Alpha Zero finds shockingly secret novelty in highly evolved Giuoco Piano vs Stockfish 8 - Game 9

♚COURSES kingscrusher.tv/chesscourses
📚 kingscrusher.tv/chessopenings
📚 kingscrusher.tv/chesstactics || kingscrusher.tv/londonsystem
#KCComputerChess
#KCAlphaZero
♚ Play turn style chess at bit.ly/chessworld
♚ Play Chess vs. Kingscrusher and others: www.chessworld.net/chessclubs...
♚ Subscribe to best KZbin Chess Video Channel : bit.ly/kingscrusher
👕 Kingscrusher Store: teespring.com/stores/kingscru...
♚Donating via Paypal: goo.gl/7HJcDq
Game quality tags: amazing, awesome, astonishing, brilliant, classic, crushing, dynamic, elegant, exceptional, excellent, exciting, fabulous, famous, fantastic, finest, flashy, greatest, important, impressive, incredible, instructive, interesting, magnificent, marvellous.
Info about Leela Zero:
en.wikipedia.org/wiki/Leela_Zero
...
Leela Chess Zero (lc0) is a free, open-source, and neural network-based chess engine and distributed computing project.
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.org/wiki/AlphaZero
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
...
►If you like my chess videos please consider become a full member here at my Chessworld Site. www.chessworld.net/chessclubs/...
#KCChess game tactics
♞ Challenge KC and others for turn style chess at bit.ly/chessworld
♚TEAM kingscrusher.tv/team
♚TWITTER kingscrusher.tv/twitter
♚DISCORD kingscrusher.tv/discord
♚TWITCH / kingscrusher
♚REDDIT kingscrusher.tv/reddit
♚QUORA kingscrusher.tv/quora
♚DONATE paypal.me/kingscrusher
♚COURSES kingscrusher.tv/chesscourses
♚COURSES kingscrusher.tv/chesscourses

[@kingscrusher]

📚 My chess courses: kingscrusher.tv/chesscourses

[@leslassiter6378]

This is the second time I have tried to down load a game when it failed. All I get is the bottom of the page, which is strange because it used to work flawlessly. Can something be done to fix this?

[@OrlOnEarth]

It's gioco piano in Italian, giucco is not even a word.

[@piyushsharma5938]

Alpha Zero was just told " if the opposite king has no squares to move on check then you win " And it replied " I am ready now "

[@kingscrusher]

Fun comment - Cheers, K

[@ped7g]

The love/hate relationship for me is, that when you ask A0, "how are you playing that and why?", it goes "you just play a6, Nh6, ..." saying the moves, but never really explaining anything... the deep-machine-learning-NN are sort of human "blind spot", not intelligent in human-like way, but still so "easy" [to create] and powerful, that they will get quickly injected into all sorts of solutions and industries... yet people will have difficult time to tell what [how/why] exactly they do, and if they go toward "wrong" (in human point of view) bias over time, it may take serious time and effort to just notice, not even mentioning correction. Then again it's probably good there's something so "alien" challenging mankind now, while it is in its infancy and not *that* powerful yet. Would it come in the form of what it will be 10-50 generations later, it would be probably completely incomprehensible, what did "hit" us.

[@jnoter]

@@ped7g your estimation of 10-50 gen for humankind? because that's a bit too long.

[@merijnvanwouden]

ped7g There’s neural networks that provide reasoning steps that humans can understand with the results that they output. This is often required for AI use cases where the reasoning behind choices is a legal prerequisite, or when the AI gives help to a certified expert to form a decision. Medical diagnosing is an example but there are many others. This just requires a major rewrite of the code and was now not a priority in A0 or Leela

[@jeffk1722]

"Don't put a knight on the edge in the opening. Also, don't double up your pawns." Alpha Zero: We'll see. Go back to your little sandbox, human.

[@matthewviramontes3131]

Okay, so I've been silent on all these AZ videos that he's been doing because it is vs Stockfish 8, and so I'm not as impressed as if it were against a more recent version of SF, however, SF8 is nonetheless rated like 3400, and this is a very common opening, so I thought I'd share my thoughts. So, a6 is a cool little novelty, but what really blew my mind was the castling by Black. I mean, *inviting* your King to be exposed with an open file AND doubled pawns?? That's unheard of! As I started to analyze the position though, it did kind of make sense. White capturing the Knight does expose the King and weaken Black's pawns, but as well White's position becomes less strong because what attack does he now have on that side? His own Knight is pinned, the Bishop that was attacking that side is gone, the Rooks and other Knight are idling on the back rank... there's just no pieces available to launch an attack! And then, AZ just totally gives up 2 pawns and allows it's Rook to be attacked, but in doing so displaces his enemy's Bishop and gains an open file. People, what you're witnessing is a new level of chess that us humans couldn't figure out in over 1,000 years of playing this game. The word "amazing" doesn't even begin to describe it.

