Crazy Stone Software

[Mivo]

I played Crazystone last night using level one. I beat it easily. But then I tried level two. OMG. Level two must be unbeatable because it literally took my breath away. No matter what I did in level two I was being killed mercilessly.

For your current strength, I would recommend Igowin HD. It plays much more gently at above SDK levels. Not weaker, just more friendly in the sense that its style is more human-like and closer to what you'll see when you play against other advanced beginners on KGS. It just plays less like a MC bot and you're likely to have more fun that way. (It just has the same awful stone clicking sounds as MFoG.)

]Regarding playing strength, I have no precise data to give a serious estimate. At some point I tried to get rated accounts on KGS to test, but I did not manage to get my accounts rated. I may try again later.

The CrazyStone bot was rated on KGS last year in spring or summer, wasn't it? I loved watching those games. Would be awesome if the bot returned. If not on KGS, perhaps on Kaya.gs where there seems to be less unnecessary bureaucracy.

[Rémi]

One last question, going a bit beyond Crazy Stone as it is: Do you think there'd be a worthwhile way to use distributed computing similar to SETI@home for Go AI? Of course I'd assume that to be way less efficient than a tightly integrated cluster, but perhaps the sheer scale could offset that. I'm sure there'd be a few thousand Go and AI enthusiasts around to offer their processing power.

I believe that even for a cluster it is difficult to scale beyond a few machines.

Rémi

[Rémi]

The CrazyStone bot was rated on KGS last year in spring or summer, wasn't it? I loved watching those games. Would be awesome if the bot returned. If not on KGS, perhaps on Kaya.gs where there seems to be less unnecessary bureaucracy.

Crazy Stone just played one game on KGS against Catalin Taranu (5p) at 4 stones. There will be another game tomorrow at 4pm German time. These are demonstration games at the European Go Congress:

http://files.gokgs.com/games/2012/8/1/egc2012pro-CrazyStone-2.sgf

Rémi

[Mivo]

Congratulations, Remi, looks like CS did really well!

Would you happen to know if the purchase of the 2012 version includes future updates? Or will the software have to be bought again when a new major version is released? I'm asking because I just saw that the 2012 version now comes with an analyse mode, but at $80 it's relatively pricey if it doesn't come with an upgrade option.

[Rémi]

Congratulations, Remi, looks like CS did really well!

Would you happen to know if the purchase of the 2012 version includes future updates? Or will the software have to be bought again when a new major version is released? I'm asking because I just saw that the 2012 version now comes with an analyse mode, but at $80 it's relatively pricey if it doesn't come with an upgrade option.

Unbalance is in charge of selling the program, and I don't wish to speak for them. This being said, they usually don't offer an upgrade option.

Rémi

[Mike Novack]

Unbalance is in charge of selling the program, and I don't wish to speak for them. This being said, they usually don't offer an upgrade option.
Rémi

The sort of thing that makes it so hard for us to compare prices (remember folks, when writing reviews, include details like that and the terms of the license (for one install, for multiple computers, etc.)

Rémi, that game just played. What was the hardware? Not just how many total cores but what chips.

[Mike Novack]

Never mind, just saw it in the game comments.

Interesting, for some reason I was picturing something even more powerful (in terms of total crunch).

[yoyoma]

From ~move 90:

RemiCoulom [3k?]: AMD Opteron(tm) Processor 8439 SE
RemiCoulom [3k?]: 24 cores

[Rémi]

There will be another game tomorrow at 4pm German time.

Sorry, the game will be at 6pm.

Rémi

[Mike Novack]

Will it get posted here like the last one was?

[Rémi]

Will it get posted here like the last one was?

Catalin won. He made some short comments at the end:

http://files.gokgs.com/games/2012/8/2/egc2012pro-CrazyStone.sgf

[Pippen]

Would be good to know more feedback from other players who have played with CZ on desktop PCs


Regards

There we go. I think if CS just is about 1d strong on a normal PC I wouldn't want to buy it since it would also analyse just as a 1d wouldn' it?

One way to solve all the discussions would be if CS was able to play as a bot on KGS. Everybody could let play "his" CS on "his" computer at KGS and would know soon its strength. It would be a huge motivational thing if I'd knew my CS is e.g. 2d KGS and I (1k KGS) could beat it maybe 9 out of 10 times (which would make me basically a 2d KGS ). Does anyone know if CS is runable as a bot on KGS?

[Mike Novack]

Describing hardware ......

An i7? An i7 what?

This is a perfect example of information that is at best partially useful. There is something like a 4.5 : 1 ratio in crunch power between the most powerful and least powerful i7's. There is something like a 3 : 1 ratio between the more powerful affordable* i7's and the least powerful ones.

* Not at all out of the question for one of us to have a desktop with say an i7-2600 or 2600k in it.

[Mike Novack]

Thanks Robert.

And I suspect that CS on your machine is probably playing stronger than 2 dan. Significantly more iron there than a typical desktop.

But I disagree somewhat about the "eveness" of the programs. MCTS is not an AI, just works, and is based on probability. For a given effective rating, probably weaker than that rating in the strategic sense but stronger that that rating in tactical fighting. All we can tell from the play of bots against a range of opponents is the effective rating from the results of games. Remember, we all make the same moves as the 9 dan for many moves of the game.