|
8748 Posts
On November 11 2010 04:34 Jibba wrote:
Their point is that even if it may be best, the results are so far from the truth that it shouldn't be packaged as a ranking. It's comparing 60% accuracy to 40% accuracy.
And yet MLG ranks its player 1-16 and, when combining tournaments, goes beyond even that for seeding. Prize money is significant for ranks 1-8. I don't know what to call it other than ranking.
My point, which I guess they missed, and which randplaty just picked up on a bit, is that the obvious and ostentatious purpose of any competition is to rank the performers as best as possible. When choosing format and rules, other considerations come into play like time, resources, what the spectators want, what the players want, etc, but you are always trying to maximize the accuracy of rankings within the restraints of all those other things.
What would suck is if there's a rule that increases accuracy of rankings but gets removed because players and spectators think it decreases accuracy of rankings (or because they don't care about the increased accuracy and they have an unjustified bias against the rule).
|
should decreasing variance be a real goal in tournaments? i don't know.
|
On November 11 2010 04:37 maliceee wrote:Show nested quote +On November 11 2010 02:57 NoXious90 wrote:On November 11 2010 01:21 maliceee wrote:On November 11 2010 00:47 NoXious90 wrote:On November 11 2010 00:40 robertdinh wrote:On November 11 2010 00:38 robertdinh wrote:On November 11 2010 00:30 Mothxal wrote:On November 11 2010 00:00 Liquid`Tyler wrote:On November 10 2010 23:31 Mothxal wrote:
You can't just say that there is one instance where single-elimination creates results that badly reflect the "real" skill-hierarchy, and that's when two higher ranked players face eachother early on. This can be fixed with seeding anyway. It is absolutely ridiculous to be okay with seeding and not with extended series. My mind is exploding about this contradiction. Seeding is justified along the same lines that extended series are except they're way more extreme and harder to defend. You want results from entirely different tournaments, played months before, against different sets of players, to give advantages in this tournament via seeding, but you don't want Person A who beat Person B in THIS tournament no more than ~50 hours ago to get any credit when playing Person B again. Wowowow whaaaaaat. Anyway, I haven't limited myself to the problems with single elimination I've mentioned. There could be a million problems with single elimination and millions of problems with brackets in general. I don't have to be comprehensive here. My point is that double elim fixes a problem with single elim and extended series fixes a problem with double elim. Explaining how double elim works as a response to why having extended series is bad just doesn't make sense. It's not a good response. Surprisingly enough there aren't angry polls about how we must destroy the menace of seeding players. There is such a thing as the integrity of a tournament; basically, if it is seen as a credible event by the players and spectators whose results reflect "something". It isn't possible to accurately measure someone's relative skill in the timespan of a tournament, yet winning them still has to be valuable and an accomplishment, and that can only be achieved if winning when it matters is worthwhile. If you start to obsess about the most truthful measuring system in such a short period then you're chasing something you can never accomplish anyway. If this now starts to be a higher priority than the actual place some players have in the brackets and such, then, as can be seen from the viewer reaction, people will start to find it ridiculous. Seeding on the other hand is a relatively harmless approach to trying to get the results more reflective of real skill, because the tournament's integrity isn't compromised in any way. The rules are still exactly the same, it's just the matches that are somewhat more balanced now. It doesn't matter what they are based on, as long as it's transparent and impartial (and based on past results of course). Even if they don't seem to help a lot, just helping a bit is still helpful and doesn't negatively impact anything of importance. The tournaments integrity is stronger with extended series than with straight double elim. Since a bo7 is the best way to determine the stronger of 2 players, over 2 bo3's. The reason it is met with such opposition is people don't truly understand it's purpose, or they have a skewed perspective of what tournament competition should be. Then there are a few who don't think players that lose games in the winner's bracket should be accountable for those losses. But as tyler sorta mentioned, if you want to take the clean-slate-in-the-loser-bracket approach you should also oppose the whole seeding system since that isn't a clean slate and gives certain players advantages based on how they placed in a previous tourney. On November 11 2010 00:36 NoXious90 wrote:
Amidst all this debate about the true higher 'purpose' of tournaments as indicators of player skill or whatever, you're overlooking one very simple thing. Tournaments are simply spectacles, forms of entertainment for people to watch and enjoy. For players, tournaments provide an opportunity for players to compete against other high level players whilst offering them a chance to win prize money.