[@mylifematters858]

These games were played against the strongest stock fish available at the time . Deep Mind claims it easily crushes 10 also

[@warrendsmith6832]

You said it baby. AZ in some of these games is playing like nothing I ever saw.

[@Th3Wick3dOn3YT]

Yes I did notice the displacing of the light square bishop being part of the trade off for the two pawn and the open-file. I thought that was super interesting, but this kind of stuff is kinda the thing you say "hey, people please dont try this at home" If that bishop is activated quickly then the compensation is lost. So, MASSIVE, I cant stress that word enough MASSIVE presicion is required.

[@asherasator]

I think these games were played before SF 9 came out.

[@knightoflambda]

I have no real experience playing chess. But from what I see of A0's playing style when it's playing Go or Chess is very, very high level. Like taken straight out of the pages of the Art of War. It LOVES limiting its opponent's strategic potential and tactical options, while expanding its own. It's willing to sacrifice ground, material, and give and give until its opponent finally realizes it's surrounded and out of good moves to play. Some pro Go player (paraphrased) said that playing A0 was like playing against an irresistible tide, one second you're going about swimmingly and then another you're lost at sea. It's written that all warfare is based in deception, and AlphaZero deceives. It will give you the best momentary advantages and also enough rope to hang yourself by. It also practices mobility and force concentration. There's this one game: kzbin.info/www/bejne/rabMYYytmsqpqrs where it was getting ready to hammer Black's king side, but couldn't marshal enough force to do it. It didn't hesitate to abandon the position and repurpose its rooks. The backbone of its strategems is strategic flexibility. I shudder to think what future human warfare would look like, run by AI 'generals'. AlphaZero writ large, canvassing a theater, in every battlefield, giving orders to every drone and soldier. If the other side didn't have an AI of their own, they probably wouldn't last long at all.

[@theguillaume73]

Morphy is back

[@badcornflakes6374]

Reincarnated

[@-_Nuke_-]

This game should go out as a shock to all the Super GMs out there including Carlsen. With their Londons and their Sicilians and their Petrovs. Here we have a Giuoco Piano played just like Chess is meant to be played. We really though back when computers starting getting into chess, that chess is going to die because computers will discover everything and leave no room for the chess romantics and everything will end in a draw. And it kinda did for a while, engines taught us how to play like them, to crave for the draws as they are the ultimate price and win only if our opponents make a slight inaccuracy that we can capitalise. But here A0 demonstrates that its ourselves and our opponents that we need to understand - not the game. A0 playing against Stockfish, unknowingly understanding Stockfish's weaknesses in its evaluation, placing Stockfish in an environment he can control, giving Stockfish just a bit of what he thinks he wants and then letting Stockfish hang himself in the process as he thinks he's winning the game... That's how chess is meant to be played and we have to thank the AI for actually saving chess right now!

[@JohnDoe-ns8ho]

I agree, except for the part about engines craving draws. When top engines play each they draw far less than top GMs, and would draw even less if their positional analysis wasn't inspired by humans at its roots. It's like engines tried to teach us how to play like them but we declined most of it because drawing every game is safer. It's why these top GM games have really dynamic play up to the point the engine analysis ends and then it's find-the-draw with winning dynamics ignored. What AZ is doing is really just hammering the original computer home, that chess isn't a dead draw.

[@-_Nuke_-]

​@@JohnDoe-ns8ho agreed! Chess is a draw if played correctly, yes but we can make games of chess to hardly ever be a draw by simply removing increments all together. You have a curtain amount of time to play the entire game and that's it, if you don't probe deep enough you loose on the board - if you probe too deep you loose on time, therefore you have to be efficient. Maybe we need to try that out in our championships and see how it works! I think the games will become more dynamic, more crazy, with many more complications with people and machines trying to maximize their efficiency under a limited time factor. I don't think the games will become worse theory wise, actually the opposite might come true, new theory will have to be invented in the start of a new era. So it's not a problem of chess that most GMs games are boring and drawish, it's the way that we play the game that is the cause of that.

