Author |
Topic: Haizhi's unconnected-step pruning (my notes) (Read 4680 times) |
|
aaaa
Forum Guru
Arimaa player #958
Posts: 768
|
|
Re: Haizhi's unconnected-step pruning (my notes)
« Reply #15 on: Feb 26th, 2009, 2:48pm » |
Quote Modify
|
on Feb 26th, 2009, 11:03am, Hannoskaj wrote:By the way, I have also played at making a position where the only winning move is making four uncorrelated steps. Here silver goals in two (ha6 mc1 ec7 cg7), but any other move is a loss in two, if I am not mistaken (typically Db2 to defend (or Cb4 if free) and Db6 Ra6). Silver: r d8 e7 d6 a5; c h7; d b6; h a7 e5; m d1; e b7 Gold: R d5 d7 g6; C b5; D b1 g4; E e6 |
| Bot-assisted analysis bears out that you have succeeded in your goal. Minor corrections are that ha7s md1w eb7e ch7w is actually a goal in three (since rd6w Ee6w Rd5w Cb5s delays the goal for one more move) and that there is one other move that is not a loss in two, eb7e ec7n Rd7w ch7w, but this still loses in three. Quote:By the way, is there any agreeable way to post a position ? |
| See http://arimaa.com/arimaa/learn/notation.html, under "Notation for Recording Arimaa Positions".
|
|
IP Logged |
|
|
|
Janzert
Forum Guru
Arimaa player #247
Gender:
Posts: 1016
|
|
Re: Haizhi's unconnected-step pruning (my notes)
« Reply #16 on: Feb 27th, 2009, 4:32am » |
Quote Modify
|
Hannoskaj It seems your first position has a related first and second step. For your first position to move Re3n: +-----------------+ 8| . . . . . . . . | 7| . . . . . . . . | 6| . . x . . x . . | 5| . . . . . . . . | 4| . . . . C e . . | 3| . . x d R R . . | 2| . . . . . . . . | 1| . . . . R . . . | +-----------------+ a b c d e f g h Re1n Re2e Ce4w Re3n But maybe I'm misunderstanding it as I did the goal example initially. Janzert
|
« Last Edit: Feb 27th, 2009, 4:45am by Janzert » |
IP Logged |
|
|
|
Hannoskaj
Forum Guru
Arimaa player #3794
Gender:
Posts: 75
|
|
Re: Haizhi's unconnected-step pruning (my notes)
« Reply #17 on: Feb 27th, 2009, 5:46am » |
Quote Modify
|
on Feb 27th, 2009, 4:32am, Janzert wrote: It seems your first position has a related first and second step. For your first position to move Re3n: +-----------------+ 8| . . . . . . . . | 7| . . . . . . . . | 6| . . x . . x . . | 5| . . . . . . . . | 4| . . . . C e . . | 3| . . x d R R . . | 2| . . . . . . . . | 1| . . . . R . . . | +-----------------+ a b c d e f g h Re1n Re2e Ce4w Re3n |
| I had thought about keeping the cat from protecting the f3 rabbit, but had forgot about the e1 rabbit. Here is a corrected version with Re1n Rf3e Ce4(wn) Re3n as only solution: +-----------------+ 8| . . . . . . . . | 7| . . . . . . . . | 6| . . x . . x . . | 5| . . . . . . . . | 4| . . . . C e . . | 3| . . x d R R . . | 2| . . . . . c . . | 1| . . . . R . . . | +-----------------+ a b c d e f g h Jonas
|
|
IP Logged |
|
|
|
Hannoskaj
Forum Guru
Arimaa player #3794
Gender:
Posts: 75
|
|
Re: Haizhi's unconnected-step pruning (my notes)
« Reply #18 on: Feb 27th, 2009, 5:55am » |
Quote Modify
|
on Feb 26th, 2009, 2:48pm, aaaa wrote: Bot-assisted analysis bears out that you have succeeded in your goal. Minor corrections are that ha7s md1w eb7e ch7w is actually a goal in three (since rd6w Ee6w Rd5w Cb5s delays the goal for one more move) and that there is one other move that is not a loss in two, eb7e ec7n Rd7w ch7w, but this still loses in three. See http://arimaa.com/arimaa/learn/notation.html, under "Notation for Recording Arimaa Positions". |
| Thank you for the correction. I must have seen the defence by Gold with loss in three but did not pay attention since I also knew it went through (just forgot to update length). It would certainly be more aesthetic if there was a way to get rid of it, though. On the other hand, I had completely missed that the cat b5 could be freed and defend for a turn. In fact I thought I had built everything to make that impossible... Anyway, moving db6 to dc5 in the initial position seems to kill that flaw. Finally, thank you for the link. I do not call that "agreeable", but with a text file as skeleton, it's at least usable. Jonas
|
|
IP Logged |
|
|
|
Hannoskaj
Forum Guru
Arimaa player #3794
Gender:
Posts: 75
|
|
Re: Haizhi's unconnected-step pruning (my notes)
« Reply #19 on: Jun 1st, 2009, 1:33am » |
Quote Modify
|
I just had an idea to define unconnected steps, that would avoid making a complete list of tests such as the one at the beginning of the topic : Four steps are unconnected if all (24) permutations are still legal and yield the same final position. This does not cover the case: * support a trap that the second piece to move was supporting if there is no piece on the trap to die during the move. Jonas
|
|
IP Logged |
|
|
|
|