|
On August 16 2013 12:34 lichter wrote:Show nested quote +On August 16 2013 12:21 Scarecrow wrote:On August 16 2013 12:10 lichter wrote:
Citing OSL as a comparison is disingenuous. They are 4man groups with the only tiebreaker possibility being a 3 way tie for first or second, with all the players having tied records against the other tiebreaker contenders. It is impossible in OSL groups to have a tie where one player has beaten both the other players. They are different kinds of groups so comparing it to OSL is irrelevant. What I meant was OSL is still essentially a best of one format with a superior tiebreaker (playing it out). Even if, due to time constraints, they had the games played unbroadcasted or even online it'd still be better than the current system where a progamer's livelihood hinges on GOM arbitrarily cherrypicking the right data. Who beat who is irrelevant when the cumulative scores are the same. If one player has already beaten both, why must he replay only those games where he already won? For a player who lost to both, why must he only play against the players he is likely to lose against? If 'who beat who' is irrelevant, why not replay the whole group until the result isn't a tie?
They are tied and should have the chance to prove who is better by playing. The player who has beaten both has also lost two games to players they beat. It'd be silly to replay the whole group when only 3 are involved in the tiebreak, especially as noone else will be motivated. Just have them play each other and have them break the tie through additional games rather than the luck of having beat the 'right' person.
|
So, is there gonna be a discussion over the fairness of the tiebreaker system every time? I thought all points should´ve been made by now.
|
I have to agree with Scarecrow. I just don't know what the better format is anyway. Almost every format has a flaw, and I'm yet to see one without complaints.
Bracket: "Lucky bracket"
Bo3 Winner-Loser Groups: "How come 4-4 advances over 4-3?" or "Head to head he's 3-2, why not advance?!"
U/D Bo1: "Useless games for some people. Head to head can work out weirdly."
Original U/D system: "Code S too stagnant"
etc etc.
Is there "the perfect system"? Why aren't we all using it already!?
|
1001 YEARS KESPAJAIL22272 Posts
It's the usual pointless argument.
A > B > C > A > B >C...
A beat B. B beat C. C beat A. But it doesn't stop there. A beat B who beat C. Bla bla bla.
The games that you believe determine who is 'better' have already been played once.
I really have to give up explaining this.
|
Whhhhaaattt? One of my favourite terrans back in business!! Fuck yeahhhhh!
Just disappointed he isnt coming to europe....
|
On August 16 2013 15:05 lichter wrote:
The games that you believe determine who is 'better' have already been played once.
And they were tied. Hence playing another round of games between those tied to determine who advances. The current system is just as fair as saying 'oh Hurricane should advance because he beat the best player in the group.'
|
|
|
On August 16 2013 16:01 Scarecrow wrote:Show nested quote +On August 16 2013 15:05 lichter wrote:
The games that you believe determine who is 'better' have already been played once.
And they were tied. Hence playing another round of games between those tied to determine who advances. The current system is just as fair as saying 'oh Hurricane should advance because he beat the best player in the group.'
I think what people are trying to say is that yes, we could take the people that were tied and play another set of games. But since these games between the players have already been played. We will use those results. This is essentially what this tie breaker system does.
It is what all H2H system does.
|
On August 16 2013 14:43 bittman wrote:
I have to agree with Scarecrow. I just don't know what the better format is anyway. Almost every format has a flaw, and I'm yet to see one without complaints.
Bracket: "Lucky bracket"
Bo3 Winner-Loser Groups: "How come 4-4 advances over 4-3?" or "Head to head he's 3-2, why not advance?!"
U/D Bo1: "Useless games for some people. Head to head can work out weirdly."
Original U/D system: "Code S too stagnant"
etc etc.
Is there "the perfect system"? Why aren't we all using it already!?
Coinflip used at DH is the best and most balanced system that allows every player have equal chances to get into Code S.
|
On August 16 2013 16:15 monkybone wrote:
To people calling it fair because it's convention and thus something everyone are equally subjected to, that's not the point. In one sense one player performed better, in another, some other player did. And that's really what matters, player performance. The conclusion is that nothing could really be considered perfectly fair in a tied score situation. The best, most practical, and fairest option would be to always play out the tiebreakers.
