TCEC S16 Final Underway; Leelenstein Wins CCC10

Last week's report on two world class engine-vs-engine competitions, TCEC S16 DivP & CCC10 Final Nearly Complete, saw both nearing the end of an advanced stage. To summarize that report:-

TCEC: The Premier Division will finish later today. Stockfish is almost certain to qualify for the final match. The other two engines are neck-and-neck. • CCC: Leelenstein and Stockfish finished tied for 1st/2nd in the 'CCC10 Semifinals', thereby advancing to the final match.

Both competitions finished their respective stages within a day after the report. What happened then?

TCEC: The Premier Division finished with Stockfish, AllieStein, and LCZero at 1st/2nd/3rd, each separated by a half-point. The following chart shows the top four of the eight participating engines.

After 22 games in the 100-game final match, Stockfish has a two point lead over AllieStein. The match will last another week and a half.

CCC: Leelenstein beat Stockfish by a score of +19-15=166 in the CCC10 final. No plans for CCC11 have been announced.

I've been tracking the two competitions weekly since Results: TCEC S14-P / CCC3 S3 (January 2019). The time has come to close the series and move on to other topics. I'll summarize the previous nine months in the next post.

[For further information from the various stakeholders in the engine-to-engine events, see the tab 'TCEC/CCC Links' at the top of this page. • NB: Leela = LC0 = LCzero]

TCEC S16 DivP & CCC10 Final Nearly Complete

Last week's post, TCEC S16 DivP & CCC10 S3 Both Underway, summarized, 'Both the TCEC and the CCC have reached the stage that will determine the participants in the finals.' More specifically:-

TCEC: The Premier Division has another week to run with three engines -- AllieStein, LCZero, and Stockfish -- within a half point of each other. • CCC: The 'CCC10 Semifinals' stage is underway with three engines -- Stockfish, Lc0, and Leelenstein -- well ahead of the other three.

The current status of the two engine-vs-engine competitions is summarized below.

TCEC: The Premier Division will finish later today. Stockfish is almost certain to qualify for the final match. The other two engines are neck-and-neck.

CCC: Leelenstein and Stockfish finished tied for 1st/2nd in the 'CCC10 Semifinals', thereby advancing to the final match. The following chart shows that Lc0 was well behind the two finalists.

Chess.com issued a report, CCC10 Finals: Leelenstein vs Stockfish. Of the 200 games scheduled for the 'CCC10 Finals', 190 games have already been played. Leelenstein has a narrow but convincing lead over Stockfish.

[For further information from the various stakeholders in the engine-to-engine events, see the tab 'TCEC/CCC Links' at the top of this page. • NB: Leela = LC0 = LCzero]

2019 US Chess Delegates Meeting

Look like it wasn't really The Last Sociology Post (*; August 2019). How could it have been when there is so much compelling material to be reviewed (the following video had 34 viewers in its first month!):-

US Chess 2019 Delegates Meeting 1.1 (2:43:37) • 'Published on Aug 22, 2019'

The description added only that the meeting was held on a Saturday:-

Part 1 of the Saturday Delegates Meeting for US Chess

For more of the same from Youtube's USChessFederation channel, see:-

• US Chess 2019 Delegates Meeting 1.2
• US Chess 2019 Delegates Meeting Part 2
• 2019 US Chess Awards Ceremony

If that doesn't satisfy your appetite for the ins-and-outs of chess administration, see last year's Delegates Meeting and US Chess Annual Awards [Full Video] (August 2018)

(*) The most recent post in this blog's series on The Sociology of Chess (November 2016)

The End of Daily Blogging

I ended last week's post, Breaking the 3000 Barrier, with a short statement of direction:-

This current post is no.2999 on this blog, making tomorrow's post [Beyond AlphaZero] no.3000. It's a goal I've been working toward since the beginning of the year, because it will be the last regular post on the blog. Some of the ongoing series need to be closed separately, but I'll look at those some other time.

I tackle those ongoing series on a weekly cycle:-

Sunday: I've been rotating monthly through eBay, Youtube, Flickr, and sociology. These are all subjects that expand on some random, chess-related topic, where I always learn something new. I'll continue with some of these until I finally close the blog.

