29 November 2019

US Chess CL Archive

This week, in US Chess Digital Archive Is Live, USchess.org announced,

US Chess has been working on digitizing its archive of Chess Life and Chess Review magazines (as well as titles such as Chess Life & Review). This project is now complete and the archive is now online and available for download by the general public.

I've referenced existing digital CL collections in at least two previous posts:-

I downloaded the files that were missing from my own collection and have started flipping through them. Unfortunately, the document scans aren't first rate -- many pages are skewed and/or clipped on the side of the magazine fold, and the text is often too light, making it difficult to read -- but who's complaining? Thanks, US Chess!

28 November 2019

Oldest Chess Piece?

How did I know that The Last Yahoo (August 2019) really was the last? In Yahoo's heyday, stories like the following would have easily been featured in Yahoo News. Now I have to rely on Google News for this sort of thing, where Google News is more-and-more a rehash of Chess.com News.

'World's oldest chess piece'? I'm sure I've seen that headline in years past.

That middle story leads to:-

2019-11-27: Chess piece found in Jordan may be world's oldest (phys.org) • 'John Oleson with the University of Victoria has reported that a small object found in Jordan made of sandstone might be the oldest chess piece ever found.'

But don't start rewriting chess history yet. The story ends with:-

Taken together, the evidence suggests that the Rook is possibly the earliest evidence of such a chess piece design, and possibly the oldest example of any type of chess piece. More work is required to verify that the stone object is, indeed, a chess piece before it can be designated as such.

I think it's a huge stretch. Huge!

25 November 2019

TCEC/CCC Bonus Events

Two weeks have passed since the previous report on the two top engine-vs-engine competitions: TCEC FRC Leagues; CCC11 R2 Underway. To summarize the situation at that time:-

TCEC: 'Cup 4' finished with Stockfish beating LCZero for 1st place (+1-0=7). After 'Cup 4', the TCEC held a series of chess960 (FRC) matches; Stockfish beat AllieStein in the Superfinal by a score of +10-0=10. • CCC: CCC11 R1 finished with Stockfish 3.5 points ahead of second place Leelenstein. R2 is underway.

What activities are the TCEC and the CCC organizing today? In a word, 'bonus' events:-

TCEC: The site is running a series of bonus tournaments; the current event is titled 'Viewer Submitted Openings Bonus 4', with eight engines. The relevant channel command informs,

!bookbonus • If you want to submit an opening to potentially be included in a future Viewer Submitted Openings Bonus (VSOB), see pinned message in #bonus-arena in !discord. Current openings at wiki.chessdom.org.

This was the first time I had seen the TCEC Wiki, so I added the it to the page of 'TCEC/CCC Links'. What's next for TCEC?

!next • 'N' bonuses where N >= 2 from {SF vs. Lc0 with S16 Sufi openings, VSOB5, ...} 30min + 5s. Then S17 testing followed by S17.

As for S17:-

!s17 • Tentatively set to December 2019. Delay due to a possible sponsored hardware upgrade.

Other commands overview S17 tentative plans: 'S17 Qualification League (unofficial)...', 'S17 L1: Not announced yet', etc.

CCC: CCC11 R2 finished with LC0 3.0 points ahead of second place Stockfish. Guest engine Lc0-dev, mentioned in the previous report, finished third, but is not allowed to advance to R3. The following chart gives a bird's-eye view of the R2 results; for a larger version, see 'TCEC/CCC Links'.

In last week's post, 'Fat Fritz' by Albert Silver, I wondered,

How long will it take the computer chess world to pit Fat Fritz against the battle-hardened competitors in the series of TCEC/CCC events?