Yet, many people on these forums argued for Bo1 OSL Ro32 because that has already been the way for OSL. Are Bo1s the fairest determination of player performance? Especially for SC2?
|
On August 16 2013 16:01 Scarecrow wrote:Show nested quote +On August 16 2013 15:05 lichter wrote:
The games that you believe determine who is 'better' have already been played once.
And they were tied. Hence playing another round of games between those tied to determine who advances. The current system is just as fair as saying 'oh Hurricane should advance because he beat the best player in the group.'
I think this is a difference of semantics.
What you are trying to argue is that the system is unfair because it does not pick the two OVERALL best players of the group. For example, if 3 people are tied at 2-2, it's obvious that they are tied as in "all three of them had the same performance with respect to the entire group"
However, try to think of it this way:
The 1st player is OBVIOUSLY qualified, there is no dispute over that
The last player is OBVIOUSLY not qualified, no dispute here too
The only question is the three remaining tied players
With tiebreakers, we're no longer trying to find out which of the three performed best in the group... we are creating a NEW group with JUST those three players. Then, we try to find out which of those three players is the best against each other. In other words, it's no longer about who was 2nd best in Group F... it's about "which of these 3 players is the best?"
Except the mini-group has already been played.
|
On August 16 2013 16:23 Entirety wrote:Show nested quote +On August 16 2013 16:01 Scarecrow wrote:On August 16 2013 15:05 lichter wrote:
The games that you believe determine who is 'better' have already been played once.
And they were tied. Hence playing another round of games between those tied to determine who advances. The current system is just as fair as saying 'oh Hurricane should advance because he beat the best player in the group.' I think this is a difference of semantics. What you are trying to argue is that the system is unfair because it does not pick the two OVERALL best players of the group. For example, if 3 people are tied at 2-2, it's obvious that they are tied as in "all three of them had the same performance with respect to the entire group" However, try to think of it this way: The 1st player is OBVIOUSLY qualified, there is no dispute over that The last player is OBVIOUSLY not qualified, no dispute here too The only question is the three remaining tied players With tiebreakers, we're no longer trying to find out which of the three performed best in the group... we are creating a NEW group with JUST those three players. Then, we try to find out which of those three players is the best against each other. In other words, it's no longer about who was 2nd best in Group F... it's about "which of these 3 players is the best?" Except the mini-group has already been played.
You see, this is cherry-picking data. Now you're giving more weight to some games than to others. Basically, you are ignoring the favored player's losses because he took more losses outside your mini-group than the non-advancing players. There is no way you can say one player deserves to advance over the others with the same score without some kind of twisted logic.
Using H2H score to determine who advances has just as much legitimacy as coin flips or rock-paper-scissors, or choosing whoever has the longest name.
|
On August 16 2013 17:09 pylonsalad wrote:Show nested quote +On August 16 2013 16:23 Entirety wrote:On August 16 2013 16:01 Scarecrow wrote:On August 16 2013 15:05 lichter wrote:
The games that you believe determine who is 'better' have already been played once.
And they were tied. Hence playing another round of games between those tied to determine who advances. The current system is just as fair as saying 'oh Hurricane should advance because he beat the best player in the group.' I think this is a difference of semantics. What you are trying to argue is that the system is unfair because it does not pick the two OVERALL best players of the group. For example, if 3 people are tied at 2-2, it's obvious that they are tied as in "all three of them had the same performance with respect to the entire group" However, try to think of it this way: The 1st player is OBVIOUSLY qualified, there is no dispute over that The last player is OBVIOUSLY not qualified, no dispute here too The only question is the three remaining tied players With tiebreakers, we're no longer trying to find out which of the three performed best in the group... we are creating a NEW group with JUST those three players. Then, we try to find out which of those three players is the best against each other. In other words, it's no longer about who was 2nd best in Group F... it's about "which of these 3 players is the best?" Except the mini-group has already been played. You see, this is cherry-picking data. Now you're giving more weight to some games than to others. Basically, you are ignoring the favored player's losses because he took more losses outside your mini-group than the non-advancing players. There is no way you can say one player deserves to advance over the others with the same score without some kind of twisted logic. Using H2H score to determine who advances has just as much legitimacy as coin flips or rock-paper-scissors, or choosing whoever has the longest name.
