Author |
Topic: 2008 World Championship Format (Read 7337 times) |
|
chessandgo
Forum Guru
Arimaa player #1889
Gender: 
Posts: 1244
|
 |
Re: 2008 World Championship Format
« Reply #15 on: Oct 4th, 2007, 10:32am » |
 Quote  Modify
|
on Oct 4th, 2007, 3:35am, camelback wrote:May be we can have previous year winner and runner-up given free entry to championship round and all interested players compete in qualifying round.
Just like FIFA World Cup winner and Host is given free entry in next world cup.
This way everybody except chessandgo and Fritzlein can have fun in the preliminary round  |
|
The runner-up is not always the second-place finisher, like this year ... if we are to attribute 2 seats, it would be more logical to give them to winner and second-place finisher, wouldn't it ?
|
|
 IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2008 World Championship Format
« Reply #16 on: Oct 4th, 2007, 11:06am » |
Quote Modify
|
Hey, it's great to see all this discussion and all the great ideas out there. We might not have started talking about the format soon enough this year. Perhaps we should start another thread for "2009 World Championship Format".
Here's a different hypothetical question. If it is too late for 2008 to have a two-stage system with a qualifier tournament and final tournament, would people prefer the unified tournament to be floating double elimination, or floating triple elimination?
A 24-player double-elimination takes seven to nine rounds, usually eight.
A 24-player triple-elimination takes nine to twelve rounds, usually ten or eleven.
Is the extra accuracy worth having the tournament be, on average, two and a half rounds longer? My answer is of course that two and a half more rounds is an advantage: more Arimaa is more better. Still, I understand if some people think it would just be too much if you had to commit to a serious game every week for two and a half months to become World Champion.
I think the idea of a longer tournament was shot down before, but if we are seriously discussing having something like five rounds of qualifying plus eight rounds of finals (or in that ballpark) then a single ten-or-eleven round tournament doesn't sound too bad.
Furthermore, one of the ideas behind a two-tier systems was intentionally to give top players a break, so they didn't have to win as many games to become champion. If we are going to level the field and make everyone qualify anyway, or everyone except last year's champion, then this motivation is reversed: in that case a single triple-elimination tournament is less hassle.
|
| « Last Edit: Oct 4th, 2007, 11:16am by Fritzlein » |
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2008 World Championship Format
« Reply #17 on: Oct 4th, 2007, 11:08am » |
Quote Modify
|
on Oct 4th, 2007, 10:32am, chessandgo wrote:| if we are to attribute 2 seats, it would be more logical to give them to winner and second-place finisher, wouldn't it ? |
|
Yes, if we qualify two based on last year's World Championship, it should be chessandgo and PMertens. However, you have to account for the fact that PMeterns likes live play against humans so much he would beg to play in the qualifying rounds.
|
|
IP Logged |
|
|
|
The_Jeh
Forum Guru
Arimaa player #634
Gender:
Posts: 460
|
|
Re: 2008 World Championship Format
« Reply #18 on: Oct 4th, 2007, 12:34pm » |
Quote Modify
|
This doesn't have much to do with the World Championship tournament, but why can't we have casual tournaments throughout the year, perhaps one per season? Each of these tournaments could be strict single-elimination and take place over the course of one day. I realize we come from diverse time zones, but because they would be casual I think it would be all right.
And then if you're concerned over the manipulation of p8 ratings and such, you could use p8 ratings based on tournament games only - previous WCT, PT, and casual tournaments. Or you could use tournament W-L-D record.
|
| « Last Edit: Oct 4th, 2007, 12:40pm by The_Jeh » |
IP Logged |
|
|
|
chessandgo
Forum Guru
Arimaa player #1889
Gender:
Posts: 1244
|
|
Re: 2008 World Championship Format
« Reply #19 on: Oct 4th, 2007, 1:10pm » |
Quote Modify
|
on Oct 4th, 2007, 11:06am, Fritzlein wrote:
I think the idea of a longer tournament was shot down before, but if we are seriously discussing having something like five rounds of qualifying plus eight rounds of finals (or in that ballpark) then a single ten-or-eleven round tournament doesn't sound too bad.
|
|
I agree with this, Karl.
|
|
IP Logged |
|
|
|
chessandgo
Forum Guru
Arimaa player #1889
Gender:
Posts: 1244
|
|
Re: 2008 World Championship Format
« Reply #20 on: Oct 4th, 2007, 1:12pm » |
Quote Modify
|
on Oct 4th, 2007, 11:08am, Fritzlein wrote:
Yes, if we qualify two based on last year's World Championship, it should be chessandgo and PMertens. However, you have to account for the fact that PMeterns likes live play against humans so much he would beg to play in the qualifying rounds. |
|
Well, anyway such direct qualification could only take place for next year, if it is included in the rules of this year's WC.
