|
The problem with MLG is that they aren't considering series to be a set of isolated games.
If I beat a player in the first round of a tournament 2-1, I won that series. If I happen to get knocked into the losers bracket later on in the tournament and my first round opponent fights his way all the way though the lower bracket to face me again, he earned the right to play me again in another isolated set of games.
Extended series really makes no sense, and I don't understand how anyone can agree to it. Even for halo, it is a terrible ruleset. I can not think of another organization that follows the extended series format. If a player gets 0-2'd into the losers bracket and then has to face his original opponent again, he essentially has to eliminate the OG player TWICE(4 wins). That is triple elimination.
Also, what happens if say the finals are a BO5. HuK vs IdrA play first round of a tournament. Idra wins 2-0, but HuK fights all the way through the losers bracket to face IdrA again. Now what? idra up 2 games in a what? Best of 9? 11? lol?
|
On November 10 2010 12:31 Liquid`Nazgul wrote:
Imagine playing the world cup of football losing 0-3 to someone in group stage and meeting them in the finals again. According to the rules you would have to start 0-3 because you already played. (Or 0-1 and have to win 2 matches). These rules just make no sense to me and I also don't think they increase the excitement of the event enough to justify such unfairness.
With unfairness you can argue that it is fair because it is the same for everyone, and this is obviously true, but that is just obvious stuff. When people say it is unfair it's just that in the isolated event of this happening to someone he is put at a disadvantage against the players who don't have to go through the extended series. In that sense it is unfair.
Let me preface by saying that I totally agree that it doesn't make the series as exciting. If the tournament is mostly for spectators and the primary goal is to be exciting, then I agree it's a silly rule. It makes series almost prohibitively difficult to win in some circumstances, which ruins the tension that you'd normally get.
However, I disagree (still) when people say that it's competitively unfair (that is, it is unfair when trying to determine the best player). Let's conceive of this situation: you play 50 games, and you win 30 of them. The odds of this happening if you are evenly matched with your opponent are 0.1013194. Fairly unlikely (not as bad as it could be, but still fairly unlikely). Therefore, there's a very good chance that you are a better player than your opponent. Now, split those wins into BoX series, and depending on the order of wins/losses, it is very possible for them to hide the fact that one player is better than the other (for example, you can split 40 Bo3 series, where the better player wins 20 with 2-0 margins, and the lesser player wins 20 with 2-1 margins). Using this data, you can mask the fact that one player is better than the other, even though it's fairly obvious with the individual game data that he is better.
Doing a separate, isolated series in the loser bracket essentially punishes the better player for losing a smaller number of games than he wins for simply playing them in the wrong order. Now, some of this comes down to the philosophy of how much you think games matter and sort of the concept of being "clutch" (that is, do games later in the series matter more and therefore are they more important and more indicative of skill?). Personally, I think this is hogwash, unless someone can point to a good amount of peer reviewed evidence that can support that claim.
I'm not necessarily arguing for the mlg rule, as I have stated earlier, I think it makes the series less exciting, but I feel a lot of people are not objectively analyzing what is going on, and are thinking with their hearts rather than with their heads.
|
On November 10 2010 12:35 dcemuser wrote:Show nested quote +On November 10 2010 12:31 Liquid`Nazgul wrote:
Imagine playing the world cup of football losing 0-3 to someone in group stage and meeting them in the finals again. According to the rules you would have to start 0-3 because you already played. (Or 0-1 and have to win 2 matches). These rules just make no sense to me and I also don't think they increase the excitement of the event enough to justify such unfairness.
With unfairness you can argue that it is fair because it is the same for everyone, and this is obviously true, but that is just obvious stuff. When people say it is unfair it's just that in the isolated event of this happening to someone he is put at a disadvantage against the players who don't have to go through the extended series. In that sense it is unfair. Exactly, Nazgul. And when you have a player like EG.iNkA come out and say 'This rule is bullshit.' AFTER he came back from a 0-2 deficit to win 4-3, you know it is pretty bad. A lesser player would say something like "No, it's pretty fine. I did it!" He had the balls to say, 'No, this is fucked up, and I say that even after I've done it'.