After CCC11 R2, the CCC ran two matches:-

+16-6=80, Stockfish - Fat Fritz
+14-5=83, Lc0 - Fat Fritz

While the results aren't particularly encouraging for AI/NN engine Fat Fritz, the matches were against two of the top-ranked engines in the world, one a traditional CPU engine, the other an AI/NN GPU engine. As the chart for CCC11 R2 shows, the final ranking in a multi-engine, round-robin tournament can depend heavily on how well the top engines perform against the bottom finishers. Let's hope Fat Fritz finds its way into the qualification rounds of the next seasons for the TCEC and/or the CCC, perhaps along with the latest version of Fat Fritz's cousin, Fritz 17.

[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]

24 November 2019

Masters of the [Chess] Arts

As the long-running series on The Sociology of Chess (November 2016), starts its fourth year, I'm no closer to understanding the subject than I was three years ago. This month's episode is from the World Chess Hall of Fame (worldchesshof.org).


Masters of the Arts: STL Chess Club's Tony Rich and US Chess' Carol Meyer (8:44) • '[Published on] Oct 21, 2019'

The description expanded on the title:-

Watch as Saint Louis Chess Club Executive Director Tony Rich interviews US Chess Executive Director Carol Meyer as they reflect on the World Chess Hall of Fame exhibition, "US Chess: 80 Years -- Promoting the Royal Game in America," on view March 6 - October 27, 2019.

I couldn't remember seeing a video in what appears to be a series of its own -- Masters of the Arts -- but it turns out there are more than two dozen of them: worldchesshof / search?query=masters (youtube.com). The video shown above was the first to appear on one of my short lists for chess videos, which stretch back to 2012. It was also the first video in the series to use the word 'chess' in the video's title. That might explain why most of the videos in the worldchesshof channel rarely reach 1000 views. SEO works, but it also works against.

***

Just after posting: On top of the inadequate SEO, the video doesn't allow embeds:-

Video unavailable • Playback on other websites has been disabled by the video owner.
Talk about clueless!

18 November 2019

'Fat Fritz' by Albert Silver

Last year I wrote a short series of weekly posts about the preliminary stages of TCEC S13:-

To make a long story short, AlphaZero begat Leela which begat Deus X. After Deus X was eliminated from S13, nothing more was heard about it until a year later:-

That article describes the evolution of Deus X into Fat Fritz. Skipping ahead a few months, a claim appeared similar to one made at the introduction of AlphaZero near the end of 2017:-

Not surprisingly, that claim was a precursor to the introduction of Fat Fritz as a commercial product:-

How long will it take the computer chess world to pit Fat Fritz against the battle-hardened competitors in the series of TCEC/CCC events?

17 November 2019

Doesn't Look Like a Photo

Why did this photo have 1615 views and 218 'faves'? Is it even a photo?


Game © Flickr user oldTor under Creative Commons.

That title didn't explain anything. The description (translated from Russian) said, 'Shot with a lens Samsung NX 18-55 f/3.5-5.6 OIS, HDR', which also didn't mean anything to me, because I'm not much of a photographer. At least I knew it was a photo and that it was in 21 groups, 3 albums, and 1 gallery. Maybe there's a clue there.

The first group was 'Grunge Art Work', which also didn't mean much to me. A Google search presented the question, 'What is grunge design?', and answered,

The term "Grunge" was first coined to stand for a specific type of music, influenced by punk, rock and heavy metal. The design itself takes on the rawness of punk and rock, and incorporates real life imagery inspired by the urban and industrial scene -- it's very stylistic, less uniform, and is easily recognizable.

Easily recognizable? Google also presented the question, 'What is grunge art?'. The answer was,

Such images are used by various designers today to help them create artworks that reflect the atmosphere of heavy rock, or underground punk music. This idea of a grunge look, like the idea of a retro style, has witnessed an expansion today.

I can't say that I'm any wiser after that. The second group was 'Opt out no limits for art work!', subtitled, 'A place to put your art no limits go wild! and have fun!' Other than a mild obsession with the female form, there was no discernible connection between the artworks in the group, many of which were impressive indeed.