|
|
IP Logged |
|
|
|
Janzert
Forum Guru
Arimaa player #247
Gender:
Posts: 1016
|
|
Re: 2008 World Championship Format
« Reply #21 on: Oct 4th, 2007, 1:41pm » |
Quote Modify
|
I had read an interesting article about a study done on different professional sports playoff structures. The things that struck me at the time were; First, across all the sports looked at the top team only had about a 30% chance of winning the championship. This was across baseball, football, hockey and some others I believe. So they had fairly widely varying playoff structures and chances of an upset for any one game. So it would seem at least from popularizing a sport to spectators absolute accuracy in picking the true top performer is not necessarily advantageous. I have feeling this may carry over at least somewhat to popularizing it to participants as well.
Second, a way suggested by the researchers to improve the accuracy was by adding some preliminary rounds. Although the article wasn't particularly clear on what form these took and I have yet to get the time to read the actual studies.
The take away I had from it was that if the ultimate goal with the world championship is to help increase the popularity of Arimaa maybe we shouldn't necessarily be striving for absolute accuracy in the tournament. And if we do want to increase the accuracy a preliminary tournament(s) may be a good way of doing so.
Janzert
P.S: Looks like the original article is here and the actual studies are here.
|
|
IP Logged |
|
|
|
jdb
Forum Guru
Arimaa player #214
Gender:
Posts: 682
|
|
Re: 2008 World Championship Format
« Reply #22 on: Oct 4th, 2007, 3:37pm » |
Quote Modify
|
Having a five round swiss as a qualifier would be a great way to encourage participation.
Under the current system, the lowest rated player has a 99% (or worse) chance to be out in two games. Who would want to join with those odds?
With a five round swiss, anyone has a half decent chance of making the top six to qualify for the final tournament. If they finish with three wins they likely have a chance at 6th. It would seem to me to be a much more inviting format to encourage lower ranked players to participate.
|
|
IP Logged |
|
|
|
PMertens
Forum Guru
Arimaa player #692
Gender:
Posts: 437
|
|
Re: 2008 World Championship Format
« Reply #23 on: Oct 4th, 2007, 8:39pm » |
Quote Modify
|
on Oct 4th, 2007, 11:08am, Fritzlein wrote:
Yes, if we qualify two based on last year's World Championship, it should be chessandgo and PMertens. However, you have to account for the fact that PMeterns likes live play against humans so much he would beg to play in the qualifying rounds. |
|
... and I totally dislike the idea of having any advantage at all for "older" players over newcomers or better players over worse.
I opposed all benefits for high-ranked players last year and I still do.
We do not need a tournament to find the best player anyway ... just let Jean and Karl slug it out 
The tournament will merely show us the luckiest player in those few specific games.
No need to further improve the chances of anyone by any rules be it folding or otherwise.
And on some totally different topic: before I beg to play in some qualifiying I will first have to consider wether to participate at all :-P
Unlike the last few year I currently see no fresh blood that could compete with the top (statement not based on ratings), so it might be a bit boring in the end ... especially if the first few rounds are only for reducing the playercount to the "significant" players.
|
| « Last Edit: Oct 4th, 2007, 8:50pm by PMertens » |
IP Logged |
|
|
|
99of9
Forum Guru
Gnobby's creator (player #314)
 
Gender:
Posts: 1413
|
|
Re: 2008 World Championship Format
« Reply #24 on: Oct 4th, 2007, 9:18pm » |
Quote Modify
|
on Oct 4th, 2007, 11:06am, Fritzlein wrote:| I think the idea of a longer tournament was shot down before, but if we are seriously discussing having something like five rounds of qualifying plus eight rounds of finals (or in that ballpark) then a single ten-or-eleven round tournament doesn't sound too bad. |
|
One benefit of a separation between qualifying and tournament is that you can play at fast time control in qualifying, but any time-rushed losses would not carry through to count for the main tourney.
