Remarkably Brilliant No-Counterplay Chess || Highly Evolved Leela vs Fire || TCEC 19

♚COURSES kingscrusher.t...
📚 kingscrusher.t...
FIDE CM Kingscrusher goes over Remarkable Game Changing Brilliant No-Counterplay Chess || Highly Evolved Leela vs Fire || TCEC 19
📚 kingscrusher.t... || kingscrusher.t...
♚ Play turn style chess at bit.ly/chessworld
♚ kingscrusher.tv - register and I will invite you for a game soon :)
FIDE CM Kingscrusher goes over amazing games of Chess every day, with a focus recently on chess champions such as Magnus Carlsen or even games of Neural Networks which are opening up new concepts for how chess could be played more effectively.
The Game qualities that kingscrusher looks for are generally amazing games with some awesome or astonishing features to them. Many brilliant games are being played every year in Chess and this channel helps to find and explain them in a clear way. There are classic games, crushing and dynamic games. There are exceptionally elegant games. Or games which are excellent in other respects which make them exciting to check out. Some games are fabulous, some are famous. Some are simply fantastic. This channel tries to find basically the finest chess games going.
There are also flashy, important, impressive games. Sometimes games can also be exceptionally instructive and interesting at the same time.
Info about Leela Zero:
en.wikipedia.o...
...
Leela Zero is a free and open-source computer Go software released on 25 October 2017. It is developed by Belgian programmer Gian-Carlo Pascutto,[1][2][3] the author of chess engine Sjeng and Go engine Leela.[4][5]
Leela Zero's algorithm is based on DeepMind's 2017 paper about AlphaGo Zero.[3][6] Unlike the original Leela, which has a lot of human knowledge and heuristics programmed into it, Leela Zero only knows the basic rules and nothing more.[7]
Leela Zero is trained by a distributed effort, which is coordinated at the Leela Zero website. Members of the community provide computing resources by running the client, which generates self-play games and submits them to the server. The self-play games are used to train newer networks. Generally, over 500 clients have connected to the server to contribute resources.[7] The community has provided high quality code contributions as well.[7]
Leela Zero finished third at the BerryGenomics Cup World AI Go Tournament in Fuzhou, Fujian, China on 28 April 2018.[8]
Info about Alphazero:
en.wikipedia.o...
AlphaZero is a computer program developed by the Alphabet-owned AI research company DeepMind, which uses an approach similar to AlphaGo Zero's to master not just Go, but also chess and shogi. On December 5, 2017 the DeepMind team released a preprint introducing AlphaZero, which, within 24 hours, achieved a superhuman level of play in these three games by defeating world-champion programs, Stockfish, elmo, and the 3-day version of AlphaGo Zero, in each case making use of custom tensor processing units (TPUs) that the Google programs were optimized to make use of.[1] AlphaZero was trained solely via "self-play" using 5,000 first-generation TPUs to generate the games and 64 second-generation TPUs to train the neural networks, all in parallel, with no access to opening books or endgame tables. After just four hours of training, DeepMind estimated AlphaZero was playing at a higher Elo rating than Stockfish; after 9 hours of training, the algorithm decisively defeated Stockfish 8 in a time-controlled 100-game tournament (28 wins, 0 losses, and 72 draws).[1][2][3] The trained algorithm played on a single machine with four TPUs.
...
Relation to AlphaGo Zero
Further information: AlphaGo Zero
AlphaZero (AZ) is a more generalized variant of the AlphaGo Zero (AGZ) algorithm, and is able to play shogi and chess as well as Go. Differences between AZ and AGZ include:[1]
AZ has hard-coded rules for setting search hyperparameters.
The neural network is now updated continually.
Go (unlike Chess) is symmetric under certain reflections and rotations; AlphaGo Zero was programmed to take advantage of these symmetries. AlphaZero is not.
Chess can end in a draw unlike Go; therefore AlphaZero can take into account the possibility of a drawn game.
AlphaZero vs. Stockfish and elmo
...
♚ Register at kingscrusher.tv and KC will invite you for a game!
♚ Or if you prefer bit.ly/chessworld
#KCChess #KCLeelaChess #KCMustSeeChess
❓ Any chess questions? www.quora.com/...
♚TEAM kingscrusher.tv...
♚TWITTER kingscrusher.tv...
♚DISCORD kingscrusher.t...
♚TWITCH / kingscrusher
♚REDDIT kingscrusher.tv...
♚QUORA kingscrusher.tv...
♚DONATE paypal.me/king...
♚COURSES kingscrusher.t...
♚COURSES kingscrusher.t...

[@kingscrusher]

My new Opening course at Udemy: kingscrusher.tv/openingtango

[@ulfnarverud1661]

What a knight-mare for black! lol

[@johnnyhyperborean1035]

Was thinking the same😂

[@alexandruandrei2045]

Very nice! Totally missed Nxd6 ...although I’ve seen Qa1+ or at least tried to make it work after Nxc5 or Nxf6. Also there were many subtleties with the white queen protecting key squares in the position. Very instructive! Nice work on the coverage!!!

[@joseraulcapablanca8564]

A true knightmre for fire. i am glad that the suffling stops others from analysing this. basman like win. only Your enthusiams and Insight into chess philosophy does justice to this kins of stuff. This weekend i saw you play some blitz and beat David Howell, if it is not too cruel you could analyse that for us. Thanks KC and keep up the good work.

[@raylopez99]

Yes, I second that opinion by my famous friend Jose Raul Capablanca. KC's win over Howell deserves a seeing.

