Ladle 18

[1200]

@Corn: Well I was talking about randomizing the rest of teams after the seeded teams have been put in place, like i was telling Pike that if the seeds are accurate they'd meet in the semis. So we are on the same page.

[owned]

It seems to me you could say CT was lucky for not having to face TR early in the brackets last ladle, so why should they be lucky again?

We played them in the first round of Ladle 16 so that doesn't really mean much.

Yeah, but you're not basing these seedings on the results of Ladle 16 do you? It's not about being lucky once, but about basing seeding on a previous 'lucky' placement and then pretend it's somehow fairer.

I'm sure that some of the wins were lucky, but the majority weren't. The majority of the time, the better team beats the worse one.

Obviously it's not 100% accurate that we're the 2nd best, but its the closest we can get. According to math, there's an approximately 50% or 75% chance we're the 2nd best, when considering the 3rd place match results. No other team is even close to that.

I'd like to see the math that shows CT was better than any of the teams in the other half of the bracket.


But I guess I'm too late arguing anyway.

Ok so as I said above, (assuming if a team beats another they are better than them) we were either automatically better than 50% or 75%
1.So we are technically better than everyone in our quarter, since we either beat them or beat somebody who beat them. So we're better than 25% atm.
2. The same thing could be said about everyone in the other quarter on our half, since we either beat them, or beat somebody who beat them or we beat somebody who beat somebody who beat them. So 25+25=50%
3. So this is where the 3rd place match comes in. (The 3rd place never really happened. Since SP didn't arrive we called it a forfeit) If SP had beaten X, that would automatically make us better than all of the teams in x's quarter. So 50+25=75% If SP had lost, it would of stayed the same as before, or 50% of the players.

Of course this is only considering that the better team always wins.

[Lackadaisical]

No you're talking about something else now. You said there was a 50% chance CT was the second best, but all you've shown now is that CT was better than 50% of the participants.

The way I see it, there are three teams (CT, X, DS) who could've claimed second place because they weren't beaten by anyone except for TR, who was clearly the best team in ladle 17. And because there isn't any way of saying which of these three teams is the best CT has a chance of 100% / 3 = 33% of being the second best. Although this is actually an OK percentage DS actually had the same chance of being the second best team, and they didn't get seeded at all. (If SP had beaten X this chance would've gone up to 50%, but the point remains: DS could claim the same thing)

This is also the 'luck' I was talking about earlier: If ladle 17 had been played with the starting positions of DS and CT switched around, there's no reason to believe DS wouldn't have been the one in the finals and CT would've been out in the first round.

And that is why I am/was opposed against seeding this time using the ladle 17 results as a the sole guideline.

[owned]

No you're talking about something else now. You said there was a 50% chance CT was the second best, but all you've shown now is that CT was better than 50% of the participants.

We're still considered the 2nd best since we are "better" than the most teams.

The way I see it, there are three teams (CT, X, DS) who could've claimed second place because they weren't beaten by anyone except for TR, who was clearly the best team in ladle 17. And because there isn't any way of saying which of these three teams is the best CT has a chance of 100% / 3 = 33% of being the second best. Although this is actually an OK percentage DS actually had the same chance of being the second best team, and they didn't get seeded at all. (If SP had beaten X this chance would've gone up to 50%, but the point remains: DS could claim the same thing)

That's ridiculous. Just because a team loses to the eventual winner, that doesn't give them the same chance of being the 2nd best as the eventual 2nd place team. We won 3 games. Ds won none. Yes it is a possibility, but I'm not gonna count it considering how small it is. Same thing with X except their chance was a bit bigger.

This is also the 'luck' I was talking about earlier: If ladle 17 had been played with the starting positions of DS and CT switched around, there's no reason to believe DS wouldn't have been the one in the finals and CT would've been out in the first round.

Kinda cut out from my last statement. There is a chance that DS or X should've been 2nd best last ladle. But the chance is smaller than us being the 2nd best.

[epsy]

Are we nitpicking here?

Yes, you already lost me p?ages ago.

[Lackadaisical]

Well sure CT won more games than DS, but DS didn't get the chance to win any games since they had to play TR in the first round.

And yeah I've only been talking about being lucky in placement, I couldn't figure out why you were bringing getting lucky in matches into the argument.

But I'm pretty sure this isn't going anywhere, and I've said what I wanted to say so this will be it from me.

[Z-Man]

No you're talking about something else now. You said there was a 50% chance CT was the second best, but all you've shown now is that CT was better than 50% of the participants.

We're still considered the 2nd best since we are "better" than the most teams.

The way I see it, there are three teams (CT, X, DS) who could've claimed second place because they weren't beaten by anyone except for TR, who was clearly the best team in ladle 17. And because there isn't any way of saying which of these three teams is the best CT has a chance of 100% / 3 = 33% of being the second best. Although this is actually an OK percentage DS actually had the same chance of being the second best team, and they didn't get seeded at all. (If SP had beaten X this chance would've gone up to 50%, but the point remains: DS could claim the same thing)

That's ridiculous. Just because a team loses to the eventual winner, that doesn't give them the same chance of being the 2nd best as the eventual 2nd place team. We won 3 games. Ds won none. Yes it is a possibility, but I'm not gonna count it considering how small it is. Same thing with X except their chance was a bit bigger.

Indeed, someone has to learn about proper probability calculations In the simplified case of four teams in a knockout tournament and assuming deterministic match results (i.e. the better team always wins), the chance that the losing finalist is indeed the second best team is 2/3. Proof: only the final loser and the loser of the first round against the final winner qualify. ANY team would have lost against the tournament winner there, so it's just a one in three chance of getting the unlucky staring bracket; therefore, chances that the first round loser is the second best team is 1/3, leaving 2/3 for the only other possibility. So while losing the final is in no way a proof of actually being second best (you'd have to do a follow up tourney of all the players having lost only to the winner), they are with the largest probability.