Monday: Since the beginning of the year, I've been following the TCEC and CCC engine tournaments. I'll continue with these for the near future, but might decrease the frequency of posts.

Tuesday/Thursday: These are free days, unrelated to any series, so I'll stop. This current post is the last.

Wednesday: I'll continue with my World Chess Championship Blog at least through the end of the current cycle. Who will challenge Magnus next year?

Friday: I use Friday posts to develop topics that can't be covered in 1-2 posts. The most recent post in the current series was 'Beyond AlphaZero', which is a good stopping point.

Saturday: Two posts a month have been for Chess960 (FRC). I'll continue at least through the climax of the FWFRCC (FIDE World Fischer Random Chess Championship) in November.

Chess is a wonderful subject and there are much worse pastimes. Whether you've been a regular or an occasional visitor to this blog, I offer a big 'Thank You!' and I hope you'll continue to stop by from time to time.

TCEC S16 DivP & CCC10 S3 Both Underway

To summarize last week's post TCEC S16 DivP & CCC10 S2 Both Underway:-

TCEC: The 'Premier Division', to determine which two engines will compete in the Superfinals, is underway with eight engines. • CCC: The 'CCC10 Quarterfinals' are underway with 11 engines.

In the world's strongest engine-vs-engine competitions, what's changed since then?

TCEC: The Premier Division has another week to run. Three engines -- AllieStein, LCZero, and Stockfish -- are currently within a half point of each other, two points ahead of the rest of the field.

CCC: As Chess.com reported in CCC10 Quarterfinals Finished; CCC10 Semis Start Now!, six engines advanced from the quarterfinals. The following chart shows a partial crosstable for those six, with Stockfish, Lc0, and Leelenstein finishing well ahead of the others. None of the other engines managed to win against the top three.

The 'CCC10 Semifinals' stage is currently underway with the same three engines again well ahead of the other three. When I gathered the basic info for this post, the schedule said the event would finish in a couple of days, but later an announcement appeared that said, 'Due to unexpected hardware issues, CCC10 Semifinals will resume [later]'.

Both events have reached the stage that will determine the participants in the finals. If the tournaments progress as the intermediate results show, we will see two AI/NN engines battling against each other -or- Stockfish against one of those engines.

[For further information from the various stakeholders in the engine-to-engine events, see the tab 'TCEC/CCC Links' at the top of this page. • NB: Leela = LC0 = LCzero]

Four Player Chess Pawns

Whether these are chess Pawns or Parcheesi pieces, their messages are identical.

Photo top middle: Mentoring © Flickr user Marco Verch under Creative Commons.

'Mentoring' and 'Support' look the same, as do 'Diversity' and 'Friends'. Is there another message here? It looks like The Last Flickr Photo (August 2019) wasn't.

Beyond AlphaZero

For the last two months, I've used this blog's Friday post to wander around topics relevant to AlphaZero:-

• 2019-07-19: Interpreting AlphaZero • 'What can we learn from black box models? '
• 2019-07-26: AlphaZero's Zeros • Engine evaluations of '0.00'
• 2019-08-02: Winning Percentage to Centipawns
• 2019-08-09: 'Game Changer' PGN
• 2019-08-16: The Value of Deep Learning
• 2019-08-23: A Chess Epiphany • 'Would Demis Hassabis have been more useful to society as a professional chess player?'
• 2019-08-30: Playing Without a Horizon
• 2019-09-06: 'Game Changer' PGN Explored

As everyone knows by now, although AlphaZero never played in a neutral, supervised competition, it was quickly copied by Leela Chess Zero (aka Lc0, LCZero), which improved rapidly to become one of the strongest chess engines in the world. Leela has since been joined by other AI/NN engines, which are also making a mark in the engine-vs-engine competitions that I've been following since the beginning of the year. Not too long ago, in TCEC S16 Qualification; CCC9 Bonuses (July 2019), I reported,

The first stage of S16, the Qualification League, is well underway. Two neural network (NN) engine newcomers, ScorpioNN and Stoofvlees, are currently leading, while a third NN, ChessFighterNN, is fighting to stay out of last place. A Leela clone running on CPUs instead of GPUS [LCZeroCPU], is in third place. TCEC chat commands point to resources describing the three NNs [see links].

