), Daniel Cremisi (2387), Peter Giannatos (2367), Aaron Balleisen (2356), Karthik Rangarajan (2338), and Vishnu Vanapalli (2323) have all exceed 2300 in the past few years. In 2009 the State Scholastic Championship had 2 players a little over 2000 rating; this year there were 5 players with National Master titles. GM Elshan Moradiabadi had students who won both the K-12 and K-8 titles this year and will have five of his students nationwide qualified for the Denker Tournament which must be some kind of record. Our student players get exposure to these FIDE-titled instructors at our local tournaments and then sign up for private lessons. The CCC not only gives hundreds of players exposure to high level instruction, it has also hosted dozens of GM/IM norm tournaments which has led to North Carolina probably leading the nation in title norm production. Strong players from all over the US now want to come play tournaments in North Carolina which provides a rich environment for our students to learn and progress.
I offer my sincere congratulations to all the strong NC players who have come up through the ranks over the last decade. I know there are many more in the pipeline and I look forward to watching them prove that North Carolina is one of the best places in the US to learn and play chess.
Soviet Chess Domination during the Cold War by Wayne H. Spon
During the cold war from 1948-1991, the Soviet Union dominated the chess world except for the brief reign of Bobby Fischer from 1972-1975. So what led to this domination? The royal game was always popular in Czarist Russia but mainly with the elites. The Grandmaster title was formally bestowed by Czar Nicholas II at the 1914 St. Petersburg Tournament to the top 5 finalists. However, the working class, referred to as the proletariat in communist terms, did not have the luxury of leisure activities such as chess. After the Russian Civil War from 1917-1921, Vladimir Lenin, an avid chess player, established the USSR and sought to modernize his country economically from the feudal czarist Russia. The Soviets disavowed pop culture in favor of ballet, classical music, realistic art and chess. Control and education of the population began in 1922 with the pioneer school system sponsored by Lenin’s wife. Both Nikolai Krylenko and Alexander Fedorovich Ilyin-Zhenevsky were instrumental in creating the Russian chess machine.
Krylenko was the commander of the Soviet armed forces and chief prosecutor for Russian tribunals. He was also the principal organizer of the Moscow International tournaments in 1925, 1935 and 1936. He encouraged chess among the military; stressing its value as a means of learning discipline, strategy, persistence and caution. Ilyin-Zhenevsky was a chess master who won the 1914 Geneva tournament. He returned to Russia after the revolution to support Lenin’s ideals. Both men scouted children with an aptitude for chess and instilled in them the communist ideology. Between 1924 and 1929, the number of registered chess players in the USSR increased from 24,000 to 150,000. Five years later, it was half a million.
The Soviet chess schools, known as Pioneer Palaces, were referred to as “a socialist chess incubator”. Chess became a political weapon for international competition between the western capitalist countries and the USSR by claiming intellectual superiority in a non military threat. Chess was also the cheapest way to provide culture, sport, leisure and intellectual development. Aside from international chess tournaments, the USSR did not compete internationally till 1952 when they competed in the Olympics. Soviet chess players were government funded and the only “athletes” allowed to travel outside of the USSR prior to 1952.
The first product of the Soviet chess school to become the World Chess Champion was Mikhail Botvinnik in 1948. Although Alexander Alekhine was Russian and designated as the World Chess Champion till he died in 1946, the Soviets did not recognize him because he fled Russia after the revolution and publicly denounced the Bolsheviks. He was also a Nazi collaborator. Ironically, the Soviets rehabilitated his status in the 50’s. Prior to 1946, the World Chess Champion match was privately organized between the challenger and the current champion. In 1946, the International Chess Federation (FIDE) formally took over the World Chess Championship title by organizing the AVRO match tournament of the six leading players to determine the World Champion. Euwe, Fine, Reshevsky, Botvinnik , Keres, and Smyslov played against each other four times. Botvinnik eventually won the championship with 14 points followed by Smyslov with 11. Both were students of Krylenko’s Pioneer Palace for Soviet chess. There was a rumor that Keres, who was from Estonia which was behind the iron curtain, agreed not to finish ahead of Botvinnik. Years later, Keres was asked why he was never a world champion. The modest Paul Keres replied, “I was unlucky”. After a pause he added, “I love my country”. Both Krylenko and Ilyin-Zhenevsky, the founders of the Soviet chess school, did not live to see their achievement. Krylenko was charged with “retarding the development of chess”. Sometime between 1938-1940, Krylenko was executed on orders from Stalin. Ilyin-Zhenevsky escaped the purges but died in a German bombing raid in 1941.
