While sleeping, watch!
United States
Unspecified
Minnesota
-
Tom Lehman says Cathedral is broken in his comment on the game. He says three is a solution where the first player always wins. Has anyone else seen this?
"This rating [7.5] is if you use my starting variant (first player goes, then the second player places *both* a piece and the neutral Cathedral; continue alternating moves normally).
Otherwise, I rate this a 2, as a somewhat interesting puzzle. Why? Because there's a start sequence that trivially allows the first player to win, no matter how the Cathedral is initially placed.
I don't post this start sequence because I think players enjoy exploring the game and finding it themselves, but I will say that I was challenged at a con by a well known and accomplished Cathedral player to demonstrate it (he even offered me $100 if I could beat him; I refused his money). I beat him 10 times running (by decent margins) and he then acknowledged that, yup, there does seem to be a problem here... "
Has anyone else figured it out? Who was the well known Cathedral player? Kind of makes me want to dig out the old polystone...
-
Russ Williams
Poland
Wrocław
Dolny Śląsk
-
The weblink http://www.hypersurf.com/~tlehmann/cathedral.html (here in Cathedral's BGG page) is broken, alas. But I found this googling:
http://sites.google.com/site/ptlehmann/gaming/hrules/cathedr...
-
W M Shubert
United States
Portland
Oregon
KGS is the #1 web site for playing go over the internet. Visit now!
Yes, I really am that awesome.
-
He is right. I'm pretty sure I found this problem also. Somebody doubted me once after I said that I had a guaranteed 1-player win, so we played online at a pbem server, and I beat him very soundly; plus, he reviewed the game, and confessed he can't see any moves he made that could have been changed to give him much chance at winning.
It's possible that I'm wrong, but once you know the strategy it is fairly easy to apply and I've never lost using it and going first. If anybody wants I can put here what my strategy is.
Edit: I followed Russ' link, and I'm pretty sure that I did indeed find the same strategy that Tom found. His description matches what I do, and his fix would probably make my strategy beatable. It's a clever fix for the problem. The only thing different is that I think there are two roughly equivalent first moves, either will work to guarantee a win. Later moves of course depend on what your opponent does, but if you know the goal then it isn't hard to figure out what your response should be, no matter how your opponent plays.
-
Russ Williams
Poland
Wrocław
Dolny Śląsk
-
William, searching my archives, I see you and I played in 2008, so perhaps you're remembering our game! This was Cathedral Board 387 at Richards PBEM Server:
1 CAnB1 2 ACeJ6
3 TOwD8 4 IFnC7
5 ACwA5 6 CSeG7
7 IFnH9 8 TOeF3
9 CSwE4 10 ABeI4
11 Resign
A B C D E F G H I J
_______________________________________
1 | | + | | x | | 1
|___| |___| |___ . . . . |
2 | + + + | x x | | 2
|___ ___|___ |___________ . |
3 | | + | | x x | x x | | 3
| | |___ |_______|___ |___ |
4 | | + | o | | o | | o | x x | | 4
|___|___| | | |___| |_______| |
5 | o o o | | o o o | x | | 5
|___ ___| |___________| |_______|
6 | | o | x | | x x x | x x x | 6
| |___| |___| ___ |___ ___|
7 | | x x x | x | | x | | x | | 7
| |___ ___|___| |___|___|___| |
8 | | x | o o | | o | | 8
| . |___|___ |_______| |___ |
9 | | o o | o o o | | 9
| . . . |___ |___ ___| |
10 | | o | | o | | 10
|___________________|___|___|___|_______|
But looking back at it now, I think I played weakly in the game. At least looking at it now, I have no idea what I was thinking. I'm sort of embarrassed posting the game record. 
Man, I haven't played Cathedral in a while. I should get it back on the table!
-
David Bush
United States
Lexington
Virginia
-
wmshub wrote: ... and his fix would probably make my strategy beatable. It's a clever fix for the problem.
I'm sure Lehman's fix would make the game more balanced, but it seems conceivable that computer analysis might some day reveal that even with this fix, one side or the other would have an advantage, albeit a more nuanced one. If placing the cathedral is regarded as the first move, then clearly the plain vanilla pie rule, a.k.a. the swap rule or one-move equalization, would not work, as no initial cathedral placement should be swapped, if I understand this correctly. But two-move equalization might be better. One player places the cathedral and one of his buildings. The other player chooses, at that moment only, whether to swap sides or not. This gives the second player a small theoretical advantage, but if there are ways to place the first two pieces which lead to an even game, then it could be argued that this protocol would be perfectly fair.
-
W M Shubert
United States
Portland
Oregon
KGS is the #1 web site for playing go over the internet. Visit now!
Yes, I really am that awesome.
-
That's the game! You did make some big mistakes, but the key part os the first move. Always as far from the cathedral as possible, giving you a chance to get one of two big corners. The "W"-shaped piece works also. Wherever the opponent plays, you go the other way; the side where you move will always end up bigger than anything they make without much difficulty.
As for the pie rule - yuck. I don't care for the pie rule at all. I like Lehman's fix much better. With his fix, then after the first player puts down a piece that threatens two corners, the opponent can block one corner with the cathedral and the other with a piece of their own. Problem solved and play continues!
-
Randall Bart
United States
Granada Hills
California
Red October
Earth is one of my favorite planets
-
The game is deterministic. Either one side has a guaranteed win or it's a tie. With Lehman's rule this game is as balanced as Chess. Without Lehman's rule, this game only works if you play pairs of games and add the scores.
How would this play with komi of 1.5?
-
-
Barticus88 wrote: The game is deterministic. Either one side has a guaranteed win or it's a tie. With Lehman's rule this game is as balanced as Chess. Without Lehman's rule, this game only works if you play pairs of games and add the scores.
The rules in my version of the game state that you are supposed to play the game this way. For some reason it seems like there are a lot of different versions of this game.
-
|
|
Posted Sat May 28, 2011 8:55 pm