Exercise for the reader: assuming larger and larger tournaments, what is the limit of the probability that the second best team indeed ends up as the loser of the final?

[Corn1]

The way I see it, there are three teams (CT, X, DS) who could've claimed second place because they weren't beaten by anyone except for TR, who was clearly the best team in ladle 17. And because there isn't any way of saying which of these three teams is the best CT has a chance of 100% / 3 = 33% of being the second best. Although this is actually an OK percentage DS actually had the same chance of being the second best team, and they didn't get seeded at all. (If SP had beaten X this chance would've gone up to 50%, but the point remains: DS could claim the same thing)

That's ridiculous. Just because a team loses to the eventual winner, that doesn't give them the same chance of being the 2nd best as the eventual 2nd place team. We won 3 games. Ds won none. Yes it is a possibility, but I'm not gonna count it considering how small it is. Same thing with X except their chance was a bit bigger.

Actually it isnt ridiculous.. Did you ever wonder why you would win money if you were beat by the winner of the tronic cup in a previous round? It's because once you get beat, your out it doesn't mean if it was suppose to be the team just a tiny bit better than you or the best team in the tournament. What's not ridiculous is the fact that how the bracket gets set up can be incredibly in your advantage or make it nearly impossible for you to win.

[Concord]

sounds like you guys want a league. . .

[Lackadaisical]

Ok, I stand corrected.




Can I still save some cred by saying the limit goes to 50%?



edit:

But it does bring me to the question: do we need seeding at all if we know there's a minimum of fifty percent chance anyway that we're getting the two strongest teams in the final.

[owned]

Exercise for the reader: assuming larger and larger tournaments, what is the limit of the probability that the second best team indeed ends up as the loser of the final?

Assuming that the tournament has a power of 2 amount of teams? Because if not, it can nearly reach 0.

[DDMJ]

@ flex: The reason the wiki Challenge Board wasn't formatted like that before (with the == Heading Here == stuff) was because that long list of Table of Contents was quite annoying. When the page is first loading you have to scroll down quite a bit. Whereas before (with just ;Heading Here), it would still be bold and recognized (the headings), but the Table of Contents wouldn't appear until much further down on the page.

[Z-Man]

Owned's result of zero is only true if you distribute the byes very badly (like, give one team a bye straight to the final). You can keep the 50% limit intact even for non-power-of-two team numbers.

Can I still save some cred by saying the limit goes to 50%?

edit:

But it does bring me to the question: do we need seeding at all if we know there's a minimum of fifty percent chance anyway that we're getting the two strongest teams in the final.

Both correct

Randomizing teams every time has the following properties:
- it mixes encounters up. In the long run, you'll fight every other team.
- it gives weaker teams a chance to get far if they get lucky with the random number generator a bit and don't face strong teams early on (see: Germany in just about every soccer tournament )

IF you seed based on previous tournaments, I suggest to do it this way:
Step 1: Take the brackets from the last tournament and resort them so that the winning team in each encounter comes from the top. So

Code: Select all

A--\
    B--\
B--/    \
         B
C--\    /
    C--/
D--/

Gets transformed to

Code: Select all

B--\
    B--\
A--/    \
         B
C--\    /
    C--/
D--/

For the next tournament, the teams that lost in the first round keep their place in the sorted brackets. Teams that won the first round get reshuffled; each of them moves up one slot, except for the total winner of course, that moves down. In the simple example, this leads to

Code: Select all

C--\
    ?--\
A--/    \
         ?
B--\    /
    ?--/
D--/

for the next tournament. This system has the following properties:
- if team X lost against team Y and team Y made it to round N, then X and Y won't meet in the next tournament.
- Especially, anyone who directly lost against the winner only last time won't meet last time's winner before the final, therefore the loser of the second iteration final (assuming consistent, non-random performance) is guaranteed to be the second best.
- Iterating that, anyone making it to the semifinals in the third iteration is guaranteed to be among the four best. And so on.
- you will meet a lot of other teams, but not as many as with the random seeding: you won't meet those on your own skill level. The two teams getting eliminated in the semis will never meet, the four teams eliminated in the quarters will never meet, and so on.
- weak teams will get eliminated early in the long run.

(Byes would be handled like regular teams that always lose. Teams that drop out from one iteration to the next get transformed to byes. New teams take the spot of a random bye, and when none are available, the brackets get extended by one round first, matching up everyone with a bye.)

Personally, I prefer random seeding, because it mixes things up and you don't have to worry about how you're going to handle "new" teams that still mainly consist of members of an old team.

[freako]

Let's just start an unseeded, double-elimination tourney

btw, isn't is time we should get an more or less centralised organisation for the ladle? I hate the idea, but it seems it doesn't work without it

[sinewav]

I like all the suggestions (Z-man's, Lack's, Lucifer's, freako's ...). I "voted" for seeding this time solely because we never tried it and a lot of people were calling for it. I like to try new things, even if they seem silly at the time.

But if a compromise can't be made, and the ladle retains it current popularity, maybe splitting it into two leagues is the best way. There are almost enough teams to do it. Each league could make their own decision on seeding vs random. Who knows, maybe one league will be more popular and suffocate the smaller one out of existence.

@freako, remember on IRC there was talk about making a poll with all the different options? Well, there are quite a few of them here. Maybe next week we can start that poll and see if it gets us anywhere (or drives us in circles again).