Tournament No.2 - construction thread

[Cisz]

Before we can set a date, we should have all basics decided and have an overall idea of what is going to happen.

I see it this way:

Step one - choose a ruleset / tournament mode and write that down.

Step two - prepare advertisement.

Step three - set a date like 3 weeks later.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

The rules so far:

We will have some sort of ffa swiss or partial ffa round robin as preliminaries ans some sort of ffa single elimination as finals. The whole thing is still planned for one day, so we cant get higher than 5 rounds max, rather 4 rounds.

Winning condition will be last one standing, game mode vh sr this time, although it's very likely the third tournament will be vhx sr, to avoid the problematic ronald elimination.

Maybe we could do the finals (semifinal and on) in vhx?

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

The advertisement:

I'm looking for sites with a tournament calendar or similiar news section to announce the tournament once we have finished the formal stuff.

Battle.net has a tournament announcement system, but it has high demands on what we have in our ruleset. Maybe to much trouble.

And inWarcraft.de has a tournament calendar, but thats german only.

Any other sites? This can't be all there is.

[Cisz]

We should ask the better players if they have time on a proposed date. Apart from the members of this group, that would be hyper, good, cyberseer.

Did I forget someone?

[ocmassreo]

I would be there if no problems

I agree with the rules that you put

the eliminating vhsr and final vhx sr may be

the standing last one it is OK for eliminatoires (more ronald's large number ?)

for the advertising i can post on french forum if u want and say tio come here for more details

[Cisz]

About the ruleset:

We should target more players this time. 32 players should be a good maximum, but if 8 players show up, thats enough for me not to cancel the event.

The best size for a match would be 4 players imo. 2-6 would be ok too (though 6 is really tough). And games of 7 or 8 players are to big to be allowed imo.

That leads to the question of how many players are advancing from a game. I'd say that in games of 4 or less players, only one should advance, and in games of 5 or 6, we could allow two players to get to the next round.

The preliminaries will be two rounds of ffa's for all players, so everyone gets at least to play this two games.

How do we determine who plays who in the second round of the preliminaries? We shouldn't sort them into games of same first result (swiss), as this will be unfair to players with one win. For a swiss to work there should be at least 3 rounds, and we don't really have time for that.

In any case, we will face problems with players with one win after the first two rounds, as we can't really find out which of them are better. So it makes sense to not only award points for the first player of a preliminary match, but also for the second (like 3 for 1st and 1 for 2nd or something similiar).

The finals will be two rounds of ffa single elimination. The winning players (one or two from each semifinal) advance to the finals. As the final can't be bigger than 6 players, there can't be more than 6 matches in the semifinal.

[Twilice]

What version?

3.0

or shall we wait for a public 4.0? (public beta)

[jolin012]

3.0

[ocmassreo]

think 3.0 is better

and make advertising for the futur public beta after

[jolin012]

to the rules we must decide if they should be played vhx sr. if not yet decided, i vote for vhx sr in finals. vh sr has too big disavdantages(ronald problem)

[Cisz]

So far we have two mayor things to develop:

How is the second round (round 2 of the prelims) arranged? I'd say, in a way that no two players meet again and otherwise random. Or rank as in swiss maybe? Jolin012 had a different idea, something like "sorted in a way that each game has one of each rank".

And which players advance to the finals? What do we do if after two rounds some players near the last rank that advances are ranked identically?

Here an example tournament, 16 players, first two rounds:

Round 1                         Round 2 

Player 1        1               Player 1        1
Player 2        4               Player 5        4
Player 3        3               Player 9        2
Player 4        2               Player 13       4

Player 5        4               Player 2        3
Player 6        1               Player 6        1
Player 7        3               Player 10       2
Player 8        2               Player 14       4

Player 9        1               Player 3        1
Player 10       4               Player 7        3
Player 11       3               Player 11       4
Player 12       2               Player 15       2

Player 13       4               Player 4        4
Player 14       3               Player 8        3
Player 15       2               Player 12       2
Player 16       1               Player 16       1

And we have the results listed here:

Player Name     Results         Rank

Player 16       1 1             1       Advances
Player 6        1 1             1       Advances
Player 1        1 1             1       Advances
Player 9        1 2             4       Advances
Player 3        1 3             5       ????
Player 12       2 2             5       ????
Player 15       2 2             5       ????
Player 8        2 3             8
Player 4        2 4             9
Player 10       2 4             9
Player 7        3 3             9
Player 14       3 4             12
Player 2        3 4             12
Player 11       3 4             12
Player 5        4 4             15
Player 13       4 4             15

I simply add the placement of each game and sort by that sum.

