Topic: Widelands tournament 2017: subscriptions started!

king_of_nowhere wrote:

... I'm gonna put kaputtnik against hasi. ...

I guess that's why he is called kaputtnik...

Also, considering both tournaments had people quit, next time the limit will be 14 players, not 12.

I don't understand why there needs to be a player limit of twelve people. Okay, maybe it is not necessary to know why. But I'd say that 13 could also replace 14...

nemesis26 wrote:

king_of_nowhere wrote:

Also, considering both tournaments had people quit, next time the limit will be 14 players, not 12.

or change the system. Make two groups (with 6 players or more), in which every player has to play against every other player in his group. 3 months time or something like that. so its no problem if one player has no time in September, he can just play in the other two months! And then KO-system....or new groups (one group for the ranks 1-6 and one with 7-12....or three groups?) to get a final ranking

That sounds like a lot of matches and a lot of time... Maybe too much... I think that the swiss system with king_of_nowheres algorithm is nice...

einstein13 wrote:

WorldSavior won. I lost. We were both playing Empire tribe.

That was really exciting match.

Oh yes, it was. You gained more land in the beginning and increased the pressure by that. During the match I wondered if I would loose it...

For sure he is better player than me!

Thanks. But maybe I'm not, and I just won, because you chatted much more than me?

Wanted to save the world, then I got widetracked

Top

Quote

kaputtnik wrote:

Me and hasi50 try to play tomorrow as the day progresses. This will be a revenge on our first tournament match then

damn. that should not be possible, as the pairing system forbids players to go against each other more than once.

On the other hand, everyone else already played, and there would be two forfaits otherwise. So I'm letting this go on account of two wrongs making one right. I also must take this as a reminder to actually check for that kind of stuff.

Top

Quote

WorldSavior wrote:

Well, actually it can happen that two players play more than once against each other - if there are enough rounds (happens usually to players with low score)

the ruling of the swiss system specifically forbids it - to be more accurate, there are many possible variations of the swiss system, generally dealing with how you resolve tie breaks and how you pair players with equal score. But all the variations I know forbids the same players from playing more than once, and certainly the variation I'm using here forbids it.

Because of that, if there are many rounds and few players there is a possibility that aturn cannot be made without breaking this rule. I don't know the mathematical demonstration - seems the kind that can be worked out, but would require a lot of calculations - but I know that, to be sure that you can make a turn, the number of players must be at least twice the number of turns. So for 6 turns we must be at least 12 to have the certainty of not breaking the system.

You assumed I picked 12 as minimum number arbitrarily, didn't you?

Also, if there are too many players and too few turns, there is the risk that more than one player will be first with full score, i.e. that two players will win all games and never face each other. In the case of no draws the number of players at full score halves at every round, rounded up (if there are 3 players at full score, 2 will face each other and the third will pick someone with lower score, so next turn there may still be 2 players at full score). So to have a meaningful resolution a tournament should also have no more players than 2^[number of turns].

Therefore, a good number of turns for a swiss tournament is lower than [number of players] / 2 and high enough that 2^[number of turns] greater or equal [number of players]. It's generally better if the value is in the middle of those extremes. Too little turns, and there will be many people at the top of the ranking who never played with each other. Too many turns, and the last few turns those at the top and bottom will have played with everyone of similar score and will have to pick players with very different score.

With 5 turns, you can have at most 32 players, that's why I preferred 6. Although for 12 players, you have a minimum of 4 rounds and a maximum of 6, so 5 would be better.

P.S. The standard formula for weekend chess tournaments is 5 turns (about 4 hours per game, one game friday afternoon, two games saturday, two games sunday). Most of those tournaments have more than 32 players. Luckily, draws are common in chess, especially at high levels (grandmasters draw almost half the times. Players of my level draw about one game in four. Of course individual gameplay makes a difference. That's because if nobody makes a mistake, the game is draw, and strong players make very few mistakes; also, in the case of the small mistakes of strong players, punishing the mistake is so hard that it would take a player of greater skill to do so), so I've never heard of people ending together at full score.

Top

Quote

Hello everybody,

had no time to look at the replays by now. I uploaded all of them to https://www.magentacloud.de/share/tu4ayusx.k

as of the crashes we had they are a mix of client and host, hope we can make something out of it.

Here are the widelands where people may dwell, walking around care that evrythings well.

Top

Quote

I updated the official thread

https://wl.widelands.org/forum/topic/2912/ with the ranking after the second round.

To make the pairing, I wait a confirmation from gunchleoc:

If she can play, I put her in, otherwise I make a pairing without her directly. there is time until the end of october, so we're not in a hurry anyway and I can wait a few days. Please guncchleoc confirm your participation or lack thereof

I am still waiting for mars-trimard replay

Top

Quote