[@GerSHAK]

+

[@SpocksBro]

A conventional chess engines weakness is that they rely on fixed position evaluation that has been pre programmed by humans to their best knowledge (weight of each piece, value of open lines etc.) AlphaZero instead started from scratch and adapts it's position evaluation dynamically through neural networks. It therefor comes to no surprise that it evolved such a surprisingly different way of playing mid game chess compared to what we knew so far.

[@bobbwc7011]

@@JohnDoe-ns8ho Engines draw less just because pure calculating power in the endgame makes the difference. humans don't normally play 100, 200, 300 moves. Chess engines are machines, playing shit out to the last move ruthlessly.

[@u.v.s.5583]

How many schoolchildren will get mated with black in this opening now?

[@lotsagrapes]

10/10 for titles lol. been watching you since about dialup era. Your thought patterns and cadence are forever burned into my memory, thanks for leading the way

[@kingscrusher]

Fun comment - Cheers, K

[@ArthurHau]

I have seen only a couple of games played by the new Alpha Zero AI. But oftentimes, it tries to keep all the minor pieces, especially its two bishops which often turn out to be attacking the king side at a remote distance eventually. I also kind of find Alpha Zero not starting to attack early on. Instead it often tries to get all its pieces out asap. It doesn't really mind a double pawn structure either as if it is ok by nature. Alpha Zero also doesn't mind giving up many pawns early on as if the pawns are actually obstructing its development and middle game attacks. The sooner those pawns are gone, the better control Alpha Zero has by its minor and major pieces. The pawns of the opponents seem to be liabilities cause they are blocking most of the possible attacks, especially by the long range pieces, rooks and bishops. Has anyone seen a game played by Alpha Zero launching some kind of queen side attack? I have not seen one so far. Alpha Zero doesn't seem to bother with queen side attacks.

[@GistTech]

Been a sub of the channel for more than two years, one of the best decisions ever made

[@joseraulcapablanca8564]

A very interesting concept here.well carried out plan too and a fine commentary thanks KC keep up the good work

[@krinkovakwarfare]

Alphazero has the imaginative strategic engagement of Bent Larsen, the accuracy of Paul Morphy, and the tactical manoever of Mikhail Tal.

[@robertstrong6639]

Watching Alpha Zero's unorthodox moves is like chess from the twilight zone, otherworldly, counterintuitive, and fascinating

[@bobbwc7011]

Alpha Zero is like Kasparow as a chess AI. I remember when i was a club player and the new issues of the chess magazines came in. I immediately went to Kasparow's games because: Always action, nonstop. People who did not grow up in the late 80s and 90s simply do not understand the impression Kasparow left on the game, the degree of innovation and playing crooked, highly dynamic, extremely sharp and unclear stuff. It was the absolute antithesis to Karpow yet positionally very sound (unlike Tal ;) for example). Alpha Zero is a deep irony in chess: Not Karpowian-Carlsen chess is the very top stuff ...Kasparowian chess is.

[@pazdziochowaty]

Someone swapped + with - on the output of power supply while AlphaZero was learning. This made it play a6 instead of Nf6 in the Italian and Nf6 instead of a6 in the Spanish

[@ericb252]

Love your passion for this game. I was waiting for your review of this one!!

[@LordoftheJamesClan]

One of my fav chess matches I seen in a while. Watching computers fight each other for my amusement is one of my favorite pastimes :p

[@paulbloemen7256]

Just a thought I would like to share here. In not a too long period of time, an AI-Zero opening book could be available of all the white and black moves AI-Zero likes, considers best: let's call it the platinum repertoire. Interesting enough for us humans, nothing really to add, or? Well, there are a lot of moves missing in this platinum repertoire, many of them considered good or interesting by us, humans. Why are they not in the platinum repertoire? These moves should then be fed to the AI-Zero, the outcome to be expected is a whole series of refutations, no doubt. My guess is, that this group of not quite good moves will be the most interesting to study, to find out where our top humans were wrong, and why. The whole "What if...?" exercise would be super rewarding, not only on the moves themselves, but also on the concepts hidden behind those moves. Human concepts in the end, only we humans need them to shape our progress. AI-Zero only looks for the moves that make it win the most often.

[@teslathejolteon8007]

Is that a kind of analyser?

