Author |
Topic: Material evaluation question (Read 3473 times) |
|
woh
Forum Guru
Arimaa player #2128
Gender: 
Posts: 254
|
 |
Re: Material evaluation question
« Reply #15 on: Apr 11th, 2007, 6:02am » |
 Quote  Modify
|
on Apr 10th, 2007, 11:04pm, aaaa wrote:| This is with both sides having at least some pieces. |
|
Yes, I only considered combinations with both players having at least one piece in those counts.
|
|
 IP Logged |
|
|
|
aaaa
Forum Guru
Arimaa player #958
Posts: 768
|
|
Re: Material evaluation question
« Reply #16 on: Apr 12th, 2007, 12:38am » |
Quote Modify
|
I've managed to come up with a collapse function that outputs a canonical combination that's in the same form as the input.
|
|
IP Logged |
|
|
|
pago
Forum Guru
Arimaa player #5439
Gender:
Posts: 69
|
|
Re: Material evaluation question
« Reply #17 on: Sep 16th, 2010, 10:50am » |
Quote Modify
|
Quote:| can someone give an example of sets such that they exhibit a cyclic relationship in the sense of preferring to have one combination with the opponent having the other, e.g. "I'd rather have pieces A versus B and B versus C, but not A versus C"? |
|
The most plausible cycle I have found at this time with the third version of my evaluator in development is:
EMHC8R > EHDD8R > EDDCC8R > EMHC8R
EDDCC8R > EMHC8R is the most controversial inequality although in that case M worth only H
|
|
IP Logged |
|
|
|
99of9
Forum Guru
Gnobby's creator (player #314)
 
Gender:
Posts: 1413
|
|
Re: Material evaluation question
« Reply #18 on: Sep 16th, 2010, 8:47pm » |
Quote Modify
|
on Sep 16th, 2010, 10:50am, pago wrote:
EDDCC8R > EMHC8R is the most controversial inequality although in that case M worth only H
|
|
You're right, that's highly controversial. I definitely disagree with this, and would prefer to play with the MH=HH than the DDC.
|
|
IP Logged |
|
|
|
pago
Forum Guru
Arimaa player #5439
Gender:
Posts: 69
|
|
Re: Material evaluation question
« Reply #19 on: Sep 17th, 2010, 2:31am » |
Quote Modify
|
Quote:You're right, that's highly controversial. I definitely disagree with this, and would prefer to play with the MH=HH than the DDC.
|
|
The controversial effect is a consequence of the relative value given to the pieces by my evaluator.
I can propose the following cycles that use relative values closer to the current consensus :
EMC8R>EHCC8R>EDDC8R>EMC8R
Or
EMDC8R>EHDCC8R > EDDCC8R > EMDC8R
They use the following relative values between pieces :
M>HC>DD > H
M>HC>DC > H
Nota : If one thinks that the relative value of M differs, it should not be difficult to build an other cycle...
|
|
IP Logged |
|
|
|
aaaa
Forum Guru
Arimaa player #958
Posts: 768
|
|
Re: Material evaluation question
« Reply #20 on: Sep 18th, 2010, 11:35am » |
Quote Modify
|
Here is what several evaluators consider a cycle, even without any rabbit ever being off: EMC8R>EHH8R>EDCC8R>EMC8R
|
|
IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: Material evaluation question
« Reply #21 on: Sep 18th, 2010, 11:56am » |
Quote Modify
|
on Sep 18th, 2010, 11:35am, aaaa wrote:| Here is what several evaluators consider a cycle, even without any rabbit ever being off: EMC8R>EHH8R>EDCC8R>EMC8R |
|
Wow, that one is a good find, especially since the differences between the successive comparisons are not very small either. If you sum them up you would think that EMC8R is quite a lot better than EMC8R.
|
|
IP Logged |
|
|
|
pago
Forum Guru
Arimaa player #5439
Gender:
Posts: 69
|
|
Re: Material evaluation question
« Reply #22 on: Sep 19th, 2010, 6:28am » |
Quote Modify
|
Quote:| Here is what several evaluators consider a cycle, even without any rabbit ever being off: EMC8R>EHH8R>EDCC8R>EMC8R |
|
I also like this one.
I wonder whether someone could exhibit a cycle with rabbits that would not be controversial
|
|
IP Logged |
|
|
|
pago
Forum Guru
Arimaa player #5439
Gender:
Posts: 69
|
|
Re: Material evaluation question
« Reply #23 on: Sep 28th, 2010, 9:17am » |
Quote Modify
|
I propose the following cycle with one additional step (4 setups) :
EMDC8R > EHHC8R > EHDD8R > EDDCC8R > EMDC8R
I have not found a cycle of 5 steps with 8 rabbits. I assume that they could be possible with an unbalanced number of rabbits but they would be more controversial because of the difficulty to evaluate the relative value of rabbits.
|
|
IP Logged |
|
|
|
pago
Forum Guru
Arimaa player #5439
Gender:
Posts: 69
|
|
Re: Material evaluation question
« Reply #24 on: Sep 30th, 2010, 5:13am » |
Quote Modify
|
Hereafter are two plausible cycles with unbalanced number of rabbits.
They are almost confirmed by the test performed by jdb with clueless (more precisely there is no obvious contradiction with jdb's results).
They are not as trivial as cycles with equal number of rabbit. So I would be curious to know if someone could explain how they work (assuming that they do exist !) and if "old" players would agree or disagree with them.
1) EDD7R > EDDCC4R > ECC8R > EDD7R
jdb’s results :
EDD7R / EDDCC4R : +3 / -2
EDDCC4R / ECC8R : +3 / -1
ECC8R / EDD7R : +2 / -2
2) ED8R > EDD6R > EDDCC3R > ED8R
jdb’s results :
ED8R / EDD6R : +2 / -2
EDD6R / EDDCC3R : +4 / -1
EDDCC3R / ED8R : +3 / -1
|
|
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print