book: Game Changer: AlphaZero's Groundbreaking Chess Strategies and the Promise of AI

Sort:
dubrovnikchessset2

whats everyones thoughts on this new book about A0

https://www.newinchess.com/game-changer

Here's some videos by the author

https://www.reddit.com/r/chess/comments/a01d7a/deepminds_alphazero_on_the_carlsencaruana_games/

I'm looking forward to more insights from A0 and a stronger LC0

quadibloc
dubrovnikchessset2 wrote:

whats everyones thoughts on this new book about A0

https://www.newinchess.com/game-changer

The book is expected out in January 2019, so I haven't had a chance to read it and therefore come to conclusions about it. The topic certainly makes it sound interesting.

liml

bump

dubrovnikchessset2

the book was released today

 

dubrovnikchessset2

Is it a...

game changer?

 

dubrovnikchessset2

Please let us know if you think there’s anything new to be learned once you’ve finished the book.

BonTheCat
iRio wrote:

Yes. I already learned, that I am very weak chess player.

Everything is relative ... Any player above E2000 is in the top 50,000 in the world. Given that there are millions of chess players in the world, and Fide has half a million registered players on their official rating list (which would put you in the top 10%), but many countries are still only beginning to organize tournaments to give players a Fide-rating, it probably means you're easily in the top 5%, probably top 3%, among the world's players. A rating of E2100 puts a player well inside the top 1% in most countries.

Don't compare yourself with monsters. Deep Blue II defeated Kasparov more than 20 years ago, and today even the 'weak' (compared to Alpha Zero which was set on a super computer) old school brute force engines on an ordinary laptop are 500 points above Magnus Carlsen. Chess is easy for computers, given enough computer power. After all, it's a limited universe with few, and very specific rules.

dubrovnikchessset2

I’m curious if any new strategies or openings are created as a result of A0 and eventually LC0, or if they will just confirm what human GMs have already figured out.

 

For example we

see that A0 has frequently played the English opening and the Nimzo-Indian defense.

 

thats nice that it affirms what modern GMs know, but we already knew those were good openings.

 

Is there anything to learn that we don’t already know?

smartbryce

huh grin.png  BonTheCat

Verbeena

BUMP - Has anyone read the book yet and can give some opinion / review about it?

kindaspongey

https://www.newinchess.com/media/wysiwyg/product_pdf/9073.pdf

GWTR

https://www.perpetualchesspod.com/new-blog/2019/2/5/episode-112-gm-matthew-sadler-and-wim-natasha-regan-1

Loseronthechessboard

i allow myself to bump this thread so i don't open a new one for nothing happy.png
Would you say that it has any use for players under 2000 since the games have no real (human) errors and the sacrifices or positional decisions by alpha far out of reach for long term planning ?
I loved some reviews by chess.com on youtube, truly amazing games, but the lessons would really be that valuable for anyone?

Cheers

IpswichMatt

Slightly off-topic, but is there any other chess-playing AI being developed apart from Google's A0 and LC0?

BonTheCat
Loseronthechessboard wrote:

i allow myself to bump this thread so i don't open a new one for nothing
Would you say that it has any use for players under 2000 since the games have no real (human) errors and the sacrifices or positional decisions by alpha far out of reach for long term planning ?
I loved some reviews by chess.com on youtube, truly amazing games, but the lessons would really be that valuable for anyone?

Cheers

To me, it seems as if Alpha Zero is playing in a more human-like style. Those long-term sacrifices are of a type that humans would play (although not with the same enormous depth).

RussBell
kaukasar wrote:

BUMP - Has anyone read the book yet and can give some opinion / review about it?

https://chessbookreviews.wordpress.com/2019/04/01/game-changer/