From 1948-1972, the Soviet Union controlled the World Chess Championship with six different champions. However, Mikhail Tal was from Latvia and Tigran Petrosian was Armenian. Both countries were behind the Iron Curtain so the USSR took credit for their titles. The famed cold war match between Spassky and Fischer was considered a humiliation in the USSR. In 1976, Spassky immigrated to France and became a French citizen in 1978. In 2012, he returned to Russia which was no longer the USSR. When Fischer refused to defend his title over match conditions in 1975, Anatoly Karpov assumed the title by default. Karpov was another product of the Soviet chess machine. He is currently serving in the Russian State Duma and has been sanctioned by the EU for the Russian invasion of Ukraine. The FIDE Council has also suspended his title of FIDE Ambassador for Life.
Kasparov is the last student of the Soviet chess school to win the FIDE World Chess Championship in 1985 but he did without Soviet backing because Karpov was their favorite son. Kasparov held the official FIDE world title until 1993 when a dispute with FIDE led him to establish a rival organization, the Professional Chess Association (PCA). In 1993, he won the PCA Classical World championship against Nigel Short but relinquished his title to Vladimir Kramnik in 2000. Since retiring, he has devoted his time to politics and writing. He is a member of The Other Russia which is a coalition opposing the administration and policies of Vladimir Putin.
The collapse of the Soviet Union in 1991 has also leveled the playing field for the rest of the chess world. Although the Russian Federation currently has the most Grandmasters in the world, government support is not reliable. The internet is also a contributing factor because of chess engines and on-line competition. There are more chess coaches, tournaments and clubs to attract new and current chess players. Approximately 30 years ago, the USCF made a concerted effort to grow scholastic chess. The chess world is catching up to the Russian juggernaut.
References: The Kings Of Chess by William Harston
Numerous Wikipedia articles
Training with and Using Chess Computers
Commonly Asked Questions and Misconceptions in Computer Chess
by Charles Roberson
This month’s article focuses on several commonly asked questions and misconceptions about chess programs. I’ve tried to keep the answers from being too technical and yet too simple. Often these answers are so simplistic to give a simple understanding that they give an incorrect understanding.
Computer Chess Ratings
- Are computer chess ratings accurate?
If we look at the CCRL rating list, we see three rating columns: one for Elo and the other two labeled “+” and “-“. The first is their Elo estimate and the next two are the +/- error margins. All we know is that the rating is estimated to be somewhere near that Elo number within the error margin. It could be at the top, bottom or somewhere in between.
- Which Program is the strongest? Is it stockfish?
On the CCRL, Stockfish is currently list #1, however it is only 15 points ahead of “Dragon by Komodo” and the error margin for Stockfish is +/- 19 with a +/- 20 error margin for Komodo. Given that these rating estimates overlap each other, we are not sure which one is best.
- How are programs rated?
A large number of games are played with various programs against other programs by a list of volunteers for each rating group. Regularly all the new games are added to a database. The new database games are now rerated typically using bayeselo – a Bayesian statistics based Elo rating program. This produces ratings from 0 to what ever. Now, a single program is picked to assign a FIDE Elo estimate. After that, all the other ratings are adjusted accordingly. The program picked is usually one close to the top.
- Is the engine playing strength the same on all platforms (my computer, laptop, cell phone)?
No. A computer chess program can gain performance with access to faster CPUs, more CPUs and faster/more memory. The baseline computer for the CCRL is an Intel i7-4770k. They limit each program to around 200 MB (this is archaically small). Lets say your computer is twice as fast as that and you can give the transposition tables 16 GB of memory, then the same program on your computer could be 150 Elo or more stronger than their estimate. Likewise, if you have a weaker computer with less memory then the same program could be significantly weaker than their estimate. So, you can NOT give a chess program a static rating without considering the hardware it is running on. Thus, the same program doesn’t have the same playing strength on all platforms/devices.
- Can I improve the strength of a program on my computer?
Yes, there are some parameters you can adjust for many programs.
- Ponder on: Most rating lists turn this off, but turning it on will add strength. It allows it to think on your time, but not useful in analysis mode.