The bottom line is, the extended series rule has an adverse impact on the entertainment value of the tournament that greatly outweights whatever insurance it provides against 'inferior' players beating 'superior' players through luck or some other perceived illegitimate method of victory. This rule is especially detrimental when it comes into play during the grand final, which is supposed to be the culmination of the entire tournament - where the two best players of the tournament face off against one another to decide the ultimate winner. If you have a grand final which begins with one player having a significant advantage over the other, the spectacle of such a match is greatly reduced.
To put it plainly the people who object to extended series are the ones that are going to be most vocal about it. There are plenty of spectators who: 1. Would prefer a tournament with as much integrity as possible 2. Don't care enough either way and just want to kick back and casually watch some SC2 and old spice commercials. The so-called integrity that would be lost if the tournament didn't follow the extended series rule would be insignificant. The GSL doesn't have an extended series rule, neither does the NFL, nor the World Cup. They all seem to do fine as far as perceived legitimacy goes, and more importantly, provide amazing spectacles which any fan worth their salt will want to see. Nfl has a regular season that establishes seeds, then has a playoff that cannot be a best of series because of physical limitations. It is one of the reasons college football does not enact playoffs or a plus 1 game, injuries plus time off from school would be a bitch to get enacted. Not only that, people question the NFL champion's legitimacy all the time, lol, who honestly thought the NY giants had a better team then the patriots? World cup has pool play followed by the final's games, but those games are also contrained by time and budget. If you knew what you were talking about you would know almost all professional football leagues have a best of series at home and away and how much you win by and score matters. So that example is shit honestly, because a team could win 3-1 at home then lose 2-0 away and they would not progress. The reason they can do that is because there aren't as many constraints as in the world cup. so, let me ask you this (because this is what my point boils down to), what do you think the majority of people would enjoy more as a spectacle, a grand final featuring two players who have to play a Bo7 to decide the winner, where the score starts out at 0-0 or a grand final featuring two players where one player goes into the match with a two game advantage? I know which one would be more exciting for me personally. as far as football goes, yes, you're correct that many round robin/knockout tournaments use an aggregate scoring format fairly similar to the one used at mlg, however, this is never ever used in the final match, not in the uefa cup, champions league or any other major tournament of that nature. This is because it would risk severely diluting the spectacle of the grand final, where the two best teams are supposed to play off in an epic match to decide the ultimate winner. Having this aggregate rule in place dampens the luster of the final match. see: idra vs select, ttone vs jinro for evidence of this. Again, there is reasoning behind them not playing a final over multiple days in the champions league or w/e, and that reason is not applicable to starcraft. In Starcraft the extended series can be played in one location at one time, in a reasonable amount of time.
Why don't they though? ask yourself that question. Why do they play the final at a neutral ground and only play one match? they could just as easily play the final over two legs at each of the team's home stadiums like they do in the Ro16, quarterfinals, semifinals,etc, but they don't. Why?
|
United States22883 Posts
On November 11 2010 04:27 randplaty wrote:I'm not saying that Tyler is correct, but InControl/Idra and even Day9's argument about tournament theory was really bad. Merely because determining the best player from one tournament is impossible, does not mean that the tournament does not even try. The tournament should still try to make the winner = the best player even if that goal is impossible. Their argument is "because it is impossible, we shouldn't try." Day9's argument is just that mathematically it is not likely that the winner will be the best player. Tyler is absolutely correct in saying that a tournament should try to determine the best player. If it is not to determine the best player, we might as well play the lottery to determine the winner. Whether or not extended series does this is still in question, but that should be the question. Does extended series help to determine the best player within the context of a double elimination tournament. There are mathematical ways to calculate this and determine this. It is not impossible to figure out if extended series helps determine the best player or not. I have not read any of any studies that studied extended series in a tournament format, but there are papers that are written about tournament formats and which tournament formats. This study http://www.cerge-ei.cz/pdf/wp/Wp303.pdf shows that the round robin tournament format is the best at determining the best player but requires the most time and resources. Second is double elimination, and the worst is a single elimination or "contest" format. So studies have been done and can be done to mathematically figure out which is the best tournament format and any discussion about tournament format that does not include this mathematical analysis is incomplete.
That takes a very sterile view of competition. Are rounds 1 and 12 weighted equally because the same players are involved?
Since the tournament is double elim, that presents a perfect example. The best player in the world faces the second best player in the world in round 2. Because they know they're better than everyone else in the tournament, neither of them is willing to unveil special strategies because they both have a good chance at running through the losers bracket, in which case they will meet again in the finals. If one of those players uses those strategies in the first round while the other doesn't, that player will win. Unfortunately, their strategies have been revealed and when the players meet again in the finals, the player that lost initially will have an edge up having not revealed their play.