ScorpioNN and Stoofvlees first survived the qualification stage, then the two league stages (finishing ahead of another NN engine, FireNNb, along the way), and are currently playing in the TCEC Premier Division. There they were joined by two other NN engines, LCZero and AllieStein, along with four non-NN engines: Stockfish, Houdini, Komodo, and KomodoMCTS.

As for the other engine-vs-engine event, the CCC, four NN engines competed in the qualification stage. I reported last week in TCEC S16 L1 Finished; CCC10 S1 in Trouble, that the tournament did not go smoothly. In spite of the difficulties, all four NN engines -- Lc0, Stoofvlees, DarkQueen, and Leelenstein -- eventually qualified to the next stage.

The upshot of this activity is that the NN engines and the non-NN engines are roughly equal in strength. Will this continue or will one of the two technologies soon prove dominant?

Breaking the 3000 Barrier

Remember this post: Breaking the 2900 Barrier (April 2019), marking both post no.2900 on this blog and a historical point in Elo chess ratings? I wrote,

For 2900 ratings, we have to turn to the world of chess engines. Here is a relevant chart signed Luke Muehlhauser and dated 2011. I've marked the 2900 crossover point with a red star. The engine under the star is Shredder 8.0 and the year was 2003 or thereabouts.

As for 'Breaking the 3000 Barrier', the same chart points to Rybka 1.2 a few years later. Looks like I won't have to write the 3000 post at all.

This current post is no.2999 on this blog, making tomorrow's post no.3000. It's a goal I've been working toward since the beginning of the year, because it will be the last regular post on the blog. Some of the ongoing series need to be closed separately, but I'll look at those some other time.

'Smart Is the New Sexy'

I don't remember where I saw the news first. Maybe it was Chessable joins the Play Magnus chess24 family (chess24.com):-

We're excited to be able to announce that Chessable is now part of the Play Magnus and chess24 group. Play Magnus will invest heavily in the fast-growing chess improvement site, and although nothing will change at first on chess24 or Chessable we’re looking forward to a great opportunity to combine our forces to provide more and better resources for chess players of all levels.

Maybe it was Chessable joins Play Magnus and chess24 – yes, it’s true! (chessable.com/blog; by David Kramaley, 'Chessable's CEO and Chief Scientist'):-

Dear Chessable members, today is simply unbelievable! We are delighted to announce that Chessable has become part of the Play Magnus & chess24 family. That’s right, The World Champion himself, Magnus Carlsen, is now part of our Chessable family! It’s been an exciting and incredible journey, and we wanted to thank you for being part of it.

Both announcements were dated a week ago, 3 September. My first reaction was, 'What is the "Play Magnus Chess24 family'"? Then I remembered an announcement from a few months ago, Chess24 and Play Magnus join forces (chess24.com; March 2019):-

Today we’re proud to announce that chess24 has merged with Play Magnus, the Norwegian company behind a family of apps building on the brand of World Chess Champion Magnus Carlsen. Play Magnus and chess24 will continue to operate independently online, but with our joint resources and know-how we believe we’re better equipped than ever to provide the very best possible chess experience.

One of the comments noted,

WildBillHickok: Chess24 owners [Enrique] Guzman and [Jan] Gustafsson will own 38% of the newly merged company according to Norwegian business paper Dagens Naeringsliv.

By one of those coincidences that often seems to be more than a coincidence, Chess24 and 'Play Magnus' both started operating within a few months of each other; see Play Magnus! (chess24.com; August 2014):-

Canadian Kate Murphy is the CEO of Play Magnus, the company behind the chess app of the same name. Her ambition is to popularize cultural and educational activities for youngsters, and Magnus Carlsen seemed like the perfect man to do this with chess. (Interview by GM Jonathan Tisdall)

For some reason the Wikipedia pages for Play Magnus and Kate A. Murphy do not link to each other. For more from the comely Canadian, see Youtube channel 'TEDx Talks', Chess and the Game of Life | Kate Murphy | TEDx UWCRCN (December 2016), 'Smart is the new sexy.'