Many groups carried the acronym 'HDR', which also appeared in the photo's description. Google presented another question, 'What is HDR photo?', and answered,

HDR is a setting on the iPhone Camera app. The letters stand for High Dynamic Range. But what does that mean? It means that your camera will process photos slightly differently than normal in order to capture greater detail from bright and dark areas in your photo.

Back to the question about views and faves, let's just add it to the long list of life's little mysteries.

14 November 2019

The Club of 14 November

Who are all of these people?


ChessCafe.com – Since 1996

And how did their ranks swell 50% in one year?

11 November 2019

TCEC FRC Leagues; CCC11 R2 Underway

Two weeks ago, in TCEC Cup 4 Finals & CCC11 R1 Underway, I decided, 'Instead of a weekly report on the two competitions, I'll report every two weeks.' To summarize the situation at that time in the two world class engine-vs-engine competitions:-

TCEC: 'Cup 4' has progressed through the semifinal round and is playing matches to settle 1st through 4th places. Stockfish is playing LCZero for 1st & 2nd places, while Komodo is playing AllieStein for 3rd & 4th -- the A/B engines vs. the AI/NN engines. • CCC: The 'CCC11 Qualification' stage finished with Leelenstein and Stockfish 1st & 2nd. The following stage, 'CCC11 Round 1', is underway with Stockfish leading.

Note that the two subjects of last week's post, AllieStein and Leelenstein, were both playing leading roles. What's happened in the two weeks since the previous TCEC/CCC report?

TCEC: 'Cup 4' finished with Stockfish beating LCZero for 1st place (+1-0=7), and AllieStein beating Komodo for 3rd place (also +1-0=7). The following chart shows the final score of both matches.

After 'Cup 4', the TCEC held a series of chess960 (FRC) matches:-

FRC GPU Bonus
FRC CPU League 2
FRC CPU League 1 Test
FRC CPU League 2 Playoff
FRC CPU League 1
FRC Sufi [aka 'Superfinal']

Stockfish beat AllieStein in the Superfinal by a score of +10-0=10. I maintain a blog on this subject (see 'Chess960 / Fischer Random' in the right sidebar under 'My Other Resources'), so I'll reserve any further commentary for that space.

At the time of writing this post, TCEC is running an 'LCZero - AllieStein Bonus' match. What about S17? The current news says, 'Tentatively set to late November 2019. Delay due to a possible hardware upgrade.'

CCC: After the previous report, 'CCC11 R1 Underway', the event finished with Stockfish 3.5 points ahead of second place Leelenstein. The top eight engines advanced to round two which is currently underway with nine engines; Lc0-dev was added to the mix. An !Command informs, 'Lc0-dev uses the very latest T40 net that is currently available at the start of each game it plays.' R2 will run for another week or so.

[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]

10 November 2019

The Idiot Is Now a Boss

On the Saint Louis Chess Club's Youtube channel, GM Wolff explains,

Twenty-two years ago I wrote 'The Complete Idiot's Guide to Chess'. It was the last thing I did as a professional chess player. The book has been in print ever since and the publisher, which is now Penguin, approached me about 9-10 months ago. They said, 'We're rebranding the whole complete Idiot's Guide series. We're beginning with a couple of titles and we'd like for your title, the chess book, to be one to begin with. They are rebranding [the series] 'How to do X like a Boss'. It's not my choice of title -- it's their new title -- so the book 'How to Play Chess Like a Boss' is on sale in the next few days.

I spent a couple months updating the material of the book and I spent particular emphasis on the chapter 'Computers in Chess' because that obviously was somewhat out of date some years later. My guide is the guide from a complete idiot to you as I taught myself quite a lot about computers in chess and in particular artificial intelligence.'

Starting with Claude Shannon in 1949, the talk covers many aspects of computer chess.


Artificial Intelligence: The Most Exciting Thing in Chess - GM Patrick Wolff (48:54) • 'Premiered Nov 1, 2019'

The video's description says,

Grandmaster Patrick Wolff talks about chess engines and the history of computers in the game. The two-time U.S. Chess Champion discusses how chess engines calculate, including an explanation of the "brute force" method, neural networks, the differences between Stockfish and AlphaZero, and more.