Since several players have the same result, they have the same rank ofc. Now how to decide which of those is better? Not all of them have played each other.

For example, if we allow an even number of players into the filnals, but less than 50%, we want 6 players to advance. But the players 3, 12 and 15 are ranked identically. Only two of them have met so far (3 and 15 in round 2).

I tried to add up the ranks of all players they won against, but again I get a tie between two of them: Player 3 and 12 beat players with a combined rank of 38 ranks. Player 15 did worse (he beat players with 48 combined ranks), so in this example I am done, as I just need to drop one of the three, the better two advance.

But what if the tie happens exactly on the border?

I assume there is no way to use the outcome of just two rounds to fairly determine a detailed ranking. Is a tie likely if we even use the ranks of the players they have beaten?

If someone can assure me that such a tie at about 1/3 or the ranks is very unlikely I'd go with this and use a cointoss in case of a tie.

Remember, we don't have time for an extra game, we allready have 4 rounds = 4 hours + extra time to plan and meet. And we can't use things like net worth, leaks or highest wave, as we play sr and the builds are not even yet.

[ocmassreo]

complicated but very interesting i like this

how you it third round would see ?

can be the 4 first in two half-final classification in vhx sr then final in vhx Sr

the 4 following (+ if equality) clashes in final of the loser

in vh Sr

it s possible to make clasement of 1 A 8

with 4 players who it are go in phase final

[Cisz]

how you it third round would see ?

What?

the 4 following (+ if equality) clashes in final of the loser

in vh Sr

I'm not sure what to think of this. Do you think the players that missed the final want to play for a ranking among them? I would expect that no, they don't, but I might be wrong. I don't want to set games for them where there is a great probability that they don't show up.

it s possible to make clasement of 1 A 8

with 4 players who it are go in phase final

What again?

[ocmassreo]

sorry lol

oki i try to use a translater but it s not good lol

1/ how u see the third round for the finalist ?

2/yep the 5nd to 8nd make a game (when semi are playing) for a classement for the 5 to 8 place

because some player are good but but some are very good (like u hyper ect...)

last tournament i loose against hyper and i dont know i f i can beat other finalist because i loose against him in first round ^^

juste a game with 5 to 8 player for see ours level

and all player stay for the final result

3/ yep to make a classification on 8 best players ^^

[jolin012]

if I know how many should advance to finals i can complete my sugguestition: i heard you say something like 6 advance if there are 16 playeres, which gives 0.375 advancing/player.

if so that gives:(p=players a=advancing)

P A

8 3

9 3

10 4

11 4

12 4/5

13 5

14 5

15 6

16 6

17 6

18 7

19 7

20 7/8

21 8

22 8

23 9

24 9

25 9

26 10

27 10

28 10/11

29 11

30 11

31 12

32 12

so, shall we have it like this or different? if like this, how many do you prefer for 12/20/28 players?

[Cisz]

A little less than 50% and an even number.

[jolin012]

a little less than 50% is that 37.5% as i showed or something else?

[jolin012]

i mean, are thus number even?

[jolin012]

however here is my sugguestition of a way to play those 2 qualifying rounds(sorry for many posts):

let's say there comes 12 players and we want 4 into the finals..

randomize the players into 4 groups of 3

Group A, B, C and D

they play, and after this round they gather new groups, not randomized however.

1st of a(the player winning first place of group a) nd 2nd of b 3rd of c in one group, we can call it group A2.

1st of b(the player winning first place of group nd 2nd of c 3rd of d in one group, we can call it group B2.

1st of c(the player winning first place of group c) nd 2nd of d 3rd of a in one group, we can call it group C2.