TCEC S16 DivP & CCC10 S2 Both Underway

A week ago, in TCEC S16 L1 Finished; CCC10 S1 in Trouble, I reported,

TCEC: TCEC admins are currently testing a new server in preparation for Premier Division, which will run for about three weeks. • CCC: In the 'CCC10 Qualification', all non-AI/NN engines have finished playing against each other, while the four AI/NN engines are currently completing their schedules.

Both engine-to-engine events have moved to the next stage.

TCEC: The Premier Division (DivP), to determine which two engines will compete in the Superfinal (SuFi), is underway with eight engines. The URL for the live broadcast changed to tcec-chess.com, and I updated the tab 'TCEC/CCC Links' to record the new address.

CCC: The 'CCC10 Quarterfinals' are underway with 11 engines instead of the 10 announced; I couldn't find an explanation for the change. The top six engines will advance to the semifinal.

The previous event 'CCC10 Qualification' is not in the archive, although a crosstable and a PGN file are available. Stockfish won, followed by Leelenstein and LC0, which were well ahead of the rest of the field. The top seven engines were joined in the quarterfinals by four others: Fire, Houdini, Komodo, and Komodo MC.

[For further information from the various stakeholders in the engine-to-engine events, see the tab 'TCEC/CCC Links' at the top of this page. • NB: Leela = LC0 = LCzero]

A Special Chess Board

It's been a few years since I last featured a DGT (Digital Game Technology) video: DGT at the Olympiad (August 2014). Here's a new one from Denmark.

What's Special about a Professional Chess Board? (3:07) • 'Published on Aug 13, 2019'

The video description didn't have much to add. In response to a comment, 'Nice commercial', the video's creator wrote,

Per Danish law we cannot make a commercial without stating it clearly in the video and description. This is not a commercial. We have not been paid for this video and neither the DGT company nor Chesshouse [NB: the distributor supplying the board] has had ANY editorial rights. Not that we would have any problems with making commercials mind you -- but any such paid content would be clearly marked both in the video and the description.

Danish law or not, that's a good guideline for anyone producing online content.

'Game Changer' PGN Explored

Last month, in 'Game Changer' PGN, I collected PGN files relevant to Sadler & Regan's book 'Game Changer', and noted,

As a bonus, the 'Annotator' tags in the NIC file match the 'Round' tags in the AZ_vs_SF files. This provides a cross-reference between the two sets of files. I'll use these files for further explorations.

The 'NIC file' is a PGN file covering the games in the book. I used the NIC file to prepare last week's post, Playing Without a Horizon. The 'AZ_vs_SF files' are two PGN files of games between AlphaZero and Stockfish provided by DeepMind to Sadler & Regan. Those games were behind the authors' motivation to write the book.

Having a little time on my hands, I used it to dig deeper into the three files. The 'NIC file' contains 106 games in the same sequence as the games presented in the book. Of these, 69 games are AZ_vs_SF, while the other 37 games are from various sources. Many of the other games were played between world class players and were used by Sadler & Regan to illustrate recurring themes from the AZ_vs_SF games.

Of the 69 AZ_vs_SF games, 48 have AlphaZero playing White. One peculiarity of the 'NIC file' is that only 51 of the 69 games can be cross-referenced to the AZ_vs_SF files. That means 18 games are from other sources. I looked at the first two of those games (no.12 and no.17 in the NIC file) and confirmed that they are indeed missing from the AZ_vs_SF files, although they are discussed in the book. Were there other game files examined in the preparation of the book?

Of the 51 games from the AZ_vs_SF files, 32 are unique, meaning that many of the games were used more than once in the book. One game appears four times. By some sort of pleasant coincidence, this is the same game I discussed in last week's 'Horizon' post.

Despite the massive effort that went into Sadler & Regan's book, there are still plenty of the original AZ_vs_SF games to explore. What other chess secrets might those games reveal?

Pal Benko 'On the Cover'

Pal Benko has been on my mind lately -- and I imagine he's been on many chess players' minds. While I was preparing the previous post, September 1969 'On the Cover', I decided to find Benko's earliest appearance on the covers of the two main American chess magazines.

The earliest Chess Life (CL) cover is shown below on the left. The year 1961 was the first with CL published in magazine format, switching from the newspaper format of previous years. By coincidence, Benko appeared on the cover of Chess Review for the same month, October 1961.

Left: 'Pal Benko - U.S. Open Champion'
Right: 'Open Champion'

Chess Life

Benko Wins U.S. Open Championship; Highest Score Ever Attained At Largest Open • Hungarian born International Grandmaster Pal Benko, of New York City, topped a record field of 198 players at the U.S. Open held in San Francisco August 14th to 26th with a record-breaking score of 11-1. By drawing only two games and winning 10, Benko's score was the highest ever achieved in the sixty-three consecutive years that the event has been conducted.

The tournament conducted in the picturesque city of San Francisco was the largest attended Open -- Milwaukee, 1953 -- with 183 players holding that first place spot for the last eight years.

Although the article mentions '63 consecutive years' of the U.S. Open, a table at the bottom of the page ('Former U.S.Open Champions') starts with '1941 Reuben Fine'. For the complete list, see Wikipedia's U.S. Open Chess Championship; the 1961 U.S. Open is given as no.62.

Chess Review

Hungarian Day at USCF Open • Two Hungarian-born players finished first and second respectively in the United States Open Championship at San Francisco. Pal Benko, international grandmaster residing in New York, went undefeated through twelve Swiss rounds to gain the crown and a $1,000 first prize with a tally of 11-1. Runner-up was his compatriot, California champion Zoltan Kovacs, 10-2, who lost only to Benko.

The earliest photo of Benko in Chess Life was the 20 January 1958 issue (CL was published twice a month at that time). It showed Benko with the 'Players and Officials at Dallas International 1957'. The players included Reshevsky, Evans, Gligoric, Larsen, and Najdorf, plus other international luminaries. Since Benko was neither a player nor an official in the tournament, what was he doing in the photo? The last CL issue of 1957 explained,

By-products of this assemblage of masters were two challenges: an exhibition match between Dallas Master Kenneth R. Smith and the recently defected Hungarian International Master Paul Benko; and a blindfold match between International Masters Miguel Najdorf and George Koltanowski.

As for Chess Review, we find Benko on the cover for March 1959 (composite photo, centered on Fischer, of U.S. Championship); and August 1959 ('Twice Champion', New York and Western Opens). He also appeared numerous times on interior pages.

September 1969 'On the Cover'

Fifty years ago, the two main American chess magazines were only two months from being merged into a single publication. Their covers featured American chess players from two different centuries.

Left: 'Marshall Rohland; USCF Secretary 1956-66, President 1966-69'
Right: 'A Morphy Sidelight'

Chess Life

Presidential Report by Marshall Rohland • At the end of one's Presidency, it is appropriate and constructive to look back over the progress in USCF during the three year term, and also the preceding years. Looking back in perspective reveals a long story of struggle and accomplishment.

For a similar report from Rohland's successor (in CL November 1972), see The USCF in Numbers (June 2014).

Chess Review

A Morphy Sidelight by Manfred Zitzman • Paul Morphy! The name remains a magic one for countless chess players past, present and yet to be born. As someone has said, it is because we remember other chess greats not in the plenitude of their powers, but as dethroned champions -- bitter, disillusioned men who, more often than not, avoided their most dangerous challengers for years before they finally played and succumbed.

Alekhine, it is true, died with the World Championship still in his possession; but his twenty years of dodging a re-match with Capablanca will forever cloud his reputation. In contrast, Paul Morphy's light blazes undiminished across more than a century of time! Here there was no dodging or evasion. Morphy actively sought out the strongest challengers he could find and blasted them off the board.

