file Optimal seatings changes

13 Nov 2014 10:33 #67386 by Ankha
Replied by Ankha on topic Re: Optimal seatings changes

You are in your hunt for the best seating for 2R+F tournament.

But I was speaking of the seatings for 3R+F tournament where for 5 players tournaments, we have player 2 who play 2nd then 1st and 1st.

Without looking for extreme cases like 5 players tournaments, in 10 players tournaments, players 2 and 7 play twice in the second place and 3 and 8 twice in the 4th place.

You're speaking about the current seatings. It doesn't mean there isn't a better seating for which rule 7 applies in full.
We must first search better seatings that enforce the current rules before thinking of changing them.

Prince of Paris, France
Ratings Coordinator, Rules Director

Please Log in or Create an account to join the conversation.

More
13 Nov 2014 10:43 - 13 Nov 2014 10:45 #67387 by Timo
Replied by Timo on topic Re: Optimal seatings changes
Why that ?

If the rules are to be changed, I guess it is for the best to do it BEFORE looking for better seatings.

I mean : if after review, one decides to change the rules (I am no sure whose role it is to change them), it would be logical to check afterward the current seatings to look for better ones...

What I mean is that the rules which FORBID to have a player playing 5th in 2 different rounds could potentially prevent a better VP/transfer repartition while having no good effect.

(I state here what is my opinion that forbiding the repetition of the 5th place while trying to share evenly the transfer and to avoid if possible the repetition of any place is redundant while potentially bad)
Last edit: 13 Nov 2014 10:45 by Timo.

Please Log in or Create an account to join the conversation.

More
13 Nov 2014 13:01 #67389 by Ankha
Replied by Ankha on topic Re: Optimal seatings changes

Why that ?

What I mean is that the rules which FORBID to have a player playing 5th in 2 different rounds could potentially prevent a better VP/transfer repartition while having no good effect.

Because the more constraints we have, the less possibilities we have. And since possibilities are exponentials, having the tighest constraints is essential to achieve computation.
The rule "COULD potentially prevent" (emphasis mine) some better results. But maybe it doesn't. On 2R it doesn't change anything, I haven't studied 3R yet.
If you have some spare time, try to find one seating where the rule prevents a better seating. If it does, okay we may think about dropping it. If it doesn't, it's a easy way to reduce the number of possible seatings.

Prince of Paris, France
Ratings Coordinator, Rules Director

Please Log in or Create an account to join the conversation.

More
13 Nov 2014 14:34 #67390 by Timo
Replied by Timo on topic Re: Optimal seatings changes
It is only a constraint of computing time ?

So you will gain some time when checking your conditions and so deleting this rule which is redundant with the rule 7 every time when you have reach a number of player big enough to avoid infringing rule 7 (always for 2R+F as you have pointed out).

And for the few cases where you have to add some odd seatings which would have been already eliminated, I guess the additional compuing time is meaningless compare to the size of the data no ?

Please Log in or Create an account to join the conversation.

More
13 Nov 2014 14:54 #67391 by Ankha
Replied by Ankha on topic Re: Optimal seatings changes

It is only a constraint of computing time ?

So you will gain some time when checking your conditions and so deleting this rule which is redundant with the rule 7 every time when you have reach a number of player big enough to avoid infringing rule 7 (always for 2R+F as you have pointed out).

And for the few cases where you have to add some odd seatings which would have been already eliminated, I guess the additional compuing time is meaningless compare to the size of the data no ?

If you inforce strictly rule 7 (7. A player doesn't play in the same seat position, if possible.), it cuts branches in the tree of possibilities.

For instance, when I must seat player 4 of a 10-players table, it removes 2 possibilities for 4. Let's say I have 5,6,7,8,9 and 10 to seat afterwards, I'm saving 2 * (5 * 4 * 3 * 2 * 1) possibilities (because there would be 5 possibilities for 5, 4 possibilities for 6, 3 for 7 etc.). A little less in truth because 5, 6, 7... also have constraints, but it's really a huge save.

Prince of Paris, France
Ratings Coordinator, Rules Director

Please Log in or Create an account to join the conversation.

More
23 Feb 2015 12:37 #69400 by Ankha
Replied by Ankha on topic Re: Optimal seatings changes
I'm done with 2R+F computations for up to 31 players. Some computations couldn't have been done exhaustively (the more 5-players tables there is, the more exponential it becomes).

I'm starting to work on 3R+F.

Prince of Paris, France
Ratings Coordinator, Rules Director
The following user(s) said Thank You: Lönkka, self biased

Please Log in or Create an account to join the conversation.

More
Moderators: AnkhaKraus
Time to create page: 0.072 seconds
Powered by Kunena Forum