1st of d(the player winning first place of group d) nd 2nd of a 3rd of b in one group, we can call it group D2.

if you don't understand it maybe easier lie this:

groups round 2:

group 1

1a, 2b,3c(winner of group a, second of group b, third of group c)

group 2

1b,2c,3d

group 3

1c,2d,3a

group 4

1d,2a,3b

ROUND 2:

the winner of every group goes to the finals.

it may seem like the second round is the only one that matters but in fact, it is not. the better you do in round one, the easier opponents you meet in round 2. if you come first in round 1 you meet a seacond a third and a forth in the next round. if you do worse, oyu meet a first a second and a third. But if you are good but mistaked in the first round, you still have a chance to advance, if you prove your skill in round 2.

as you see there is not more than one a/b/c/d so you won't meet the same player again as did first round(if not played with less groups than players in each group, but there is no way to avoid this, not just in this way of playing the 2 rounds)

if there was something i didn't make clear, please ask, i am a bit tired now when im writing this x_x

[Cisz]

"ROUND 2:

the winner of every group goes to the finals."

This does not work out at all imo.

I a player wins first round and gets second on round two, he is not allowed into the finals? He had the harder game on round one, if you are corect. And a player that lost round one but wins round two advances? Allthough his combined result (3,1) is worse han the other players (1,2)?

And what about tournaments where the number of finalists is not a divider of the total players, as is 16 -> 6 or 20 -> 8?

You still need to convince me.

[jolin012]

ok, my reply is too big to fit in here^^ so I'll d o like this, a simple answer here for those wh ounderstand quick and don't want any details, nor answer to any questitions but the 16players->6finalists one, and i will upload the doc filewith the entire text.

Group A

(player ID) A1(winner)

A2

A3

A4(looser)

Group B: B1(win),B2,B3,B4(lose)

Group C: C1,C2,C3,C4

Group D: D1,D2,D3,D4

ok round two looks like this(players keep same ID to know where they came from(i. e. playername a1 means they come from winning group a))

Group A2(2 advancing): A1,B2,C3,D4,C2

Group B2(2 advancing): B1,A2,D3,C4,D2

Group C2(1 advancing): C1,A3,B4

Group D2(1 advancing): D1,B3,C4

LONG ANSWER: http://rapidshare.com/files/56798237/to ... t.doc.html

"You still need to convince me. " still not convinced? (don't answer until you read all the long answer muahahaha )

[jolin012]

Round 1 Round 2

Player 1 1 Player 1 1

Player 2 4 Player 5 4

Player 3 3 Player 9 2

Player 4 2 Player 13 4

Player 5 4 Player 2 3

Player 6 1 Player 6 1

Player 7 3 Player 10 2

Player 8 2 Player 14 4

Player 9 1 Player 3 1

Player 10 4 Player 7 3

Player 11 3 Player 11 4

Player 12 2 Player 15 2

Player 13 4 Player 4 4

Player 14 3 Player 8 3

Player 15 2 Player 12 2

Player 16 1 Player 16 1

probably won't make a big differance but i think you made a little miss here and on your scoreboard the same miss repeated.

[Cisz]

MagicalHacker did yet another cool tool, a tournament simulator.

It's meant to give a quick preview of his version of a four round tournament with 2 rounds prelim, semifinal and final.

>> Click me, I'm a tournament tool <<

Edit - Totally forgot: Jolin012 and MagicalHacker developed this method for conducting the prelims together, and I think it was Jolin who came up with it in the first place.

[ocmassreo]

i try with 16 players

name of players (a,b, ......o, p)

Test 1

[jolin012]

that's 14 players:P (as you have chosen not to use i or n)

Edit: Oh, i looked at the link and it sais you have chosen 16 players, and i tried myself, if i type 16 players, name them, and in step 3 of 5 choose the fourth option, something is bugged and it'll only show there are 14 players in the tournament, player "i" and "n" are removed.

[MagicalHacker]

Fixed. The link you posted is still broken, as you made a choice in step 3 that shouldn't have been possible at all.

[Cisz]

Hm. If this works as intended it would provide a working seeding w/o manual interference in just two rounds.

This sounds to good to be true.

MagicalHacker - do you believe this is as good as I just outlined? And there is no catch?

If so, we should go public and post that ruleset into the forums.

Edit - I fed your tool some numbers and I don't think the method works with less than 22 players.