The CR cover shows the October 1859 cover of 'The Chess Monthly', published almost 110 years before the CR issue. We saw the earlier magazine on this blog a few months in Five Volumes of 'Chess Monthly' (May 2019), and looked at Morphy's role in Paul Morphy, Editor (ditto). The author of the three page Morphy article also featured on the CR side of December 1965 'On the Cover' (December 2015).

TCEC S16 L1 Finished; CCC10 S1 in Trouble

Another Monday, another report on the two ongoing engine-vs-engine competitions. Last week, in TCEC S16 L1 Wrapping Up; CCC10 S1 Underway, I noted,

TCEC: 'League 1' is almost finished. • CCC: The first stage, 'CCC10 Qualification', is already more than one-third completed and should finish this week.

Since 'almost finished' and 'should finish' are only forecasts, what's the current status of the two events?

TCEC: After 'League 1' finished, the top four engines participated in a playoff to determine which two engines would advance to the Premier Division (aka 'DivP'). The two AI/NN engines, Stoofvlees and ScorpioNN, finished 1st/2nd in the playoff. The results of both events are shown in the following chart.

TCEC admins are currently testing a new server:-

TCEC webserver has proven to not have been stable always. Therefore we are currently configuring a completely new webserver, and will continue with DivP when ready, which should be within a day (at most). This new webserver will eventually completely replace the old one.

The Premier Division will run for about three weeks.

CCC: Last week's forecast -- 'CCC10 Qualification should finish this week' -- ran into some kind of a brick wall. All non-AI/NN engines have finished playing against each other, while the four AI/NN engines are currently completing their schedules. During the past week, the CCC ran a 'Sanity Test GPU Server' with those same four engines. I couldn't find a discussion of the problem.

[For further information from the various stakeholders in the engine-to-engine events, see the tab 'TCEC/CCC Links' at the top of this page. • NB: Leela = LC0 = LCzero]

1974 Spassky - Byrne Poster

We don't see many posters in this blog's series for Top eBay Chess Items by Price (March 2010), and I'm not sure why. Is it because there aren't many chess posters that command three or four digit prices? Or because there aren't many chess posters in existence? Or because of some other reason.

The item shown below was titled 'Genuine 1974 Vintage Chess Poster Boris Spassky vs. Robert Byrne in Puerto Rico'. It sold for US $599.95 'Buy It Now'.

The description added,

Here we offer a scarce (perhaps singular) relic from the apogee of the 1970's worldwide chess craze: an original silk-screened poster promoting the January 1974 Cold War chess match between USSR champion Boris Spassky and American grandmaster Robert Byrne, held at the College of Engineers, Architects, and Surveyors, San Juan, Puerto Rico.

The poster is signed in the plate, "AM". I purchased this poster 35 years ago from the front window of the renowned Village Chess Shop in Greenwich Village, NYC. Though it's got some minor tears, stains, pin holes, and creases, it is printed on very thick stock and looks absolutely wonderful. The design is remarkable and worthy of a centerpiece location! Measures 18.25 x 25.75 inches.

The match is documented on my World Championship page 1973-75 Candidates Matches as:-

Quarterfinal match; San Juan, Puerto Rico; I, 1974; Spassky - Byrne; +3-0=3 (4.5-1.5)

For another eBay poster from the same era, see 1972 Fischer - Spassky Poster (March 2013). By coincidence, an alternative item for this current post was also from the 1972 Fischer - Spassky match. Its decription said,

Exceptional memorabilia from the closing ceremony of The World Chess Championship in Reykjavik, Iceland 1972 where Bobby Fischer and Boris Spassky played the match of the century in Reykjavik, Iceland 1972. An item like this only comes on the market every 10-15 years.

A signed conclusional banquet menu from the closing ceremony of the World Chess Championship 1972. With official stamp and cancel of the match, dated September 3rd 1972. Signatures of the newly crowned world champion Bobby Fischer and the newly defeated champion Boris Spassky and the chief arbiter Lothar Schmid. Size approx. 24 x 10,5 cm (folded).

The menu sold for US $1525 after five bids from two bidders.