Another point I'd like to make is that just because a pro says it, doesn't make it true. I appreciate how they are able to play at a very high level, but that doesn't make their critical thinking ability any more accurate. It really carries no weight at an intellectual level.
|
8748 Posts
On November 10 2010 12:48 Siffer wrote:
The problem with MLG is that they aren't considering series to be a set of isolated games.
If I beat a player in the first round of a tournament 2-1, I won that series. If I happen to get knocked into the losers bracket later on in the tournament and my first round opponent fights his way all the way though the lower bracket to face me again, he earned the right to play me again in another isolated set of games.
Extended series really makes no sense, and I don't understand how anyone can agree to it. Even for halo, it is a terrible ruleset. I can not think of another organization that follows the extended series format. If a player gets 0-2'd into the losers bracket and then has to face his original opponent again, he essentially has to eliminate the OG player TWICE(4 wins). That is triple elimination.
Also, what happens if say the finals are a BO5. HuK vs IdrA play first round of a tournament. Idra wins 2-0, but HuK fights all the way through the losers bracket to face IdrA again. Now what? idra up 2 games in a what? Best of 9? 11? lol?
It's not really fair to say that you've identified the problem with MLG. They are perfectly aware of the perspective you have just outlined and they choose to have extended series. If you can't understand extended series ("[they] make no sense") then I think it's a little bit silly for you to be discussing this, since MLG understands both sides and you only understand your own side.
I think you first need to resist the urge to think too much about how double elimination tournaments are run but instead think about why double elimination is favored over single elimination. It's a format for playing more games so that results reflect performance more accurately.
There are two ways for results to kind of get messed up in a tournament.
The first way is you get unlucky with the people you have to play against. If you are scheduled to play the best player in the tournament the first round, and you are the second best player in the tournament, and he eliminates you, then the results will show you placed 65th-128th, and that's really not very accurate. Double elimination fixes that. With double elim, he knocks you out, you beat everyone you play against after that and win the losers bracket, while he beats everyone he plays and he wins the winners bracket, and you place 2nd. It's not a perfect solution, since you can still get unlucky if you are 3rd best or worse, but it certainly helps.
But what happens when you lose to a guy and then end up meeting him in the losers bracket in a double elimination tournament? He's already proven himself to be better than you. It's a "dividing by zero" kind of moment because the purpose of the losers bracket kind of disappears and your reason for still being in the tournament is kind of gone. So, extended series! The second way that tournament results can be inaccurate is when the series are just too short. If you win 55% of your games against someone, he still has a pretty damn good chance to win a bo3, but the longer the series, the lower that chance. You played him in a bo3 but maybe your loss was a fluke so let's extend the series and see if you beat this guy in a bo7.
That's one way extended series make sense. The other way is... he beats you 2-0 in the winners bracket, you beat him 2-1 in the losers bracket, and now you're 2-3 against the guy, but you're determined to be the better player because... why? The arbitrary order of the games? That doesn't make any sense. Your only explanation for why that does make sense is because it's a strict following of the double elimination format which is the solution for a different problem. Extended series fixes one (or two?) of the flaws of a straight double elimination tournament.
|
Tyler's last paragraph explains perfectly why I like the extended series rule. How it works is a little weird, but it prevents you from eliminating someone from the tournament despite having a losing record against him.
|
8748 Posts
So basically, if you lost to a guy and you're in the losers bracket, this is where you're at: You have a chance to show how good you are against everyone else in the tournament. You recognize that you've already been bested by that one guy. If it happens that the brackets set it up so you play that guy again, you have a chance to show that you're actually better than that guy. Tying that guy 3-3 or 2-2 (1-2, 2-1 or 0-2, 2-0), or getting super lucky with game order and actually going 2-3 against him (0-2, 2-1), is NOT showing you're better than that guy.
|
so THAT's the meaning of double elimination tournament
|
Damn, Tyler makes too much sense!
|
On November 10 2010 13:20 Liquid`Tyler wrote:
So basically, if you lost to a guy and you're in the losers bracket, this is where you're at: You have a chance to show how good you are against everyone else in the tournament. You recognize that you've already been bested by that one guy. If it happens that the brackets set it up so you play that guy again, you have a chance to show that you're actually better than that guy. Tying that guy 3-3 or 2-2 (1-2, 2-1 or 0-2, 2-0), or getting super lucky with game order and actually going 2-3 against him (0-2, 2-1), is NOT showing you're better than that guy.
Yep, people just need to realize that in that tourney all that matters is your performance against the people you are specifically matched up with in the brackets. Third party performance is irrelevant to each specific match.
|
On November 10 2010 13:20 Liquid`Tyler wrote:
So basically, if you lost to a guy and you're in the losers bracket, this is where you're at: You have a chance to show how good you are against everyone else in the tournament. You recognize that you've already been bested by that one guy. If it happens that the brackets set it up so you play that guy again, you have a chance to show that you're actually better than that guy. Tying that guy 3-3 or 2-2 (1-2, 2-1 or 0-2, 2-0), or getting super lucky with game order and actually going 2-3 against him (0-2, 2-1), is NOT showing you're better than that guy.
Which is all good and great, if extended series just fixed this it would be cool, but unfortunately it has side-effects that are not so cool. Because it actually introduces bracket-luck-based unfairness, complicates the rules for very little benefit. Actually much "fairer" solution to the problems that extended series solves would be extended series paired with dynamic (but strictly defined) pairing of people in the losers bracket. That would eliminate luck introduced by standalone extended series, you would basically pair people only with those they did not played before as long as possible, that means that extended series would come into effect only close to LB finals and it would come to effect for all players nearly simultaneously. But it has its own problems, even more complexity, worse spectator experience and it still does not fix the unfairness completely just lowers it.
|
On November 10 2010 19:57 mcc wrote:Show nested quote +On November 10 2010 13:20 Liquid`Tyler wrote:
So basically, if you lost to a guy and you're in the losers bracket, this is where you're at: You have a chance to show how good you are against everyone else in the tournament. You recognize that you've already been bested by that one guy. If it happens that the brackets set it up so you play that guy again, you have a chance to show that you're actually better than that guy. Tying that guy 3-3 or 2-2 (1-2, 2-1 or 0-2, 2-0), or getting super lucky with game order and actually going 2-3 against him (0-2, 2-1), is NOT showing you're better than that guy. Which is all good and great, if extended series just fixed this it would be cool, but unfortunately it has side-effects that are not so cool. Because it actually introduces bracket-luck-based unfairness, complicates the rules for very little benefit. Actually much "fairer" solution to the problems that extended series solves would be extended series paired with dynamic (but strictly defined) pairing of people in the losers bracket. That would eliminate luck introduced by standalone extended series, you would basically pair people only with those they did not played before as long as possible, that means that extended series would come into effect only close to LB finals and it would come to effect for all players nearly simultaneously. But it has its own problems, even more complexity, worse spectator experience and it still does not fix the unfairness completely just lowers it.
Extended series more accurately determines the stronger player between 2... as unfair as it is to even have to play an extended series where others may not have to... it is only slightly less unfair to get "unlucky" and play a person that beat you in a second bo3 running into him in the loser's bracket anyway.
That little bit of unfairness, which actually helps to determine who is the stronger player of the two, is worth it because it makes for a more accurate tourney, so that no player can get elim'd by a player despite having a winning record vs him.
|
Also treating maps (as opposed to a series) as a unit of performance is not good if players can choose maps, because in that case maybe player's A loss 2:1 in the WB was caused by bad luck with first map, which as this tourney showed is a very big problem. For example PvZ player A(Zerg) loses first map on Steppes then wins on Scrap then loses on Kulas  In LB he wins on Metalopolis and Delta Q, I think he proved enough in that case
|
I dont get how people such as incontrol can say that facing people in tournaments is an isolated event and then talk about seeding as if its necessary. Surely if every game is an isolated event you dont need seeding cos the winner will arise anyway. Which is the point of a tournament anyways yes?
But MLG is not looking to make a string of isolated open tournaments with every game and tourney isolated. Its obvious that they want to tour the country. They want regular appearances to make seeding matter.
