Author |
Topic: Endless Endgame Event? (Read 81986 times) |
|
Fritzlein
Forum Guru
Arimaa player #706
Gender: 
Posts: 5928
|
 |
Re: Endless Endgame Event?
« Reply #15 on: May 15th, 2012, 6:03am » |
 Quote  Modify
|
on May 14th, 2012, 6:42pm, browni3141 wrote:
Perhaps 31g is responding to the threat that you would switch your dog and rabbit on a3 and b3? (That better not have been a goal in two on 31s, since you don't miss those  )
|
|
 IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: Endless Endgame Event?
« Reply #16 on: May 15th, 2012, 9:36am » |
Quote Modify
|
on May 15th, 2012, 6:03am, Fritzlein wrote:
Perhaps 31g is responding to the threat that you would switch your dog and rabbit on a3 and b3? (That better not have been a goal in two on 31s, since you don't miss those ) |
|
Indeed 31s db3n db4w ra3e da4s is a forced win, but according to the bot it's a goal-in-3:
31s db3n db4w ra3e da4s
32g Ed2s Ed1w Rb1w Ec1w
32s Rf5n Rf6x df4n De3n ed3e
... and the goal threat is unstoppable so silver goals on 33s.
So browni3141 did not miss a goal-in-2 here, but in another game he played yesterday he did miss a goal-in-2 against him at 18g:
http://arimaa.com/arimaa/gameroom/comments.cgi?gid=230097
18g Cat to f1 seems to defend for the time being.
|
|
IP Logged |
|
|
|
browni3141
Forum Guru
Arimaa player #7014
Gender:
Posts: 384
|
|
Re: Endless Endgame Event?
« Reply #17 on: May 15th, 2012, 3:21pm » |
Quote Modify
|
I'd like to revise my previous statement and say that goals in two should be easy. I could argue that that game you cited was a blitz game, and that 15s isn't quite enough time to find all goals in two, but I'm sure you could find instances where I miss goal in two in longer games. Even if I went as far as saying that it only counts if the game is at least 45s per move and the goal missed was in a critical position (the miss potentially would affect the outcome of the game), you could probably find one. I think if we practiced tactics more perhaps we could be nearly immune to these mistakes. Perhaps becoming a tactical monster will help me on my quest to become WC.
|
| « Last Edit: May 15th, 2012, 3:23pm by browni3141 » |
IP Logged |
|
|
|
browni3141
Forum Guru
Arimaa player #7014
Gender:
Posts: 384
|
|
Re: Endless Endgame Event?
« Reply #18 on: May 16th, 2012, 5:54pm » |
Quote Modify
|
It looks like briareus just allowed a preventable goal in one in this game on move 45g
http://arimaa.com/arimaa/gameroom/comments.cgi?gid=230281
I guess humans aren't the only ones that need to work on tactics . Of course it's probably just a bug.
|
|
IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: Endless Endgame Event?
« Reply #19 on: May 17th, 2012, 3:42am » |
Quote Modify
|
on May 16th, 2012, 5:54pm, browni3141 wrote:
Thanks, replied in the game comments 
|
|
IP Logged |
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender: 
Posts: 359
|
|
Re: Endless Endgame Event?
« Reply #20 on: May 20th, 2012, 12:48am » |
Quote Modify
|
on May 14th, 2012, 7:50am, hyperpape wrote:
Why not just grab a bunch of interesting positions from actual games? Take a game that had a complicated endgame and back up 5, 10 or 15 moves to reach a point where you're not just looking for goal in two threats, and play from there.
That way, you ensure that the positions are ones that plausibly arise in real games.
There is a worry that it may be hard to find even games, and therefore it might be hard to build a rating system. In my opinion, the rating system is the least interesting part of this exercise, but I suspect others would differ. |
|
I like the idea of using real-game positions for this, but it seems to bring up a number of issues:
-How would positions be selected for addition to the lineup?