[@kingscrusher]

Here it is: kzbin.info/www/bejne/gGLKo5Kca5msiLM - Cheers, K

[@tomv7357]

This high level shuffling is beyond me.... At how many moves did the game end? Feels inhuman

[@AndrewBackhouse1]

I was loving the Basman comments!! Haha

[@brettaspivey]

Just looked at the Leela vs Scorpio game. Absolutely insane, continual positional sacs, you will want to look at that one, chess from another planet!

[@davidchang8428]

link?

[@pablosartor6715]

Awesome game by Leela! Very well explained as usual. Thanks KC!

[@Bobby-fj8mk]

Great game. The Knight is the only piece that can move in the 3rd dimension which is the most useful in a closed position. That's what we saw. Those 2 Knights were making fools out of the 2 Rooks.

[@MoonBurn13]

On the pop quiz where the line’s critical move is QQR1ch, I got the answer, partly perhaps because I’m familiar with the observation that the horizontal move of the Queen is (for human players) “the 2nd most overlooked move in the game”. Which brought to mind: It would be great, KC, if you could do a vid, or even a series, on overlooked moves, and compare human error with machines’.

[@gregainsborough9866]

Thanks for the great commentary!

[@stevensturgillschesstv4821]

Fascinating KC!

[@pavanlulla]

You help take chess to the next dimension...

[@SmithnWesson]

Why does Leela do all that high level shuffling anyway? Do you every see that when Stockfish (non-NN) plays against itself?

[@robharwood3538]

It's due to the way she's self-trained with some small amount of randomness thrown into the training (which is important so that the training doesn't get stuck at 'local minima'). This means that her 'attitude' is that "There's always a chance the opponent could blunder," so she gives them every opportunity to blunder before making a committal pawn move or capture. She's not optimized for fastest wins, but rather for highest probability of winning. Of course, when playing against a top top engine like Stockfish (or Fire in this case), the chances of such a blunder is almost zero, but Leela doesn't know that. All she knows is from her training games, where blunders/inaccuracies due to slight randomness are always possible. One way to understand it is to imagine that maybe all of a sudden a human takes over for Fire. While engines playing nearly-optimal end games don't blunder, and so you get 'high level shuffling', the odds of a human blundering (even a Kasparov or Carlsen) are much much higher. So, if she can give them more opportunities to blunder, that would increase her probability of winning. So she shuffles. Also, it often actually *does* trigger engine 'blunders' in clocked games where the extra moves end up burning down the clocks, and Leela's 'intuition' starts to shine while A/B engines start to struggle getting enough search depth to outplay Leela. (This only works in Leela's favour if there are still several pieces on the board, such that the opponent doesn't get as many end-game tablebase hits as when there are only a few pieces left. But it still works for her often enough to be a noticeable pattern in her winning play.) KC doesn't usually cover the details of the time troubles and engine evaluation graphs that are clear factors during the live games, but that information is available in the PGN files available from the TCEC website if you load it up in your favourite Chess GUI. Also you can just use the Archive on the TCEC website to view the games directly from there without having to download the PGN. If you review the games in more detail this way, you can often see how inaccurate evaluations (for a long time Stockfish used to have serious trouble correctly evaluating 'fortresses' and 'pseudo-fortresses', for example) and time troubles, and EGTB lookups can significantly influence the game play, and help explain why Leela's shuffling does have a kind of counter-intuitive logic to it.

[@andromeda4995]

I knew kingcrusher only from string dogg and... The hair

[@tron103]

Great job, your videos are academic and edutainment sir

[@kingkura]

That endless shuffling is actually unimpressive I guess the NN will be trying to win on time in drawn positions with the nothing shuffle.

[@LeopardBella]

Basman taught Leela everything she knows!!

[@fairytalejediftj7041]

If you don't go for the fawn pawn, are you even chessing?

[@richardfredlund3802]

ha ha 'a few hundred moves of shuffling then the knights are much better than the rooks...' 'that's the basic recipe' (made me chuckle anyway). Why open the h-file when you can install a thorn pawn and wait 300 moves to get a better position. :)

[@laynemccormic9102]

Thorn pawn? form pawn?

[@robharwood3538]

Fawn pawn is the meme these days. 😅

[@johnpetkos5686]

Am I the only one in the whole universe who hears something like "phone pawn'' or "foam pawn" or, even worse, "fon pawn''????

[@robharwood3538]

It's a running joke in the TCEC community (e.g. on the live chat during games) to call them 'fawn pawns'. It's just part of KC's regional dialect/accent where 'th' sounds very close to 'f' for other English speakers. Many English folk (I'm not sure the specific region) have the same accent. But it's become a kind of meme now, along with his 'flamboyant' or 'bombastic' video titles like "highly evolved Leela the mighty Stockfish". 😅

[@nicholasperkins4655]

How did Leela "know" to keep the pawns symmetrical on the queen side? In most games that is considered drawish.

[@AndrewBackhouse1]

I guess it calculated a closed position would favour the knights.

[@robharwood3538]

By playing millions and millions of games against herself, and learning useful patterns from them, basically. But, to be fair, most games really *are* drawish. KC typically only covers decisive games between engines (with a few exceptions here and there), so we're getting a slightly skewed view of those games where her tactics/strategies end up actually leading to a win rather than a draw. Still, she does seem to show a lot more care for pawn structure than most other engines, particularly the traditional A/B kind. I guess it's just a good pattern?

[@martinmartin6300]

Was Besmont secretely a genius?

[@stevebloom5606]

He oozed genius.

[@adampolcz3692]

.

[@lancejira9903]

Am fourth.grrrrrrrrr