I don't necessarily agree that extended series or even double elimination is the best way to run a tournament or that even bo3 is good for tournaments. But I do think you should have an advantage by staying in the winners bracket.
|
On November 10 2010 13:10 Liquid`Tyler wrote:
But what happens when you lose to a guy and then end up meeting him in the losers bracket in a double elimination tournament? He's already proven himself to be better than you.
He's proven nothing of the sort. All he's done is take 2 games off you in a single BO3. The simple fact that sometimes the player who loses the first game goes on to take the BO7 disproves your statement. How many games have you played with your practice partners where one of you has gone 3, 5, or 7 games in a row at some point? Does that prove that one of you is better? Of course it doesn't. That sort of thing can only be proven by consistent performance over time.
Also, if the purpose of double elimination in general is to allow more consistently-good performance to result in better tournament result, and if the purpose of the extended series is to ensure that earlier good performance is rewarded when you get knocked down to the losers' bracket, then shouldn't those players who perform well in the winners' bracket get the same advantage (or lack of it) when they get knocked down to the losers' bracket, regardless of whether they're playing somebody for the first or second time?
Say IdrA beats HuK in round 1 of the winners' bracket. Then IdrA loses to iNcontroL in round 2. IdrA then meets Huk again in the losers' bracket, with HuK winning the BO7 extended series. iNcontroL then loses to Kiwikaki in round 3 of the winners' bracket. Why should IdrA be rewarded for performing well, but iNcontroL is not rewarded despite performing even better (assuming we're taking wins in the winners' bracket as a measure of good performance)?
It seems to me that extended series, while well-intentioned, don't reward good performances consistently, whereas a 1-0 headstart in a BO5, given to the player who got further in the winners' bracket, would achieve this goal, and would also result in nobody being eliminated despite being up on games against their opponent.
If it happens that the brackets set it up so you play that guy again, you have a chance to show that you're actually better than that guy. Tying that guy 3-3 ... is NOT showing you're better than that guy.
True, but it's also NOT showing that you're worse than him, and the simple fact at that point is that you're 3-3 against each other, but he's lost two series compared to your one, and he's blown a 1-0 headstart in the 'rematch'. In the confines of the tournament, there's no pressing reason for him to remain in.
|
I thought double elim tournaments went like this:
Final 2 winner bracket players play
Loser faces winner of Losers bracket
Winner of that match faces winner of winners bracket (finals)
|
I've been to MLG since 2005 playing Melee. There's been debates since on that validity of extended series on the tournaments. I think for Melee, the extended series worked just fine, players in particular didn't have a huge problem with it. The reason for this is that I think games are more stable in fighting games or FPS's.
When you have a Bo3 in a SC2 match where each match can be very unstable with cheeses and the lower player can advance much more easily than a Melee match, then I think an extended series starts coming into question where "is player A who beat player B really better than player B?" Is it ok to just extend the series? I'm not so sure about that.
|
SO if you think someone in your group might go the distance, dropping a game to someone you know you can beat and taking the looser bracket route before you meet anyone that might be challenging could be considered a valid decition?
That way nobody you face will ever have this sort of shitty rules applied to them? Atleast thats what id do if what i hear is correct.
Basically this logic emerges very fast when it all comes down to this:
DO you want to play an extended series where you are already down 2 games vs people who are fucking amazing at this game? or do you want to play them straight up with no disadvantage on the table.
Anyone that are not playing vs idra etc. should consider tossing their first games and go for the looser bracket Select style. The farther you go into the winners bracket the tighter the noose tightens against your neck.
|
8748 Posts
On November 10 2010 20:18 The Touch wrote:Show nested quote +On November 10 2010 13:10 Liquid`Tyler wrote:
But what happens when you lose to a guy and then end up meeting him in the losers bracket in a double elimination tournament? He's already proven himself to be better than you. He's proven nothing of the sort. All he's done is take 2 games off you in a single BO3. The simple fact that sometimes the player who loses the first game goes on to take the BO7 disproves your statement. How many games have you played with your practice partners where one of you has gone 3, 5, or 7 games in a row at some point? Does that prove that one of you is better? Of course it doesn't. That sort of thing can only be proven by consistent performance over time.
It's proven to the bracket. Of course I understand the limitations of a bo3 when it comes to knowing the players' skills. But when playing in a bracket, the guy who wins the match, whether it's bo1 or bo7, is considered the better person by the bracket, so he advances. For single elim, the bracket says "ok you lost to this guy in a match, losing to him means you're worse than him, only one of you can go on so he does, you're out" and for double elim the bracket says "you lost to this guy, you're worse than him, but I want to see if you're better than anyone else, go play in the losers bracket"
On November 10 2010 20:18 The Touch wrote:
Also, if the purpose of double elimination in general is to allow more consistently-good performance to result in better tournament result, and if the purpose of the extended series is to ensure that earlier good performance is rewarded when you get knocked down to the losers' bracket, then shouldn't those players who perform well in the winners' bracket get the same advantage (or lack of it) when they get knocked down to the losers' bracket, regardless of whether they're playing somebody for the first or second time?
The extended series doesn't need to reward earlier performance to have a purpose... how is playing a regular bo7 rewarding earlier performance? The extended series throws away any bias about which games came earlier or later (something that a 2nd bo3 doesn't do) and allows you to play a straight up bo7. The bracket just wants to know which one of you is truly better and I think from your first paragraph we agree that a bo7 is better than a bo3 for that.
Looking at double elim without extended series, the situation is like this: if the guy who won the first time is going to win again, he's going to need 4 total wins. He is always going to have to win a bo7 between himself and his opponent if he wants to survive. His opponent can advance by losing a bo5, or by tying 2-2 or 3-3. So when MLG speaks of "why should we punish the guy who won in the winners bracket" this is what they're talking about. So they put it on fair terms by making both of them have to win a bo7 to advance. They are not giving the guy who won in winners bracket an advantage at all. They're taking away the advantage of the guy who lost in winners bracket and making it fair.
On November 10 2010 20:18 The Touch wrote:
Say IdrA beats HuK in round 1 of the winners' bracket. Then IdrA loses to iNcontroL in round 2. IdrA then meets Huk again in the losers' bracket, with HuK winning the BO7 extended series. iNcontroL then loses to Kiwikaki in round 3 of the winners' bracket. Why should IdrA be rewarded for performing well, but iNcontroL is not rewarded despite performing even better (assuming we're taking wins in the winners' bracket as a measure of good performance)?
IdrA isn't being rewarded here. iNcontroL is being rewarded. When incontrol loses to kiwikaki, he's put into round 4 of the losers bracket. Idra has been in the losers bracket since round 1. By winning 2 more rounds in the winners bracket than idra, incontrol plays 3 less elimination rounds in the losers bracket than idra.
On November 10 2010 20:18 The Touch wrote:
It seems to me that extended series, while well-intentioned, don't reward good performances consistently, whereas a 1-0 headstart in a BO5, given to the player who got further in the winners' bracket, would achieve this goal, and would also result in nobody being eliminated despite being up on games against their opponent.
Avoiding playing a bunch of elimination rounds is the reward for staying in the winners bracket. That is actually a related reward. Giving someone a free win in a particular series for staying in the winners bracket is a contrived bonus reward that is totally undeserved.
On November 10 2010 20:18 The Touch wrote:Show nested quote +If it happens that the brackets set it up so you play that guy again, you have a chance to show that you're actually better than that guy. Tying that guy 3-3 ... is NOT showing you're better than that guy. True, but it's also NOT showing that you're worse than him, and the simple fact at that point is that you're 3-3 against each other, but he's lost two series compared to your one, and he's blown a 1-0 headstart in the 'rematch'. In the confines of the tournament, there's no pressing reason for him to remain in.
There is no 1-0 head start, neither with extended series nor without them. Where do you get that from? It's two bo3's. The first one determines who gets knocked down to the losers bracket. The second one determines who gets to stay in the losers bracket and who gets eliminated. Winning the first bo3 doesn't make you 1-0 in bo3's. That stat doesn't matter at any point. Extended series is a clever way to make the deciding match a bo7 instead of a bo3 which is something we agree is better for deciding who the better player is.
|
At first I didn't like the idea of extended series, but now I'm sort of warming to the idea if it's done properly.