- Transposition Table Memory: Transposition Tables are used to remember the eval of a previously seen position. When a position is seen again, then the saved eval is used and that position is not searched again.
- Endgame info: Endgame Tablebases are used by most rating lists and take effect when the position is down to 6 pieces (both sides total). Some programs gain significantly from these. Not using them weakens the program.
- Threads: All of the top programs are tested and rated using 4 CPUs. This increases their playing strength relative to using just 1 CPU. If you have more than 4, you could gain strength by using more of them. Often the program default is 1. Changing that will increase the strength. If you have only one or two CPUS then the program will be weaker on your computer than the one tested with 4 CPUs.
- How weak are some programs?
Some chess programs are very weak. The weakest one on the CCRL 40/15 (40 moves in 15 minutes) list is BACE 0.46. It is 490th on the list with an estimated FIDE rating of 1376. The weakest chess program on their Blitz (2 minutes +1 second) list is SxRandom 1 which is 736th on the list with an estimated FIDE rating of 188. One above that is ACE 0.1 which an estimated ratting of 329. Remember all of these rating estimates are on the target hardware at the target time control, thus the chess programs could be stronger or weaker on your computer.
Common Computer Chess Questions and Misconceptions
- How long does it take to code a GrandMaster level chess program? Can it be done in a weekend?
Assuming you don’t cheat by copying somebody else’s work and slapping your name on it, it could take quite a while. Think of it this way. A typical GM level program can be at least 10,000 lines of code or more. In paperback form, that would be around 330 pages of code. So, how long would it take you to type 330 pages of anything and make it grammatically and syntactically correct? Now, answer that question again considering you are writing highly detailed instructions on how to do something very complicated.
- Is there a good way to use a chess program as a training partner at nearly any strength?
Yes, I covered that in detail in the February 2022 and March 2022 newsletters.
- Chess has been solved / Alpha Zero plays perfectly…
No. Chess has not been solved and we don’t think Alpha Zero plays perfectly. Checkers has been solved and it was solved correctly: every possible checkers game was played out and looked at such that perfect play for both sides is known. Due to this we know, checkers is a draw with perfect play. To do the same analysis on chess would require thousands of years of compute time with a super computer.
- The best program for analysis is the one at the top of the rating list.
The best program for analysis is the one that you can understand and produces accurate results. Most people don’t understand Stockfish’s analysis. However, some claim that LC0 (an attempt at replicating Alpha Zero) produces understandable analysis. Some claim that Komodo’s analysis is more human and more understandable. So, the answer is not the same for everybody.
- I just ran 10 games and A beat B 9 of them so A is better.
Maybe and maybe not. It takes a lot more than 10 games to get statistically significant results. If you want to estimate a program to the accuracy of +/- 1 Elo then you will have to run it through around 32,000 games. In many testing efforts over the last several decades, I’ve seen several times when A beat B 9 out of the first 10, but then ended up being worse. Running just 10 games is statistically insufficient.
- Do programs search all the way to the end of the game for middlegame positions or opening positions?
No, not typically. Programs will search as deep as they can given the time constraints, but they will not see all the way to the end of the game except in nearly completed games. Given that they search deeply but not all the way to the end, there exist a chance that they could be wrong. The deeper they search the lesser the chance that they are wrong but the possibility still exists. This issue is called the “horizon effect” in Game Theory. You can look as far as you want, but you never know what is just over the horizon unless you’ve already been there.
- What does a score of -1.87 really mean?
All chess AIs do their math in whole numbers for speed. It is much faster to do integer math (187) than floating point math (1.87). By the UCI protocol (Universal Chess Interface – governing communication between chess AIs and chess GUIs (Graphical User Interfaces)), chess AIs are required to report everything in “centipawns” thus 1.87 pawns is reported as 187. The GUI then divides the value by 100 and reports 1.87. So, 1.87 means that white is up by 1.87 or will be up by 1.87 pawns if predicted play is followed. A score of -1.87 means that black is up by 1.87 pawns. Likewise a score of +5.42 means white is up or will be up by the equivalent of 5.42 pawns. A score of 5.42 pawns could mean a bishop and 2.42 pawns, a rook and 0.42 pawns, a rook and a pawn but down positionally by 0.58 pawns or any one of several other combinations.