That may be a tidy little hypothetical example, but players hide their builds all the time, or they get extra preparation and ideas for later rounds. Since later rounds are more important, the two sets of matches cannot be combined as if it's a single bo7 series, because it's not. It's a less important bo3 and a more important bo3.
|
Top 16 @ MLG .. a few matches were forfeits or coin flips from 9-16.
1. Liquid`Jinro
2. FnaticTTOne
3. Pain.User
4. Liquid`Tyler
5. Liquid`TLO
6. EG.Machine
7. Liquid`Ret
8. ROOT.Drewbie
9. LiquidHuK
10. ROOT KiwiKaki
11. Dignitas SeleCT
12. ROOT SLusH
13. ROOTqxc
14. EGiNka
15. EGiNcontroL
16. aTn Socke
to repost..
|
On November 11 2010 04:44 Mereel wrote:
i have a short question about all the artosis bashing.
so on the liveon3 show everyone bitched over him with autism and whatnot, idra said he is retarded ( i thought there are friends?) and now on the sotg show nony bitched over him and later on incontrol too.
so is that all trolling or what.....i dont really understand whats going on
Artosis has frequently said and done some very puzzling things over the course of his time in the SC community. Most or maybe even all of these things have been benign or silly, but damn does he make a good target/source of humor 
|
iNcontroL
USA29055 Posts
easily the most boring conversation in this thread post a SOTG
I need to work harder at offending people 
|
On November 11 2010 04:50 trancey wrote:
Top 16 @ MLG .. a few matches were forfeits or coin flips from 9-16.
1. Liquid`Jinro
2. FnaticTTOne
3. Pain.User
4. Liquid`Tyler
5. Liquid`TLO
6. EG.Machine
7. Liquid`Ret
8. ROOT.Drewbie
9. LiquidHuK
10. ROOT KiwiKaki
11. Dignitas SeleCT
12. ROOT SLusH
13. ROOTqxc
14. EGiNka
15. EGiNcontroL
16. aTn Socke
to repost..
Liquid invasion of the top 10 indeed  Tyler the oracle
|
On November 11 2010 04:52 {88}iNcontroL wrote:easily the most boring conversation in this thread post a SOTG I need to work harder at offending people
Your A Nazi?!!?! How dare you incontrol!! They hurt many people!!!
This podcast gone too far!!
|
On November 11 2010 04:51 theqat wrote:Show nested quote +On November 11 2010 04:44 Mereel wrote:
i have a short question about all the artosis bashing.
so on the liveon3 show everyone bitched over him with autism and whatnot, idra said he is retarded ( i thought there are friends?) and now on the sotg show nony bitched over him and later on incontrol too.
so is that all trolling or what.....i dont really understand whats going on Artosis has frequently said and done some very puzzling things over the course of his time in the SC community. Most or maybe even all of these things have been benign or silly, but damn does he make a good target/source of humor
so what did he said...
|
United States22883 Posts
On November 11 2010 04:45 Liquid`Tyler wrote:Show nested quote +On November 11 2010 04:34 Jibba wrote:
Their point is that even if it may be best, the results are so far from the truth that it shouldn't be packaged as a ranking. It's comparing 60% accuracy to 40% accuracy. And yet MLG ranks its player 1-16 and, when combining tournaments, goes beyond even that for seeding. Prize money is significant for ranks 1-8. I don't know what to call it other than ranking. My point, which I guess they missed, and which randplaty just picked up on a bit, is that the obvious and ostentatious purpose of any competition is to rank the performers as best as possible. When choosing format and rules, other considerations come into play like time, resources, what the spectators want, what the players want, etc, but you are always trying to maximize the accuracy of rankings within the restraints of all those other things. What would suck is if there's a rule that increases accuracy of rankings but gets removed because players and spectators think it decreases accuracy of rankings (or because they don't care about the increased accuracy and they have an unjustified bias against the rule). It's ranking the top performers of that specific tournament, not the top players at any given time. Being the best player doesn't necessarily mean you will have the best performance.
MLG is a league that may have an end goal of ranking according to skill, but it does so through a composite of all the performance rankings.