Winners Round 1 - Nadagast vs. ZpuX
Game 1 - Xel'Naga Caverns
Game 2 - Lost Temple
Game 3 - Steppes of War
Losers Round 4 - Nadagast vs. ZpuX - Extended Series
Game 4 - Xel'Naga Caverns (WTF? they already played on that)
Game 5 - Metalopolis
Game 6 - Lost Temple (Again, they already played this)
Winners Round 1 - Sung123 vs. Mamerhet
Game 1 - Xel'Naga Caverns
Game 2 - Metalopolis
Game 3 - Steppes of War
Losers Round 4 - Sung123 vs. Mamerhet - Extended Series
Game 4 - Xel'Naga Caverns (as above)
Game 5 - Steppes of War (as above)
Game 6 - Delta Quadrant
Winners Round 1 - AhhBoxxah vs. Agh
Game 1 - Xel'Naga Caverns
Game 2 - Metalopolis
Game 3 - Lost Temple
Losers Round 4 - AhhBoxxah vs. Agh - Extended Series
Game 4 - Xel'Naga Caverns (as above)
Game 5 - Metalopolis (as above)
Game 6 - Lost Temple (as above)
Game 7 - Blistering Sands
Winners Round 2 - Fenix vs. InKa
Game 1 - Lost Temple
Game 2 - Xel'Naga Caverns
Losers Round 6 - Fenix vs. InKa - Extended Series
Game 3 - Scrap Station
Game 4 - Metalopolis
Game 5 - Steppes of War
Game 6 - Desert Oasis
Game 7 - Lost Temple (as above)
BTW grats to InKa for coming back from 0-2 down to take the series 4-3.
Looking at the replays, the first map of the extended series isn't chosen by the loser of the match in the upper bracket? Why not? If it's supposed to be a follow on from the previous games shouldn't the player who lost the last game be selecting a map remaining in the map pool?
Another thing is I'm not sure if the same map exclusion by the players in the first match is maintained in the extended series, if not, why not?
|
On November 11 2010 04:18 Liquid`Tyler wrote:
The extended series doesn't need to reward earlier performance to have a purpose... how is playing a regular bo7 rewarding earlier performance?
It just seems to me that this is exactly what it does. You get an easier time in the losers' bracket compared to other players, based on your earlier performance in the winners bracket. I know that's not the intention of the rule, but I think that's the actual result.
Your earlier performance is rewarded through you having the luxury of having to lose 3 or 4 games in the losers' bracket in order to be knocked out, compared to other players only having to lose 2, even if they did as well or better than you in the winners' bracket. That's an advantage you have over them, and it's based on how you performed earlier in the tournament. The advantage isn't over your oppoenent, but over players in other matches.
I'm not saying that the extended series is undeserved - I absolutely agree that it's counterintuitive to be knocked out by somebody you have a winning record against during the tournament. But if good performance in earlier rounds is going to be rewarded, I think it should be done so as a general rule, rather than only if you happen to play somebody you've already beaten.
The extended series throws away any bias about which games came earlier or later (something that a 2nd bo3 doesn't do) and allows you to play a straight up bo7. The bracket just wants to know which one of you is truly better and I think from your first paragraph we agree that a bo7 is better than a bo3 for that.
Yes, absolutely. The extended BO7 is definitely better than a second BO3 in that respect. I just think it throws up a problem of its own as well.
There is no 1-0 head start, neither with extended series nor without them. Where do you get that from?
Sorry, it was from my own suggestion - I saw your 3-3 and my brain made the connection. I was saying I think it'd be a good idea to have whoever got further in the winners' bracket have a 1-0 headstart in a single BO5 when they meet in the losers' bracket. Better performances in the winners' bracket would result in a consistently applied advantage in the losers' bracket. The worst that could happen is you gettting knocked out by somebody you went 3-3 with overall (2-0 in winners' and 1-3 in losers'). You will have lost two series and your opponent only one.
My goal was a system that still didn't result in you being knocked out by people you had a winning record against, while at the same time applying a consistent reward for winners' brackets performance. I'm sure such a system would have its own flaws, but I would personally prefer it to the extended series rule, from a spectator's point of view.
|
|
|
|
|
|
EPT