[@paulbloemen7256]

TeslaTheJolteon As far as I know, there are no special tools yet to distill favourite openings from games played by AI-Zero machines. It would be great to have such a kind of harvesting machines, based on thousand or millions of games, with all kinds of statistical data, needed to find out which moves are best. Right now, the effort is probably solely focussed on machines gaining playing strength; teaching tools etc. will come later, no doubt. Entering "moves on request, to be analyzed" would be super to build that refutation knowledge base. I can imagine that these tools are built by other parties, using the AI-Zero machines as input on request.

[@dustsettler4596]

@@paulbloemen7256 this idea is brilliant and it would definitely work.

[@timmy1729]

Agreed

[@benjaminmiddaugh2729]

@@teslathejolteon8007 I suspect the easiest way to do this is run a statistical analysis of Alpha Zero (or Leela) to determine their favorite choices (sorted by opposite opening and whatever other considerations would be necessary) from a giant database of games. Then train another instance of the neural net against an original only playing a chosen opening and see what happens. If there are ways to win, that method would find them, although the training criteria might need adjustment to compensate for the consistent opening moves by the opponent.

[@charmainattwood977]

Incredible game. Thank you.

[@chessloversonly]

Great vid!! Very interesting!!!

[@eshneto]

Cool video. The channel is getting close to 100k subscribers!

[@klimtkiller]

alphazero the rebellious teen going against social norms

[@kingscrusher]

Fun comment - Cheers, K

[@ArthurHau]

There is no such thing as social norms in Alpha Zero. It has no human knowledge input and hence it is playing using different strategies that are beyond human knowledge or even human understanding. We should all learn from Alpha Zero, not the other way around. Alpha Zero WILL become our social norms. LOL

[@dannygjk]

@@ArthurHau You missed the point.

[@echoeversky]

Millennials are done with our stuff, including our millennial AI’s. :3

[@casias002]

Hopefully it doesn't "grow up".

[@xyzct]

This shows that well trodden territory can become a mental rut.

[@paulbloemen7256]

Nothing short of AMAZING! Again high level AI-Zero shows its point, freeing us humans from our stifling concepts, making our world both bigger and better. GO-GO-GO AI-Zero!

[@solosaroman]

Stokck fish is not obrying the rulle:"to take is mistake".

[@peterpetrov6522]

Awesome!

[@amberw.1701]

Reminds me of a Fischer game in two knights defense, playing nh3 then casting, against bisguier I think

[@boriserjavec6470]

ahhhha yes the classic double mouse click at the end of video :D

[@dickbrazen]

You said Ba7 was a blunder, but paused at 11:31 I'm guessing 1.◼e5xd4 2.◻c3xd4 ◼Ba8xf3 3. either/or ◻g2xf3/Rf1xf3/Nxf3 ◼Ba7xd4 and if I'm not getting confused black would have a much stronger position.

[@teslathejolteon8007]

There is also a game between AlphaZero and Stockfish where Stockfish is white and plays the Italian game, but instead of AlphaZero playing Nh6 in the opening, it played Qf6.

[@WillToWinvlog]

Nice knife f4 reference!!

[@kingscrusher]

Cheers, K

[@tomzhangg]

@@kingscrusher its knife f5!!!!

[@cheya111]

In retrospective, when you look at it... Black have lost the dark bishop, and the knight on the king side is pinned. In other word, black king is safe and that is a novelty. And it is black using the open G file to launch the attack.

[@swxqt6826]

AlphaZero is the perfect proof for why the Bishop pair is better in the mid game than the rooks

[@Zardoz37]

I'm curious to know what happens if White declines the sacrifice of the pawns after 13...Ne7. The a2-g8 diagonal looks like a good one for White's Bishop - can White take a leaf out of Black's book and play for piece activity instead of material with a move like Ba2, on either move 13 or move 14?

[@KILLERAOC]

Whoa! I’m such an amateur I know this position well! This is how it’s done!

[@mumu135mumu554]

00:40 It really did surprise me that you said castling is the common move here. c3 is the main move, not castling.

[@magnetcar]

This is just awesome novelty

[@dieghito525]

Very accurate analisys

[@kingscrusher]

Cheers, K

[@dannygjk]

4.d3!? that's why AlphaZero's opening plan is sound.

[@zac8766]

ironic variation considering giuoco piano means "slow game"

[@KrishanBhattacharya]

13:09 FORN PAWN!!

[@simohayha6031]

Two questions. I looked at the cccc schedule at chess.com I saw that game 307 or something is between leela#3 vs Houdini #1. But I didn't see a date. You know it? Also is it possible to have Leela analyze and learn from all the A0 vs SF games?