- How can a position be up by +0.42 pawns?
A chess AI will include many things other than raw material count in its eval. It will include many positional features such as space, mobility, passed pawns, isolated pawns, doubled pawns … Each of these features can have a positive or negative effect on the position evaluation. Remember a positive for black is a negative for white and vice versa. So, a score of +0.42 could mean several things:
- Material will be even at the end of the predicted line plus white has some positional advantages amounting to 0.42.
- Material will be even at the end of the predicted line plus black has some positional penalties amounting to -0.42 for black which is +0.42 for white.
- At the end of the predicted line, white will be up a pawn but down 0.58 worth of positional penalties (1.0 – 0.58 = +0.42).
- At the end of the predicted line, white will be up a pawn but black will be up 0.58 worth of positional bonuses: thus +1.0 – 0.58 = +0.42.
- At the end of the predicted line, white will be up a pawn but black will be up 0.28 worth of positional bonuses and white down 0.30 in positional penalties: thus +1.0 – 0.28 – 0.30 = +0.42.
- And many more combinations …
- How did it come up with the number?
A chess AI does two things: search and static eval. The two together create the dynamic eval that the chess AI returns. Simplistically stated, the search looks at many possible lines of play from the current position. At the end of each line, a Quiescent Search (Q-search) is performed to eliminate all good captures from the position. The moves in each line are the ones considered to be the best responses for both sides. Never expect your opponents to make a bad move, but be ready to punish them if they do. Now, the static eval is run on the new position at the end of a line. This is considered the static value of that end position but the dynamic value of that line. This is done for each line. The line with the best dynamic value is considered the dynamic value of the current position.
- How can I see what it bases that on?
All programs can produce a PV (primary variation) for a position. This is it’s planned line of thought. If you start at the current position and play out all of the moves in the PV. Then you can see the key position that produced the key static eval. Most chess GUI’s have a clickable button which will show that position for you. This position is considered the best achievable position for the side to move given that both sides will play their best moves.
- How important is nodes per second?
Nodes per second is a general Graph Theory term. In chess AIs, it is just chess positions per second. The strength of a chess AI comes from a combination of many things. Simplified it comes down to the knowledge vs speed trade off. It is a trade off because adding knowledge usually slows down the positions per second and reducing knowledge often increases positions per second. But, sometimes a reduction in speed allows so much more accuracy that the chess AI prunes the search tree significantly better resulting in smaller tree that more than compensates for the reduction in positions per second.
Example: lets say a program can look at 10 million positions per second and we make it smarter but that cuts the speed to 8 million positions per second. Have we made it worse or better? It is smarter but it is slower. Assuming we didn’t make any mistakes in our programming and the tree size for many test positions drops by say 50%, then we took a 20% decrease in speed but a 2x reduction in the number of positions in the typical search tree. Thus, the new version is stronger.
- Programs get all their strength from depth of search, thus they only improve via speed?
No. There have been numerous improvements to chess knowledge in chess programs since the 1970s and lots of that in the past 2 decades. Much of a programs playing strength comes from that knowledge. Of course, speed helps and increased depth also increases playing strength. The early programs from Fidelity and others didn’t even have pawn structure knowledge; they mostly just counted material. It was fairly easy to get an outside passed pawn, trade off pieces, then have an uncatchable passer to win the game. Part of that was due to not having pawn structure knowledge and part of it was due to lack of search depth. In the following years, pawn structure knowledge was added: passers, connected passers, isolated pawns and doubled pawns, but speed also improved. Many people just noticed the speed improvement and attributed all the strength gains to that.
About 15 years ago when programs were around 2700, I devised a test to see how much playing strength came from knowledge increases alone and I published my results on talkchess.com. Dr. Bob Hyatt noticed it and emailed me questioning the methodology. He decided that was too good not to try and he did using a super computer to run around 32,000 games with modifications to his program Crafty. His results concurred with mine that the knowledge accrued was worth at least 600 Elo out of the 2700. Back then a 2700 Elo program without all that knowledge would be faster, but 600 Elo weaker resulting in a 2100 rated program. So, speed isn’t everything and complexity can make things better.
NC Chess Interview
Interview with Yang Yu and his parents by Charles Roberson
Yang Yu is a NC chess player in the first grade. This year he won the first grade National Chess Championships and the NC K-1 State Chess Championships. NC Chess interviewed Yang and his parents for this month’s newsletter.
|What do your parents think of chess, you playing chess, tournaments …?