For another recent video about computer chess from the Saint Louis club, see AI/NN: 'Here to Stay?' (February 2019).

05 November 2019

November 1969 & 1994 'On the Cover'

I closed last month's 'On the Cover', October 1969 'On the Cover', saying,

And so ended one of the most important chapters in American chess journalism. Chess Review had been published continually since January 1933. And so ends this series on American chess in the 1960s.

In November 1969, Chess Life and Chess Review were merged for the first issue of Chess Life & Review. Let's continue the 'On the Cover' series with CL&R from 50 years ago, together with CL from 25 years ago.


Left: 'World Junior Champion Anatoly Karpov (USSR)'
Right: 'Gata Kamsky Advances in Race for PCA and FIDE World Titles!'

Chess Life & Review (50 Years Ago)

The six-page tournament report had been carried in the October issue of Chess Review, the last issue of its publication, under the title 'Tenth Junior World Championship; Stockholm, August 10-29, 1969; Related by Dr. Peter Trifunovich'. It started,

The tenth world championship for those under twenty was contested in Stockholm. Thirty-eight participated, two from the host country, one each from others, the contender from the Dominican Republic not appearing. It is expected that more than forty will participate in 1971.

Karpov finished 1st, with A.Adorjan & A.Urzica 2nd-3rd. The American representative, Kenneth Rogoff, failed to qualify for the championship section, finishing 3rd in his preliminary section behind Ulf Andersson of Sweden and Aurel Urzica of Romania.

Chess Life (25 Years Ago)

The photo shows Gata Kamsky on the right and his father Rustam Kamsky on the left. CL's 'On the Cover' introduction said,

Once upon a time, when one mentioned the "Chess Genius from Brooklyn," there could be but one interpretation. Today, we must preface such remarks carefully.

As the Professional Chess Association (PCA) and FIDE, continue to bicker, or not, one person may unify both World Championship claims, by winning both titles. Gata Kamsky will play former PCA Challenger Nigel Short in Linares, and then meet Valery Salov in January, in the next round of FIDE Challenger matches.

Just as the "good cop, bad cop" routine is undoubtedly disorienting to a suspect, "good press, bad press" concerning the Kamskys, must be equally confusing to the armchair aficionado. Lev Alburt and Larry Parr attempt to clear away some of the confusion, in our feature story.

The feature story was under the header 'The Human Side of Chess' and was titled 'Making it in America; A Conversation with Gata Kamsky'. It started,

At age 20, Grandmaster Gata Kamsky has become the first serious American contender for the world chess crown since Bobby Fischer a quarter century ago. He has just defeated India's Viswanathan Anand, 6-4, on the latter's home court in a place called Singhi Natar [sic; 'Sanghi Nagar'], and will probably be rated number three in the world in the next PCA rankings. He would also enjoy a similar distinction on the FIDE scale if that organization rated such PCA events as the young American's recent, devastating crush of Vladimir Kramnik in New York.

I expect we'll be seeing the evolving story in future editions of 'On the Cover'. For a sneak preview, see my page World Chess Championship: FIDE/PCA Chronology.

04 November 2019

AllieStein and Leelenstein

In last week's post about two world class engine-vs-engine competitions, TCEC Cup 4 Finals & CCC11 R1 Underway, I decided,

Instead of a weekly report on the two competitions, I'll report every two weeks. That rhythm is sufficient to follow the tournaments without having to spend too much time on them. It also allows for an occasional post on the off-week, to go a little deeper into the underlying issues.

