Skip to main content
Welcome guest. | Register | Login | Post

Determining the universal gamefest winner

4 replies [Last post]
libervisco's picture
Offline
Joined: 2006-05-04

This is a continuation of this thread which has become heavy and long. Let's make this one brief and final, just ironing out the details based on what we've come to so far.

Basically, those eligible for the title of the universal winner of the game fest are the winners of each game tournament. If a winner is a team, one team member is chosen for eligibility either by team vote or a game specific score. Hence, if we have 5 tournaments, we will have 5 players to choose the universal winner from. The question now is, how to choose between them most fairly.

The current idea, based on the discussion I had with Jose in the last thread, is this. Each game component will have its own points system awarding points to wins, losses and draws. This system is to be relied on for the qualifications stage mainly, but will be tracked through the tournament stage as well. In the end, the one with most points wins.

Jose might notice here that I've pretty much ditched the idea of ratios, multipliers and all that which was introduced to deal with some of the following issues. I'll try to address them in a different and what I feel to be a simpler way.

1. Size differences between tournaments (game components)

Basically, the perceived problem here is that in a tournament which has less players and hence less matches to win it is obviously easier to win the tournament. So the proposed solutions were to have a minimum "at-bats" matches played before player is eligible, or later proposed to have a multiplier which penalizes smaller tournaments. However this is all out dated now, if we don't count each tournament player as eligible for universal title (but only the winners) and if we don't use ratios. When we just do a straight count of points based on wins, losses and draws, smaller tournaments are self penalizing.

Therefore, the desired result is already achieved. Less effort needed to win also means less chances to win the universal title. If that sounds lame then, well, we just have to make sure that each tournament has a big number of players so that we can have a lot of games and a more exciting race. But this depends not only on the organization, but game communities as well.

2. Differences between game specific point systems

The concern here is that if in one tournament we give 4 points for a win and in another 2, the system is skewed favorably towards the tournament with 4 points award per win.

But I think this should be fairly simple to solve. Just standardize. Let the win/loss/draw be worth the same in each tournament no matter how we define a "win", "draw" or "loss" in any of them. This is fairly simple in all team vs team and 1vs1 player matches no matter what game we're talking about. If it is a "fragfest" however (classic deathmatch style) the points can be a fair approximation. It is already laid out in the wiki.

Of course, if we are to stick with a "you must get at least 1 point for each match you participate in" then we can extend that 1 point for the fourth best in a fragfest to all of the rest.

3. Teams

For all intents and purposes within the team-based tournament a team is as one, a player. However, once the team-based tournament is over, only one member of the team is eligible for the universal winner title. That one can be chosen by vote or by a game specific score, depending on the game. In Tremulous it can be a vote (because in tremulous, a score is not easily related to how good a player is) while in Nexuiz and OpenArena it can be the number of frags.

In the end, the chosen player takes the points of his team as his own.

4. Special games with non-traditional matches

In case a certain game is of such nature that it would be difficult to have a larger number of matches, such is for example freeciv which has matches lasting for days, the solution is to simply increase the number of points for wins/losses/draws to "catch up" with other games. So if a match lasts 24 hours (not necessarily in one pieces mind you Eye ) and an average match of other games is about 2 hours we can give 12 times more points for the match that lasts 24 hours (basically, a win would then be 48 points).

In summary:

No ratios, no funny business. Just standardize the point numbers, adjust them if necessary for special situations and then count them up - immediately having our universal winner.

libervisco's picture
Offline
Joined: 2006-05-04
In case someone asks "isn't

In case someone asks "isn't this just a return to the old points system" the answer is "not exactly". The old system treated each played individually, even if they play in teams. The new one doesn't. This is a big difference. Also, the old system was imposed with less sensitivity to the uniqueness of each game while the new one allows for this and compensates for the change by simply adjusting the point numbers (see the freeciv example above).

And lastly, the new system is used to determine the universal winner only among tournament winners instead of everyone regardless. Smiling

Offline
Joined: 2007-09-11
I like this system. But,

I like this system. But, how many players are confirmed to show for the tournaments?

libervisco's picture
Offline
Joined: 2006-05-04
Sorry for late reply. Right

Sorry for late reply. Right now we have about 50 individual players signed up overall, which are not signed up as teams and the sign up did require people to sign up when they are seriously interested in participating, so most of them should show up. We'll make sure we have the player count we need for each tournament. If it happens that for some we don't, then it will either be reduced in size and adapted accordingly or cancelled (if no-shows are really that extreme, which I'd doubt).

Note that those 50 above are just individual players. Counting everyone in clans we are well around 100 or more players. Also, we will let some time after the final and most significant announcement for more people to join in fresh.

Cheers

libervisco's picture
Offline
Joined: 2006-05-04
Here are some calculations

Here are some calculations of number of matches played if we assume 12 players or teams and a standard 4 players/teams per qualification group:

OpenArena, Nexuiz, Tremulous, regardless of mode of play:

- With single elimination: 23 matches
- With double elimination: 28 matches

Armagetron Advanced (with a system described here):

If one match per group member: 20 matches
If number of matches equals number that would be played if each played 1on1: 26

I think it is easy to conclude that we have only two ways to go about standardizing.

- Single elimination all over and the first option in AA: around 20-23 matches per 12 players
- Double elimination all over and the second option for AA: around 26-28 matches per 12 players

Of course, if there are more than 12 players there are also more matches, but this is not a problem. It may be "lame", but it is fair and fair is sometimes lame. Less players = less competition = less matches = less effort = less deserving of the main fest prize. If one wants to increase chances of winning the ultimate title best way to do so is to participate in multiple tournaments. Recommended would be, one smaller and one bigger tournament.

I say we standardize on single elimination. 20 matches on 12 players is plenty (especially considering we don't do them all at once, that we still have 2 maps per match in FPS games, a lot of rounds per match in AA etc) and double elimination is a pain. Eye

I predict