My personal favourite is roughly 5 rounds of open qualifying swiss, followed by a double-elimination between the top 8 (which is reasonably short for most people), seeds also determined by the swiss.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2008 World Championship Format
« Reply #25 on: Oct 4th, 2007, 9:32pm » |
Quote Modify
|
on Oct 4th, 2007, 9:18pm, 99of9 wrote:| followed by a double-elimination between the top 8 |
|
Then the final would be six more rounds, or five if someone wins undefeated, for 10-11 rounds total, just like a single, open triple elimination. But maybe having the five-round qualifier first is friendlier, allowing everyone to get in some serious games while providing a fair basis for reducing the field to eight and seeding that field on something other than ratings.
|
|
IP Logged |
|
|
|
99of9
Forum Guru
Gnobby's creator (player #314)
Gender:
Posts: 1413
|
|
Re: 2008 World Championship Format
« Reply #26 on: Oct 4th, 2007, 10:02pm » |
Quote Modify
|
The first 5 rounds would be at a short time control so don't require as major time commitments. Are you suggesting the same for the FTE?
I'm also still happy with a straightforward open double elimination, but I am getting more wary of the effects of seeding.
|
|
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: 2008 World Championship Format
« Reply #27 on: Oct 5th, 2007, 2:55am » |
Quote Modify
|
on Oct 4th, 2007, 12:34pm, The_Jeh wrote:This doesn't have much to do with the World Championship tournament, but why can't we have casual tournaments throughout the year, perhaps one per season? Each of these tournaments could be strict single-elimination and take place over the course of one day. I realize we come from diverse time zones, but because they would be casual I think it would be all right.
And then if you're concerned over the manipulation of p8 ratings and such, you could use p8 ratings based on tournament games only - previous WCT, PT, and casual tournaments. Or you could use tournament W-L-D record. |
|
Yes, I think the time zone differences is what makes this difficult. But we've never tried it before, so we don't know for sure.
|
|
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: 2008 World Championship Format
« Reply #28 on: Oct 5th, 2007, 4:17am » |
Quote Modify
|
on Oct 4th, 2007, 9:18pm, 99of9 wrote:
One benefit of a separation between qualifying and tournament is that you can play at fast time control in qualifying, but any time-rushed losses would not carry through to count for the main tourney.
My personal favourite is roughly 5 rounds of open qualifying swiss, followed by a double-elimination between the top 8 (which is reasonably short for most people), seeds also determined by the swiss. |
|
My favorite is still the swissKnife format 
http://arimaa.com/arimaa/forum/cgi/YaBB.cgi?board=talk;action=display;nu m=1114794077;start=84#84
What we do know from experiments we've done in the past is that if you have an accurate rating system you can just use the rating to pick the best player and it beats every other format by a long shot. But in the absence of an accurate rating system the Floating Elimination formats provide the best alternative. However its main drawback is that it doesn't provide many games for the weaker players.
Our present rating system is far from accurate at predicting ones performance against the field and I would never consider using that for the original proposal of this thread. I feel the p8H or some variant of it will be much better, but the feedback from the community has reminded me that it still needs to be tried out and proven before being used. So the original proposal of this thread is perhaps a bit premature given the state of our rating system. Thus, I think it makes a lot of sense to have all eight of the contestants in the final WC tournament determined by the preliminary swiss tournament.
|
|
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: 2008 World Championship Format
« Reply #29 on: Oct 5th, 2007, 5:19am » |
Quote Modify
|
So I think the basic format for this year will be a swiss preliminary to select 8 players for the final Floating Elimination tournament. There are still a lot of details that need to be sorted out.
One thing I would like to determine is if we should use the double elimination or the triple elimination format. Now that we know there will be 8 players I ran some simulations to see how the two formats compare. Here are the numbers:
run2 'formats/floatTripElim' 2000 8 500 200 999999
1 51.0%
2 25.1%
3 13.8%
4 6.3%
5 2.2%
6 1.2%
7 0.4%
8 0.1%
average number of rounds = 9.05
average rating from best = 36.3
run2 'formats/floatDoubleElim' 2000 8 500 200 999999
1 42.5%
2 32.5%
3 12.6%
4 8.8%
5 2.3%
6 0.8%
7 0.3%
8 0.2%
average number of rounds = 5.65
average rating from best = 43.1
Each simulation was done with 2000 tournaments with 8 players having a true rating distribution of 500 points and measured ratings having an inaccuracy of 200 and chance of draw games 1 in 999999. The simulation programs can be downloaded from here:
http://arimaa.com/arimaa/tourn/compare/sim.tar
or in ZIP format:
http://arimaa.com/arimaa/tourn/compare/sim.zip
The FTE picked the player with the highest true rating 51% of the time and the FDE picked the player with the highest true rating 42.5% of the times. But to gain that 9% it adds about 3.5 rounds (almost one month with one round per week) to the schedule.
Though I was originally thinking to use the FTE since the number of players is less. After looking at these numbers, FDE still looks tempting. Would like to know what the players who plan to participate think.
|
| « Last Edit: Oct 5th, 2007, 5:20am by omar » |
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print