In 2017, we saw the AI/NN engine AlphaZero burst into the top rank of chess engines. In 2018, it was joined by Leela. In 2019, as my post TCEC/CCC 2019 Q1-Q3 Summary from a couple of weeks ago revealed, they were joined by AllieStein [playing in the TCEC] and Leelenstein [in the CCC]. It's instructive to chronicle the progress of those last two engines in 2019. Let's start with AllieStein's progress in the TCEC:-

  • 2019-03-18: TCEC S15 Div4 Finishes; Stockfish Wins CCC6 • 'TCEC S15 division 3 is currently underway. The two top placed engines in divisions 4a and 4b all met in a playoff, from which the two top engines qualified into division 3. The AI/NN engine AllieStein (Allie + Stein) was one of the two qualifying.'
  • 2019-03-25: TCEC S15 Div3 Finishes; CCC7 Starts • 'S15 division 2 is currently underway. The AI/NN engine AllieStein was one of two qualifying from division 3.'
  • 2019-04-01: TCEC S15 Div2 Finishes; CCC7 Nears Final • 'S15 division 1 is currently underway; the AI/NN engine AllieStein was one of two qualifying from division 2.'
  • 2019-04-08: TCEC S15 Div1 Finishes; CCC7 Final Starts • 'KomodoMCTS and AllieStein finished 1st and 2nd in division 1 to qualify into the premier division with six other engines.'
  • 2019-04-29: TCEC S15 Final: SF vs. Leela; CCC8 S1 at 3/4 Mark • 'Among the also-rans, the result of AI/NN engine AllieStein was the most remarkable. Starting the season in division 4, the lowest of the five divisions, it consistently finished in the top-2 for each successive division, eventually qualifying into the premier division. There it tied for 4th/5th with Houdini. [...] AlphaZero was the first AI/NN engine to compete at the level of the traditional engines; Leela was the second; and AllieStein is the third.'
  • 2019-09-13: Beyond AlphaZero • 'ScorpioNN and Stoofvlees first survived the [S16] 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.'
  • 2019-09-16: TCEC S16 DivP & CCC10 S3 Both Underway • 'The [S16] 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.'
  • 2019-09-30: TCEC S16 Final Underway; Leelenstein Wins CCC10 • 'The Premier Division finished with Stockfish, AllieStein, and LCZero at 1st/2nd/3rd, each separated by a half-point. [...] After 22 games in the 100-game final match, Stockfish has a two point lead over AllieStein.'
  • 2019-10-14: Stockfish Wins TCEC S16; CCC11 Underway • 'Stockfish beat AllieStein by a final score of 54.5-45.5, more specifically +14-5=81.'

The TCEC provided the first mention of Leelenstein; after that it was all CCC:-

  • 2019-03-04: TCEC S14/S15 Interlude; CCC6 S2 Wrapping Up • 'After the finish of [TCEC] S14, while S15 preparation was underway, the site conducted three short competitions. The first event was a 48-game three-way match that lasted a day or so. It finished with the result Stockfish 17.5/32, Leelenstein 16, Antifish 14.5.'
  • 2019-03-25: TCEC S15 Div3 Finishes; CCC7 Starts • 'The same [Chess.com] report outlined the plan for CCC7. [...] That makes four AI/NN engines: Lc0, Antifish, Leelenstein and, Allie.'
  • 2019-04-08: TCEC S15 Div1 Finishes; CCC7 Final Starts • 'The preliminary event 'Blitz Bonanza' finished with a surprise result. Although Stockfish finished first, it was ahead of three AI/NN engines -- Leelenstein, Leela, and AntiFish -- in that order.'
  • 2019-05-06: TCEC Cup 3 Nears QF; CCC8 S1 90% Finished • 'Leelenstein and Antifish will be fighting each other for one of the last two places in the final.'
  • 2019-05-13: Leela Wins TCEC Cup 3; CCC8 S1 99% Finished • 'Since the rules specified, 'Top 4 engines advance to final, only two can be from a single, or shared, code-base', Leelenstein beat Antifish for the second qualifying spot in the Leela family.'
  • 2019-05-20: TCEC S15 and CCC8 Finals Both Underway • 'In the final stage, with about two-thirds of the games already played, the order of the four engines is Stockfish, Leela, Leelenstein, Houdini. The point spread between each successive place is almost enough to guarantee that this will be the final standing.'
  • 2019-07-15: TCEC S16 Starts; CCC9 Finishes • 'CCC9 Gauntlet Final (200 game match); 1) 103.5 Stockfish; 2) 96.5 Leelenstein'
  • 2019-09-16: TCEC S16 DivP & CCC10 S3 Both Underway • 'Six engines advanced from the quarterfinals, with Stockfish, Lc0, and Leelenstein finishing well ahead of the others. [...] The 'CCC10 Semifinals' stage is currently underway with the same three engines again well ahead of the other three.'
  • 2019-09-23: TCEC S16 DivP & CCC10 Final Nearly Complete • 'Leelenstein and Stockfish finished tied for 1st/2nd in the 'CCC10 Semifinals', thereby advancing to the final match. [...] Of the 200 games scheduled for the 'CCC10 Finals', 190 games have already been played. Leelenstein has a narrow but convincing lead over Stockfish.'
  • 2019-09-30: TCEC S16 Final Underway; Leelenstein Wins CCC10 • 'Leelenstein beat Stockfish by a score of +19-15=166 in the CCC10 final.'
  • 2019-10-28: TCEC Cup 4 Finals & CCC11 R1 Underway • '[The following chart] captures a portion of the final crosstable for the 'CCC11 Qualification' stage. Only the full results of the first six engines, Leelenstein 1st, are shown.'