[@ramdamdam1402]

Short answer : no

[@matthewviramontes3131]

A date for what? The game will be played when all the others before it have finished. There's no set date for when they play.

[@simohayha6031]

@@matthewviramontes3131 oh so it's just one game at the time and one finishes the next one in line starts right away?

[@matthewviramontes3131]

Yes

[@sergiu2325]

So alpha zero can become even better than he is now just by playing against itself.

[@ericb252]

At the start instead of white doing d3 what if they do e3 and then d4. My stockfish engine on lichess is showing massive advantage for white when black responded to e3 with A5. Was I suppose to do to do something other than A5 in response to e3?? Something went horribly wrong in my noob game. I love this idea but failed hard on it lol What does Leela think? I understand our tools to analyze opening books might be limited when working with alpha zero

[@bruceli9094]

To Alphazero every chess piece is the same value. Even the King can be sacrificed!

[@GerSHAK]

lol :)

[@badcornflakes6374]

Morphy like

[@marcusfossa6695]

A7: the alpha zero variation.

[@Durian_Defense]

Highly evolved video title.

[@kingscrusher]

Cheers, K

[@kamshik]

Was gonna make a comment that every title is highly evolved and everything is a shocking novelty revelation, but you beat me to it :D

[@pand9293]

This will be named the "Alpha Zero" variant of the Giuoco Pianissimo !! Mr Gavriel I sent you a question on Quora about Alpha Zero, would you like to answer please?

[@GerSHAK]

+

[@offrampt]

Fings that feery hadn't fort abaht.

[@dickbrazen]

I do a6! This is big news for me because even though I've suddenly been learning very quickly, I'm not historically a regular winner at chess. After the beginner level, there's always been a big dropoff.

[@dickbrazen]

I've been exactly there for black at d6. I can't wait to find out why it failed.

[@henrry9099]

At 13:54 why didn’t King take E3 can someone explain?

[@ckr1011]

Could you speed it up some?

[@budesmatpicu3992]

Is it somehow known "how far" upfront is AlphaZero "seeing" the game? Perhaps not 100% accurately, but somewhat "roughly", "human like"... it has to estimate that it has good attack, perhaps even "qualitatively" evaluating posible moves, e.g. "that corner bishop will not be stopped/obstructed so queen can do damage with knight..." (without any precise variant calculations).

[@ImAllInNow]

As interesting as these AI games turn out, are they really that instructive for normal players due to the intense amount of calculation necessary to navigate that middle game? So many things could have gone wrong in that middle game and the moves needed to be exact to make your way through it.

[@bobbwc7011]

The last player who played like Alpha Zero was Kasparow. So a few very top players will be able to utilize these innovations but they have to excel in dynamic positions. Which is quite an irony: In the 1970s Karpow and his mechanically engineered, positional chess was the gold standard. It took a Kasparow to dethrone this pure positional play and replace it with something new, something very rewarding but something very dangerous: sharp, crooked positions which are highly dynamic and actually minefields filled with wrong moves. During the unprecedented reign of Kasparow, other creative aggressive players had a chance to bloom. Then the computing power available at home became higher and higher and higher so that the chess engines started to absolutely dominate GM training, analyses and preparation. This led to a bloated elite of cautious players with a Karpowian beast at the top. But Carlsen also implemented elements of Fischer's style because chess engines play very tactically and dynamically...the type of player which faded away after Kasparow retired. In the future, a player type like Kasparow is the only way to handle the computer-dominated evolution of chess, and such a type of player is far more rare than a Karpow-type player.

[@chessmaster9070]

stockfish got fool incredible game

[@strangevideos3048]

2:27 white should play bishop g5 to eliminate that anoying knight.....

[@tomzhangg]

it's knife f5!!!!

[@Th3Wick3dOn3YT]

I think, I am missing some more IQ points to fully understand A0, it gives too many pawns for activity. Usually we give 1 maybe 2 and look for compensation but this 3 or 4 pawns down its just not practical. Like Elon Musk would say " I need an increase by a factor of 2".

[@asherasator]

I've played like this naturally & realized a long time ago when you give up 3-4 pawns or a piece for activity or space, it's hard for the opponent to maintain the material advantage because if you played well with a plan it becomes dangerous for the other side & end up giving the material back or at least some of it.