We think chess is great for Yang. It helps him to do calculations, think logically and keep focused; more importantly, playing chess helps him to learn how to deal with loss.
Do your parents have a favorite tournament?
We prefer local OTB tournaments and the free USCF-rated online tournaments on chess.com.
Do your parents have advice for other players and/or parents?
Keep the kids interested and remind them to practice. It’s important to keep practicing, Yang has done 27000+ puzzles and played about 2500 games by now.
Do you have a favorite tournament?
All sorts of tournaments at Coach Bill’s Triangle Chess.
How does it feel to be a national champion?
I am happy.
How long have you been striving for it?
I think it’s pretty long, 3-4 months.
Were you really sure you were going to achieve it this time?
I was very sure I could achieve it.
How do you celebrate a victory?
Nothing special, I am just be happy.
How do you deal with a loss?
I take a break and try to win the next game.
What other titles have you won?
NC K-1 champion for 2021 and 2022.
How/who did you get into chess?
Mom taught me to move pieces; I learned basic checkmate patterns from playing with stockfish level 1 on Lichess.
How long have you been playing chess?
About two years.
How old were you when you started playing chess?
I started when I was 4.5 years old.
Were there breaks and when did you get serious?
I am serious all the time.
What was your first chess rating?
885 online, 1031 OTB.
What was your peak chess rating?
Regular 1424 and Blitz 1652.
|What is your favorite thing about chess?
Play and learn.
What do you like about chess tournaments?
I got to win more games at tournaments sometimes.
Where would you like to go with chess from here?
I want to become a grandmaster.
Do you have advice for others?
Who is your favorite player?
Which players would you like to meet?
Magnus Carlsen, Anatoly Karpov and Garry Kasparov. Dad told me it’s impossible to meet Bobby Fischer now.
What tournaments would you like to spectate?
National and state K-12 championship.
What tournaments would you like to participate in?
National and state K-12 championship.
Do you have a favorite historical chess game?
Bobby Fischer’s 21-move Brilliancy against Byrne.
Do you have a favorite game of your own?
I don’t have a favorite game of my own.
Have you moved around (States, Cities…)?
No. I am in the Triangle all the time.
What are your favorite online chess sites (playing and training)?
Lichess, chess.com, chesskid and chesstempo.
What is/are your favorite chess book?
100 endgames you must know.
Who is/are your favorite chess streamer?
Is/Are there chess club(s) that you’d like to visit?
I like Caribou Coffee at Cary but I don’t have much time to visit now.
To Move and Win! Presented by John Timmel
1947 USSR Championship held in Moscow. Black to move and win Efim Geller
Nürnberg, 1888 White to move and win Metger
The following is Yang Yu’s favorite historic game (see Yang’s interview above)
Robert Byrne vs. Bobby Fischer in the 1963/4 US Championship.
Black just played 21…….Qd7. Two GM commentators of the games in progress stated that Bryne was winning. To say they were surprised by Byrne’s reply would be an understatment.
What did white do here? answer at the bottom of the page
Solution to last month’s quiz
Dan Meinking / Chess Life 1982
Solution: descending and ascending staircase maneuvers
1.h7 b3! 2.h8Q b2 3.Qg8+ Ka1 4.Qg7 Ka2 5.Qf7+ Ka1 6.Qf6 Ka2 7.Qe6+ Ka1 8.Qe5 Ka2 9.Qd5+ Ka1 10.Qd4 Ka2 11.Qc4+ Ka1 12.Qxa4! Ka2 13.Qc4+ Ka1 14.Qb3 a2 15.Qc3 Kb1 16.Qxd3+ Ka1 17.Qd4 Kb1 18.Qe4+ Ka1 19.Qe5 Kb1 20.Qf5+ Ka1 21.Qf6 Kb1 22.Qg6+ Ka1 23.Qg7 Kb1 24.Qh7+ Ka1 25.Qh8! Kb1 26.Qh1# 1-0
Bryne – Fischer
If 22. Qf2 Qh3+ 23. Kg1 Re1+! 24. Rxe1 Bxd4 or 22. Ndb5 Qh3+ 23. Kg1 Bh6 wins
Copyright – 2022 North Carolina Chess Association