Both AllieStein & Leelenstein entered the competitions at the same time and both reached the top ranks at about the same time. Does that tell us something about AI/NN engine development?

03 November 2019

Vasarely Set, Board, or Both?

We don't see many chess sets here on Top eBay Chess Items by Price (March 2010). The previous set appears to have been Chess with Rare Woods (February 2017), and that wasn't just a set, it was a table, which I introduced with 'We don't see many chess tables here on...'

I never have a problem if I want to feature a chess set. Every time I start to compose an eBay post, there are dozens of expensive sets to choose from. That's why I don't often choose one: they would inundate everything else. The set shown below was titled 'Vasarely, Victor (1906-1997). Chess Color Serigraph in Plexiglas Hand Signed', and sold for US $1520 after a single bid.

The description said,

Medium: Serigraph & Silkscreen • Originality: Limited Edition • Features: Framed, Signed • Style: Abstract / optical • Unused / complete, with custom clear plexiglass dust cover.

The item came with a 'CERTIFICAT' that said,

We hereby certify that THE VASARELY CHESS SET is issued in an edition limited world-wide to only 1,500 (one thousand five hundred) copies. Each VASARELY CHESS SET is an individually numbered original work, with the chessboard personally signed by Victor Vasarely in his own hand.

The photo of the item also had four other photos showing how the board and set were designed. It looks like some sort of manual raytracing executed at a time when computers and software were either big budget items or even non-existent.

Why did I decide to feature this set for this post? Vasarely produced other chess related works. Cataloging them would make a good rainy day project.

01 November 2019

Fide.com Archive

Last month, in a post on my World Championship blog about the Status of the Women's World Championship, I wrote,

The recent design change of FIDE's web site has rendered many (all?) bookmarks unusable ('404 Page Not Found; The page you were looking for doesn't exist'), making any research problematic.

I just noticed that new design includes a link called 'Archive' that leads to:-

old.fide.com

It seems slow, but it works. Its most recent news article is dated '19 September 2019'. To acess bookmarked articles, I have to change the subdomain for a link from www.fide.com to old.fide.com. I don't know why FIDE can't do this automatically with a '404' error handler, but I'm thankful that it works at all. For the record, a corresponding Archive.org link is:-

archive.org/web/*/www.fide.com • currently: 'www.fide.com : Saved 2,428 times between November 4, 1996 and November 1, 2019'

I'm not the only blogger to have an issue with FIDE's redesign. See also Today’s Pet Peeve (spraggettonchess.com; 3 October 2019).