[@allanfifield8256]

Alpha Zero loves the Bishop pair. Seems to "value" the the Bishops at about 4 each with the Bishop pair even with what appears to be a bad bishop. I am starting to "doubt" Stockfish 8 in these positions; it seems to play aimlessly with no plan while Alpha Zero slowly mobilizes and improves.

[@allanfifield8256]

Actually rephrasing seems to value the bishops as much or more than the rooks with the bishop pair and no useful mobility for the opposing rooks.

[@zakaria3557]

Great attacking chess

[@kingscrusher]

Cheers, K

[@FMDD168]

No mention made that the Actual Losing move for Stockfish was 51. Ke2? when 51.Bd1! Rg8 52.Kf4 Bd2+ 53.Ke5 would have held the balance in the game. The Devil is in the Details.

[@ingeniouswild]

interesting analysis but super dense ads destroy the viewing :/

[@davezick800]

The play is somewhat similar to an old steintz idea noted by fischer in his my sixty book,,,the line is in a panawn sac variation in the fried liver declined...or whatever it is called..i have forgotten in the intervening thirty years..e4 e5 nf3 nc6 bc4 nf6 ng5?! d5 exd5 na5 bb4check c6 dxc bxc and now be2 (bf1?!) ..the knight gets hit with h6 and, to avoid e5, hops to h3 either before or after castling...i have played this many times thirty years ago when i was something of a blitz hustler...fine play, no problems...so the basic idea is rather old

[@davezick800]

the bishop often redeploys to g4;;;fischer noted the fine play,,different variation, true, but a similar idea

[@LucilleMilo]

i sort of enjoy your videos. often it is unclear who is playing which color pieces. this one you made a quick reference in the beginning, so not so bad.

[@kingscrusher]

Haha many thanks!

[@dannygjk]

It has been the convention for generations that the white player is mentioned first, black second.

[@magicstix0r]

This is novel? I play this opening all the time.... and lose...

[@peteralainszpiriev4750]

Two old patzers try to to play chess but they cant.

[@dannygjk]

Stockfish plays Bxh6. Judge Alpha: "I knew you'd play that." kzbin.info/www/bejne/q16YZYKmetx4Ztk

[@savvytravvi8660]

Great analysis as always. However, is it possible to hear less of your throat noises? Like experiment with microphone placement or something?

[@dannygjk]

I think he smokes sounds like a typical long-term smoker.

[@RollingGod24H]

Hahahahahahahahahahahahahhaha

[@kingscrusher]

Cheers, K

[@VenturaPiano]

Sort of misleading to say deepmind sent you these games don't you think? (Unless ofcourse they did send it to you directly)

[@kingscrusher]

Email contact was made - yes.

[@VenturaPiano]

@@kingscrusherahhh okay then I apologize

[@fleecemaster]

@@VenturaPiano Deep mind sent games to all the chess commentators, other commentators have also spoken about it

[@humanrightsadvocate]

Could you please keep the titles of your videos descriptive and abandon the tabloid wording? It cheapens an otherwise good video. And quit with the "highly evolved" stuff. Alpha Zero is the only game in town. There is nothing to compare it to. It's not "more evolved" than anything else. It's only "more evolved" then earlier versions of Alpha Zero.

[@GaussianEntity]

The title is incredibly descriptive if you're aware of the history of Chess and the Giuoco Piano opening. The a7 move by Black has long been considered subpar and even awful by some Chess books so the move is shocking in itself that it led to a win against a powerful engine. The highly evolved portion refers to the opening, not Alpha Zero. Anyone familiar with the opening knows that the games look nothing like this, hence the descriptive of highly evolved. We have a lot to learn from these openings still. So what's wrong with the title?

[@humanrightsadvocate]

Yeah, sure! The title is "very descriptive". In fact, we can say that kingscrusher found a shockingly secret way to title his videos. Let's all hope he brings back the "megaexciting" adverb. Don't you miss that? Then, the title can be something like, "Highly Evolved Alpha Zero finds shockingly secret novelty in megaexciting Giuocco Piano. That would be very descriptive and not tabloid-like at all. A descriptive title for a chess game contains _only:_ - name of the event - date of the event - names of the players - name of the opening - fucking nothing else @@GaussianEntity

[@GaussianEntity]

@@humanrightsadvocate Yeah sure, let us know how that works out. I'm sure you'll rake in the views with that rigid formula lol.

[@humanrightsadvocate]

You're the kind of a stupid fuck that needs a tabloid click-bait title in order to click on a video. @@GaussianEntity