You're completely right - we ARE cherry-picking because the other games DON'T matter - all other players have had their fates decided. Group winner is through, group loser is out, those matches do not matter anymore!
The only important matches are the ones still in contention for that Code S spot... meaning the games between themselves. That's all.
|
You're not fully understanding why H2H is unfair and arbitrary. You say the games outside the tied group don't matter. By saying that you're basically ignoring the games that Kangho went 0-2 and Hurricane went 2-0. Isn't that unfair to Hurricane? Why do you only choose the games that Kangho won to say that only those games matter? Of course Kangho won against the other tied players. So what? That doesn't make his wins any more "valuable" than Hurricane's wins or Effort's wins. To say that those wins are more valuable is completely arbitrary. You can also make the argument that whoever has the best record against the top player (jjakji) should advance, but that would also be arbitrary. There is no reason to value some games more than others.
Btw, I have no interest in any of these players. From a purely logical standpoint using H2H score just does not determine who played better or "deserves" to advance.
|
|
|
On August 16 2013 17:57 pylonsalad wrote:
You're not fully understanding why H2H is unfair and arbitrary. You say the games outside the tied group don't matter. By saying that you're basically ignoring the games that Kangho went 0-2 and Hurricane went 2-0. Isn't that unfair to Hurricane? Why do you only choose the games that Kangho won to say that only those games matter? Of course Kangho won against the other tied players. So what? That doesn't make his wins any more "valuable" than Hurricane's wins or Effort's wins. To say that those wins are more valuable is completely arbitrary. You can also make the argument that whoever has the best record against the top player (jjakji) should advance, but that would also be arbitrary. There is no reason to value some games more than others.
Btw, I have no interest in any of these players. From a purely logical standpoint using H2H score just does not determine who played better or "deserves" to advance.
Actually, if we want to determined the 'better' player or the player that 'deserves' to make it. Then we would need to play the entire group again. Why? Match ups. Let's say the tied was between the 3 zergs (Shine, KangHo, Effort).
Say Effort v Z specialist and got 2-2 because he lost against the 2 non zergs and beat both zergs.
Say Kangho sucks in vZ and got 2-2 by beating both T and P and lost against Shine and Effort.
If they just play another set of tiebreaker, Effort would win most of the time. But does that make him a 'better' player? or the player that deserves to go through?
Yes,in H2H, some wins will be more 'valuable'. But in play again tiebreakers, you still have match up 'imbalance'. A lot depends on who you are tied with. If Effort was tied with the 2 non zergs, he wouldn't make it. But he was 'lucky' and tied with the zergs, he gets to go through.
|
On August 16 2013 17:57 pylonsalad wrote:
You're not fully understanding why H2H is unfair and arbitrary. You say the games outside the tied group don't matter. By saying that you're basically ignoring the games that Kangho went 0-2 and Hurricane went 2-0. Isn't that unfair to Hurricane? Why do you only choose the games that Kangho won to say that only those games matter? Of course Kangho won against the other tied players. So what? That doesn't make his wins any more "valuable" than Hurricane's wins or Effort's wins. To say that those wins are more valuable is completely arbitrary. You can also make the argument that whoever has the best record against the top player (jjakji) should advance, but that would also be arbitrary. There is no reason to value some games more than others.
Btw, I have no interest in any of these players. From a purely logical standpoint using H2H score just does not determine who played better or "deserves" to advance.
Imagine it like this:
Up & Downs - jjakji, Shine, EffOrt, KangHo, Hurricane - Round Robin Bo1
1st - jjakji - Advance directly to Code S
5th - Shine - Fall directly to Code A
2nd-4th - KangHo, EffOrt, Hurricane - Advance to Up & Down Round 2
Up & Down Round 2 - Hurricane, KangHo, EffOrt - Round Robin Bo1
KangHo > Hurricane
KangHo > EffOrt
EffOrt > Hurricane
1st - KangHo (2-0) - Advance to Code S
2nd - EffOrt (1-1) - Code A
3rd - Hurricane (0-2) - Code A
In this scenario, would you consider it unfair? Suddenly all other games don't matter because a new group was played... Hurricane's victories over Shine and jjakji were ignored, and he dropped out in the tiebreaker group.
Now, in the interest of saving time, let's simply assume that the second group will yield the same results as when they first played. If KangHo beat Hurricane the first time, we assume that he will win again. In this manner, the second group need not be played, time is saved, and we have our two players advance.
H2H is not so absurd as you may think.
|
On August 16 2013 19:10 Entirety wrote:Show nested quote +On August 16 2013 17:57 pylonsalad wrote:
You're not fully understanding why H2H is unfair and arbitrary. You say the games outside the tied group don't matter. By saying that you're basically ignoring the games that Kangho went 0-2 and Hurricane went 2-0. Isn't that unfair to Hurricane? Why do you only choose the games that Kangho won to say that only those games matter? Of course Kangho won against the other tied players. So what? That doesn't make his wins any more "valuable" than Hurricane's wins or Effort's wins. To say that those wins are more valuable is completely arbitrary. You can also make the argument that whoever has the best record against the top player (jjakji) should advance, but that would also be arbitrary. There is no reason to value some games more than others.
Btw, I have no interest in any of these players. From a purely logical standpoint using H2H score just does not determine who played better or "deserves" to advance. Imagine it like this: Up & Downs - jjakji, Shine, EffOrt, KangHo, Hurricane - Round Robin Bo11st - jjakji - Advance directly to Code S 5th - Shine - Fall directly to Code A 2nd-4th - KangHo, EffOrt, Hurricane - Advance to Up & Down Round 2 Up & Down Round 2 - Hurricane, KangHo, EffOrt - Round Robin Bo1KangHo > Hurricane KangHo > EffOrt EffOrt > Hurricane 1st - KangHo (2-0) - Advance to Code S 2nd - EffOrt (1-1) - Code A 3rd - Hurricane (0-2) - Code A In this scenario, would you consider it unfair? Suddenly all other games don't matter because a new group was played... Hurricane's victories over Shine and jjakji were ignored, and he dropped out in the tiebreaker group. Now, in the interest of saving time, let's simply assume that the second group will yield the same results as when they first played. If KangHo beat Hurricane the first time, we assume that he will win again. In this manner, the second group need not be played, time is saved, and we have our two players advance. H2H is not so absurd as you may think.
All other games do matter, they ended up getting them into the tiebreak in the first place.
@ the 'they already played their round robin' argument: H2H basically counts some games twice (which is just too much weight to put on single Bo1's), that's why it's fairer to play it out again with the tied players.
"If KangHo beat Hurricane the first time, we assume that he will win again," is just ridiculous. He won a Bo1 so we essentially give him a second free-win to break the tie. In your example Kangho advancing is fair because he has won more games and they are no longer tied. Their fate is in their hands rather than being lucky enough to win the Bo1 that gets counted twice... I can't believe some are arguing H2H is superior or even equal to playing actual tiebreaker games in anything but timesaving.
The absolute fairest way (in terms of mu's) is of course playing out the entire group again, but that's not realistic for several reasons. So playing out a smaller tiebreaker group with the relevant, motivated players is a nice compromise between that and arbitrarily ending it on H2H.
|
The other games do not "suddenly don't matter". They matter in that they allowed these players to get into the 2nd round tiebreakers. Shine doesn't get to play the 2nd round because those games mattered. No matter how you want to twist your logic, H2H doesn't make sense.
Why would you just assume that Kangho will beat Hurricane again? That is a ridiculous and ARBITRARY assumption to make.
H2H is absurd and you keep twisting your logic to try to make a circle fit into a square.
|
Something tells me that some people do not watch real sports. Head to head is the most common way to break ties in pretty much every sport.
Everything is arbitrary. Not just the tiebreakers, but every tournament structure in every tournament. People need to stop crying about perceived unfairness every time the results aren't what they want. And stop with the BS that you don't care about the results you are whining about. That's the least believable argument you can possibly make.
Life isn't fair. Life has time constraints. Tell me, do you have a better idea to break ties without additional games being played?
|
|
|
|
|
|
EPT
