BOTW: 2017.11.21 Speed Chess Championship
Nov. 21, 2017. (10am Pacific, 1pm EST)

Magnus Carlsen plays Alexander Grischuk in the first semi-final of the Speed Chess Championship. Magnus is the defending Champion and huge favourite, but Alex is a former World Blitz Champion and a notorious time-trouble addict, which has given him more than the usual preparation for a blitz match.

The winner goes to the final against the winner of the match between Hikaru Nakamura and Sergey Karjakin.

The format for the matches is 3 hours of online play, broken into four formats:

  1. 90 minutes of 5+2 blitz, 
  2. 60 minutes of 3+2 blitz,
  3. 30 minutes of 1+1 bullet,
  4. one chess960 game in each time control. 

Live Games (IM Danny Rensch and GM Robert Hess)

More Magnus

While in Germany last week for the Play Magnus Live Challenge, the World Champion took part in two additional PR events.

The serious one was a live interview with Die Zeit journalist Ulrich Stock on artificial intelligence:

 The stupid one was to play a game against a self-professed super-learner who claimed he had set himself the challenge to learn one challenging skill per month, and his last challenge was learning chess in one month well enough to beat the World Champion.

The fact that the event happened -- Magnus played him -- and that it got some serious-faced coverage (see below) is so ludicrous as to be fascinating. A few thoughts:

  • Imagine a novice claimed to have spent a month learning tennis and challeneged Roger Federer to a set. Any interest in seeing that?  From his first awkward serve it would be obvious to even the tennis-ignorant that he had no chance. Is it just because chess skills are harder to see than physical skills that someone might imagine chess could be mastered in a month? What do you suppose people who believe this think about other types of intellectual skills -- e.g. in business or politics -- do they think those can be mastered in a month?
  • Why set the bar so rediculously high: wouldn't a draw against an IM be impressive enough?
  • Imagine a club-level chess player who thought this was possible: what would that say about their opinion of lowly FMs and IMs who have spent years studying chess but still have no chance against Carlsen? 
  • The self-promoting super-learner claimed he was developing  "an algorithm" that would enable him to beat Carlsen... but hadn't finished it at the time of the match. But such algorithms already exist -- you can download the source-code to one of them here: -- but just like the very first chess program ever made (by Alan Turing) it runs way too slowly on wetware. So this fellow wasn't just fantasizing about besting Carlsen in a month, but also about his ability to create a world-class chess-playing program -- which would be so simple that it could be run by a human brain -- also in only one month.

At the risk of underestimating the gung-ho super-powered self-improvement of the Bay Area entrepreneur jump-off-a-cliff-and-build-your-airplane-on-the-way-down crowd, I will ask the rude question: wouldn't such accomplishments take at least a month-and-a-half?

Palma de Mallorca, Grand Prix

November 16-25, 2017.

An 18 player, 9-round Swiss featuring some of the world's best players and which will determine the last two qualifiers for the Candidate's Final. 

After 5 rounds, Lev Aronian leads with 3.5. MVL (3/5) and Radjabov (2.5/5) still have chances to qualify.

The event is sponsored by Agon, which is restricting live video to its pay $ite

Post game interviews can be found for free on its facebook page:

TCEC Season 10 Superfinal: Houdini vs Komodo

November 20 - ??, 2017.

The TCEC Superfinal started late Sunday night/Monday morning: 100 games between Houdini 6.03 and a development build of Komodo on  an exceptionally powerful 44 core Xeon server. After 5 games, Houdini leads Komodo +1 =4 -0.

Houdini and Komodo tied atop the eight engine quad-round robin with 18.5/28. Last year's champion Stockfish was 1/2 a point behind. Interestingly, Stockfish was the only undefeated engine, and beat Komodo head-to-head +1 =3, but gave too many draws to the lower finishers. Perhaps a lesson about raising the "contempt" value during the round robin!?


The only computer chess event that matters is the Thoresen Chess Engine Competion. Engines play on identical hardware and use identical opening books, making it the best test of playing strength. It's also the event that the engine developers take seriously, with the top ones submitting development builds of their best programs, including Stockfish, Komodo, Houdini, and Fire.

This year's big innovation is the upgraded hardware: an exceptionally powerful 44 core Xeon server, more than double the power last year's host.

Like it or not, this tournament will have a strong claim to featuring the best chess ever played.

TCEC Season 10 consists of three stages:  

  1. Stage 1: 24 engines; single round robin; TC: 60m + 10s. 
  2. Stage 2: The top 8 engines from stage 1 play a quadruple RR; TC: 90m + 10s.
  3. Superfinal: Top 2 from stage 2 play a 100 game match.

With the exception of stage 1, engines play both sides of each opening book position.


Live Games  24/7: