These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE Alliance Tournament Discussion

 
  • Topic is locked indefinitely.
 

Final stages of tournament

First post First post
Author
BillNyeTheEVEguy
Touch of Madness
#1 - 2012-11-25 03:24:36 UTC
So how is the final stage of the tournament going to work? That is, what's going to happen once there are four teams remaining?

Let's imagine XXXMity and Africa's Finest are the last two teams remaining in their bracket, and Asine Hitama's Team + Much Crying are the last teams in theirs. At that point, who plays each other? Do XXXMity and Asine Hitama get an advantage because they're playing from the Winner's Bracket? Help me out here, CCP.
CCP Veritas
C C P
C C P Alliance
#2 - 2012-11-25 10:54:25 UTC
Sure thing.

The winners of the Winners bracket play the winners of the Losers bracket. At this phase, we're still dealing with two double-elimination tournaments, so if the Losers dudes win the first match, they need to play again as that means the Winners guys have only lost once. However, if the Winners dudes win the first match then it's done, as the Losers guys will have lost twice.

That's how the two brackets resolve.

On the final day, the winners of both brackets play each other in a best of 5 to determine the tournament 1st and 2nd place. The losers of the last match of each bracket will play each other in a best of 3 to determine the tournament 3rd and 4th place finishers.

CCP Veritas - Technical Director - EVE Online

tofucake prime
The Hatchery
RAZOR Alliance
#3 - 2012-11-25 13:38:48 UTC  |  Edited by: tofucake prime
This is not how double elimination tournaments work and it is confusing and bad.

This is two tournaments (left and right), and then a showmatch for third and fourth, then a Best of 5 showmatch for first and second.
X ATM092
The Hatchery
RAZOR Alliance
#4 - 2012-11-25 13:58:24 UTC  |  Edited by: X ATM092
Some basic theory for you guys at CCP.

The double elimination format is effective because it reduces the impact of the random brackets by ensuring the best two teams make it into the final. Even if a top 2 team get prematurely knocked into the losers brackets they should win through there and then return to the finals. What would happen in a tournament run correctly is the left and right sides would continue to be run simultaneously as part of the same tournament knocking people down into the losers until the winner of the left and the winner of the right play each other. Whoever loses that would play the losers bracket champion to ensure that the two teams in the grand finals were the best 2. Then they go into the final with a one game deficit.

What you have done is created two unrelated tournaments and then a showmatch. The left hand side of the bracket produces a #1 and a #2 as does the right. Then you arrange a showmatch between the two #1s and the two #2s and then proclaim the loser of the #1s as the 2nd place finisher for the tournament as a whole. This is nonsense as if they came from the right hand side of the bracket then they never had anything to do with the left hand side, they could be way, way worse than the #2 from the left hand side. What you have done isn't a problem for us because rather than play Verge or Ronin or Hydra in a bo5 we instead get to play any of the right hand side guys but I think you ought to clarify what you're doing so people don't think this is an actual tournament. It's not, it's two unrelated tournaments and then a showmatch between two teams who have proved themselves in completely different and unrelated environments for laughs and $. Your format guarantees that two teams on the right hand side of the bracket will always achieve at least 2nd and 4th, regardless of whether or not they actually merit it because your format is an abomination which undermines everything the double elimination format attempts to achieve.

This is not a problem for us as if we win today then we only have to win 1 game with a game in hand against the best teams in the tournament in order to guarantee top 2 and a bo5 with some random incursion bears but it is possible that people misunderstood exactly how the format works. Also, although this is just me being cynical here, is it possible that whoever drew the brackets just drew them badly and you are trying to subsequently honour them for whatever reason? I can't imagine any reason why someone could be aware of double elimination and then run this mess instead, it seems much more likely that this is just like the time someone forgot to write "medal" in the prize of AT X and then it subsequently became enforced rather than just saying "my bad, that was an oversight".

Also is there any chance that you could explain how it's run on the stream today, we can't be the only people surprised to learn that it's actually two tournaments and then a misleadingly named showmatch.
CCP Veritas
C C P
C C P Alliance
#5 - 2012-11-25 14:20:45 UTC
X ATM092 wrote:
Lots of words

Hi, I set up the format and wrote the program which performed the seeding. I suppose you hate me and my computer's Python installation now.

The reason it's set up like this is because we wanted a best-of-five for the top, which is an idea fundamentally incompatible with double-elimination. Instead of tossing the whole idea out and going with a weaker system leading up to the final, we hybridized it as best we can - having two double-elim tournaments leading into a separate finals. This does have the problem of having a possible misordering between #2 and #3 (and even 4th), but that's just something we're going to have to live with this time. This would be less of a problem if we had some metric to base seeding on instead of straight randomness.

In case you think we're way off in left field with this format, it's also similarly used in the College World Series.

CCP Veritas - Technical Director - EVE Online

X ATM092
The Hatchery
RAZOR Alliance
#6 - 2012-11-25 14:30:22 UTC
CCP Veritas wrote:
X ATM092 wrote:
Lots of words

Hi, I set up the format and wrote the program which performed the seeding. I suppose you hate me and my computer's Python installation now.

The reason it's set up like this is because we wanted a best-of-five for the top, which is an idea fundamentally incompatible with double-elimination. Instead of tossing the whole idea out and going with a weaker system leading up to the final, we hybridized it as best we can - having two double-elim tournaments leading into a separate finals. This does have the problem of having a possible misordering between #2 and #3 (and even 4th), but that's just something we're going to have to live with this time. This would be less of a problem if we had some metric to base seeding on instead of straight randomness.

In case you think we're way off in left field with this format, it's also similarly used in the College World Series.

Any chance you could clarify that it's two tournaments and a showmatch on the stream today so nobody gets confused. Unfortunately the College World Series has insufficient cultural weight to make it apparent what people are talking about when the words double elimination are misleadingly used.
CCP Veritas
C C P
C C P Alliance
#7 - 2012-11-25 14:32:26 UTC
I'll talk to the dudes who are on camera about it, yup.

And more people should go to the CWS. Omaha is an underrated town.

CCP Veritas - Technical Director - EVE Online

Cartheron Crust
The Tuskers
The Tuskers Co.
#8 - 2012-11-25 14:39:45 UTC
What?

If you were that set on having a best of five for the final you could run a normal double elimination and just have the last two left do a best of 5. So instead of the team coming from the winners bracket having to only win once or the winner from the losers bracket having to win twice as is usual for the final matches in a double elimination you have your best of five.

Or you can use the version you are currently which ignores logic.
X ATM092
The Hatchery
RAZOR Alliance
#9 - 2012-11-25 14:42:04 UTC
CCP Veritas wrote:
I'll talk to the dudes who are on camera about it, yup.

And more people should go to the CWS. Omaha is an underrated town.

Thank you, also thank you for responding with more courtesy than I perhaps deserved. I was somewhat shocked when I eventually worked out exactly how this was being run.
Aurelia Septima
Hedion University
Amarr Empire
#10 - 2012-11-25 17:01:54 UTC  |  Edited by: Aurelia Septima
CCP Veritas wrote:
. This would be less of a problem if we had some metric to base seeding on instead of straight randomness.


but you could have noticed that you put all top runners from the latest 2 alliance tournaments in one side and run your program once more?Roll
X ATM092
The Hatchery
RAZOR Alliance
#11 - 2012-11-25 17:58:18 UTC
Aurelia Septima wrote:
CCP Veritas wrote:
. This would be less of a problem if we had some metric to base seeding on instead of straight randomness.


but you could have noticed that you put all top runners from the latest 2 alliance tournaments in one tournament and then told them that only one of them could be in the top 2 while guaranteeing that a scrub team also got top 2Roll

fixed
CCP Veritas
C C P
C C P Alliance
#12 - 2012-11-25 17:59:24 UTC
Aurelia Septima wrote:
but you could have noticed that you put all top runners from the latest 2 alliance tournaments in one side and run your program once more?Roll

I could have done that, but then we couldn't say with a straight face that seeding was random. It'd be random-except-if-we-have-some-problem-with-how-it-came-out.

You just roll the dice and live with what comes out, that's how it has to be if you're doin' randoms.

CCP Veritas - Technical Director - EVE Online

X ATM092
The Hatchery
RAZOR Alliance
#13 - 2012-11-25 18:33:34 UTC
It's just sad because Africa's finest going out at this stage is not something they deserve but they were in a seriously stacked bracket.
BillNyeTheEVEguy
Touch of Madness
#14 - 2012-11-25 18:48:33 UTC  |  Edited by: BillNyeTheEVEguy
CCP Veritas wrote:
Aurelia Septima wrote:
but you could have noticed that you put all top runners from the latest 2 alliance tournaments in one side and run your program once more?Roll

I could have done that, but then we couldn't say with a straight face that seeding was random. It'd be random-except-if-we-have-some-problem-with-how-it-came-out.

You just roll the dice and live with what comes out, that's how it has to be if you're doin' randoms.

Other esports tournaments rig seeding to an extent. GSL (the biggest SC2 tournament) and the OSL (the biggest SC:BW tournament) have both done it. They don't admit it, but the evidence is there. Maybe you have a moral problem with doing it yourself, but there's no doubt that tweaking the brackets makes for a greater chance of a finals where the best two teams play each other.
Trebor Daehdoow
The Scope
Gallente Federation
#15 - 2012-11-26 01:15:25 UTC
Perhaps the simplest way to do it is have a single double-elimination bracket, with the final being a best-of-5 with the winner's bracket winner going into the best-of-5 with 1 automatic win (since they have not lost yet). So they have to win 2 matches to win, and their opponent has to win 3 matches.

3rd and 4th would be a best of 3 between the last 2 teams eliminated from the loser's bracket, but since they have 2 losses, they start out even.

In both cases, once a team has gotten to the final 4, they are knocked out when they have 4 losses.

As for seeding, I could be evil and suggest a multiround blind auction where the only info teams get is whether or not they have bid enough to get a slot, and where seeding and potential byes are determined by how much a team bids. Thus, the teams that bid the most don't meet until the final round of the winner's bracket (assuming they survive). Lots of lovely metagaming possibilities, don't you think? Twisted

Private Citizen • CSM in recovery

Velicia Tuoro
Light Speed Interactive
#16 - 2012-11-27 14:52:27 UTC
Well, I'm still a little confused...

This is my attempt to interpret what the weekend will look like:

Day 1
Losers Bracket Deciders
match 1 - The Expendables vs XXX Mity
match 2 - Reputation Cartel vs Goggles

Resolving the Winners VS Losers Brackets

match 3 - R0Nin and Pixies vs Match 1 Winner
match 4 - Asine Hitama's Team vs Match 2 Winner
match 5 - The R0Nin vs Match 1 Winner (only if losers bracket team won first match)
match 6 - Asine Hitama's Team vs Match 2 Winner (only if losers bracket team won first match)

Day 2
3&4 place best of 3
match 7
match 8 - losers from matches 3/5 and 4/6
match 9

1&2 place best of 5
match 10
match 11
match 12 - winners from matches 3/5 and 4/6
match 13
match 14

Am I anywhere near right? What?

Senior Representative Light Speed Interactive http://www.lightspeedinteractive.net

Velicia Tuoro
Light Speed Interactive
#17 - 2012-11-28 12:45:43 UTC
Woo, looks like I'm not as dumb as I feel!

https://forums.eveonline.com/default.aspx?g=posts&m=2235243#post2235243

Senior Representative Light Speed Interactive http://www.lightspeedinteractive.net