-What can be done about positions which display a significant advantage for one side?
-Have there been enough such positions that this method could be implemented without position-specific theory developing?
-How far back from the end of the game should the positions be taken?
|
|
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender:
Posts: 359
|
|
Re: Endless Endgame Event?
« Reply #21 on: May 20th, 2012, 12:51am » |
Quote Modify
|
on May 14th, 2012, 9:27am, mistre wrote:Another idea might be to set up a mob vs bot postal game with reduced material. That way everyone can learn collectively about a bot's strengths and weaknesses in the endgame at a slow pace.
I think this idea will give the most "bang for the buck" without a great deal of setup and customization that a more detailed proposal such as yours would require. |
|
This would indeed require much less setup effort, but I strongly suspect it would also yield much less progress. How long will it take? Can the mob handle two games at once? How likely is the mob to play a risky experiment? Will a bot learn the mob's strategies and try specifically to counter them? How useful will reviewing the game be when the silicon opponent has nothing to add?
|
|
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
hyperpape
Forum Guru
Arimaa player #7113
Gender:
Posts: 80
|
|
Re: Endless Endgame Event?
« Reply #22 on: May 20th, 2012, 7:32pm » |
Quote Modify
|
Quote:| How would positions be selected for addition to the lineup? |
|
Someone looks and chooses. I suppose there is no great way to organize that.
Quote:| What can be done about positions which display a significant advantage for one side? |
|
Try to avoid them. For that reason, we'd want solid players to choose the positions.
Quote:| Have there been enough such positions that this method could be implemented without position-specific theory developing? |
|
There have been a lot of Arimaa games. I'd be surprised if there weren't enough.
Quote:| How far back from the end of the game should the positions be taken? |
|
As far back as they "look like an endgame".
|
| « Last Edit: May 20th, 2012, 7:32pm by hyperpape » |
IP Logged |
|
|
|
mistre
Forum Guru
Gender:
Posts: 553
|
|
Re: Endless Endgame Event?
« Reply #23 on: May 21st, 2012, 9:09am » |
Quote Modify
|
on May 20th, 2012, 12:51am, clyring wrote:
This would indeed require much less setup effort, but I strongly suspect it would also yield much less progress. How long will it take? Can the mob handle two games at once? How likely is the mob to play a risky experiment? Will a bot learn the mob's strategies and try specifically to counter them? How useful will reviewing the game be when the silicon opponent has nothing to add? |
|
All valid concerns. It was just meant as a suggestion and I wouldn't be the one to pursue it anyways as I have plenty of other stuff keeping me busy.
It appears that being able to play Bot_NoMHH in the gameroom is a first step in the direction you want to go.
|
|
IP Logged |
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender:
Posts: 359
|
|
Re: Endless Endgame Event?
« Reply #24 on: May 23rd, 2012, 7:32am » |
Quote Modify
|
on May 20th, 2012, 7:32pm, hyperpape wrote:
Try to avoid them. For that reason, we'd want solid players to choose the positions. |
|
Ah, then you see the problem here: There are not yet any solid players in the endgame.
on May 20th, 2012, 7:32pm, hyperpape wrote:| There have been a lot of Arimaa games. I'd be surprised if there weren't enough. |
|
I wouldn't be surprised if there were, but relatively few games trade down into an almost-even endgame. I'd estimate that at least 100 of them would be needed, and considering how few games trade down evenly, it would also not surprise me if we could not yet find enough using only real-game positions. This is something to consider before starting to scan through the long list of played games, not after.
on May 20th, 2012, 7:32pm, hyperpape wrote:
As far back as they "look like an endgame". |
|
But when does a position "look like an endgame?" I suppose "when neither player can defend without putting themselves at a large disadvantage" would be a criterion, but who is currently capable of judging this accurately?
|
|
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
hyperpape
Forum Guru
Arimaa player #7113
Gender:
Posts: 80
|
|
Re: Endless Endgame Event?
« Reply #25 on: May 23rd, 2012, 2:56pm » |
Quote Modify
|
Quote:| Ah, then you see the problem here: There are not yet any solid players in the endgame. |
|
There is no one who can look at an endgame position and guess whether it's equal with at least some reliability? Bear in mind that the penalty for a false positive is some degree of wasted time, not something disastrous. If your goal is to make an accurate rating system, you need to be terribly fearful of uneven setups. If you just want interesting games that teach about the endgame, then you don't.
Quote:| I wouldn't be surprised if there were, but relatively few games trade down into an almost-even endgame. I'd estimate that at least 100 of them would be needed, and considering how few games trade down evenly, it would also not surprise me if we could not yet find enough using only real-game positions. This is something to consider before starting to scan through the long list of played games, not after. |
|
I don't see why you'd even need 100. Not to start with, at least. And I'm not so sure there aren't games that trade down. I feel like they're pretty common (the games don't have to be expert games, since you're just snapshotting a position). But if there were a script to verify the existence of materially even games, that would be great.
Quote:| But when does a position "look like an endgame?" I suppose "when neither player can defend without putting themselves at a large disadvantage" would be a criterion, but who is currently capable of judging this accurately? |
|
I'm not sure I understand why you'd choose this criterion.
Players and commentators talk about games being in the endgame all the time. This thread depends on us knowing an endgame when we see one--otherwise how would we have noticed that bots are better at them?
Honestly, it seems like this is the strange situation where you're proposing a great idea while also potentially bikeshedding it.
So as not to do my own bikeshedding, let me say that if you insist on random setups, that's not necessarily a disaster. It just has the problem that you won't have setups that are likely to ever arise in real games.
|
|
IP Logged |
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender:
Posts: 359
|
|
Re: Endless Endgame Event?
« Reply #26 on: Jul 11th, 2012, 5:50am » |
Quote Modify
|
on May 13th, 2012, 11:42am, clyring wrote:| I don't want to participate in a "wouldn't it be neato if" discussion, but I suppose I deserve this one for listing it as "the first issue to resolve" when in fact it interferes only with the convenience of the players- as things are, the players could simply suicide pieces en masse to create a position I give them before the game and then begin play, but that would be a great hassle for all involved and something that I would like to avoid if possible. |
|
In a few hours, it will have been 60 days since I first posted this topic. I take this as clear evidence that avoiding this inconvenience will not be possible anytime soon. However, I still have a strong desire to make this event happen. Anyone still interested in participating is free to contact me via PM on the forum before 2359 UTC Monday.
-I don't think I've yet addressed the issue of time controls. If there are no objections, I'll proceed with 2/0/100/0/90t/6.
|
| « Last Edit: Jul 11th, 2012, 5:54am by clyring » |
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
supersamu
Forum Moderator
Forum Guru
Arimaa player #7523
Gender:
Posts: 140
|
|
Re: Endless Endgame Event?
« Reply #28 on: Jul 17th, 2012, 12:48pm » |
Quote Modify
|
I have a proposal: Silver should start with two moves; Then gold does 4 moves.
After this the game progresses as usual. This would minimize the starting advantage of gold.
Since we cannot start a game at this point, we would have a bit of time in reserve as soon as we start the game from this point on. Is this intended?
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: Endless Endgame Event?
« Reply #29 on: Jul 17th, 2012, 5:43pm » |
Quote Modify
|
on Jul 17th, 2012, 12:48pm, supersamu wrote:I have a proposal: Silver should start with two moves; Then gold does 4 moves.
After this the game progresses as usual. This would minimize the starting advantage of gold. |
|
Interesting: Until now I have always seen the reverse proposal, namely Gold starts with two steps and then Silver continues with four. Since the setup is mirrored, the effect is identical. I guess the only reason to choose one over the other is that there is a tradition of having Gold move first.
|
|
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print