What is the disclaimer made whenever KeSPA rankings come out? It's not a ranking of skill, it's a ranking of performance over the past 3 months. The two might be similar but they are distinct.
|
On November 11 2010 04:46 NoXious90 wrote:Show nested quote +On November 11 2010 04:37 maliceee wrote:On November 11 2010 02:57 NoXious90 wrote:On November 11 2010 01:21 maliceee wrote:On November 11 2010 00:47 NoXious90 wrote:On November 11 2010 00:40 robertdinh wrote:On November 11 2010 00:38 robertdinh wrote:On November 11 2010 00:30 Mothxal wrote:On November 11 2010 00:00 Liquid`Tyler wrote:On November 10 2010 23:31 Mothxal wrote:
You can't just say that there is one instance where single-elimination creates results that badly reflect the "real" skill-hierarchy, and that's when two higher ranked players face eachother early on. This can be fixed with seeding anyway. It is absolutely ridiculous to be okay with seeding and not with extended series. My mind is exploding about this contradiction. Seeding is justified along the same lines that extended series are except they're way more extreme and harder to defend. You want results from entirely different tournaments, played months before, against different sets of players, to give advantages in this tournament via seeding, but you don't want Person A who beat Person B in THIS tournament no more than ~50 hours ago to get any credit when playing Person B again. Wowowow whaaaaaat. Anyway, I haven't limited myself to the problems with single elimination I've mentioned. There could be a million problems with single elimination and millions of problems with brackets in general. I don't have to be comprehensive here. My point is that double elim fixes a problem with single elim and extended series fixes a problem with double elim. Explaining how double elim works as a response to why having extended series is bad just doesn't make sense. It's not a good response. Surprisingly enough there aren't angry polls about how we must destroy the menace of seeding players. There is such a thing as the integrity of a tournament; basically, if it is seen as a credible event by the players and spectators whose results reflect "something". It isn't possible to accurately measure someone's relative skill in the timespan of a tournament, yet winning them still has to be valuable and an accomplishment, and that can only be achieved if winning when it matters is worthwhile. If you start to obsess about the most truthful measuring system in such a short period then you're chasing something you can never accomplish anyway. If this now starts to be a higher priority than the actual place some players have in the brackets and such, then, as can be seen from the viewer reaction, people will start to find it ridiculous. Seeding on the other hand is a relatively harmless approach to trying to get the results more reflective of real skill, because the tournament's integrity isn't compromised in any way. The rules are still exactly the same, it's just the matches that are somewhat more balanced now. It doesn't matter what they are based on, as long as it's transparent and impartial (and based on past results of course). Even if they don't seem to help a lot, just helping a bit is still helpful and doesn't negatively impact anything of importance. The tournaments integrity is stronger with extended series than with straight double elim. Since a bo7 is the best way to determine the stronger of 2 players, over 2 bo3's. The reason it is met with such opposition is people don't truly understand it's purpose, or they have a skewed perspective of what tournament competition should be. Then there are a few who don't think players that lose games in the winner's bracket should be accountable for those losses. But as tyler sorta mentioned, if you want to take the clean-slate-in-the-loser-bracket approach you should also oppose the whole seeding system since that isn't a clean slate and gives certain players advantages based on how they placed in a previous tourney. On November 11 2010 00:36 NoXious90 wrote:
Amidst all this debate about the true higher 'purpose' of tournaments as indicators of player skill or whatever, you're overlooking one very simple thing. Tournaments are simply spectacles, forms of entertainment for people to watch and enjoy. For players, tournaments provide an opportunity for players to compete against other high level players whilst offering them a chance to win prize money.
The bottom line is, the extended series rule has an adverse impact on the entertainment value of the tournament that greatly outweights whatever insurance it provides against 'inferior' players beating 'superior' players through luck or some other perceived illegitimate method of victory. This rule is especially detrimental when it comes into play during the grand final, which is supposed to be the culmination of the entire tournament - where the two best players of the tournament face off against one another to decide the ultimate winner. If you have a grand final which begins with one player having a significant advantage over the other, the spectacle of such a match is greatly reduced.
To put it plainly the people who object to extended series are the ones that are going to be most vocal about it. There are plenty of spectators who: 1. Would prefer a tournament with as much integrity as possible 2. Don't care enough either way and just want to kick back and casually watch some SC2 and old spice commercials. The so-called integrity that would be lost if the tournament didn't follow the extended series rule would be insignificant. The GSL doesn't have an extended series rule, neither does the NFL, nor the World Cup. They all seem to do fine as far as perceived legitimacy goes, and more importantly, provide amazing spectacles which any fan worth their salt will want to see. Nfl has a regular season that establishes seeds, then has a playoff that cannot be a best of series because of physical limitations. It is one of the reasons college football does not enact playoffs or a plus 1 game, injuries plus time off from school would be a bitch to get enacted. Not only that, people question the NFL champion's legitimacy all the time, lol, who honestly thought the NY giants had a better team then the patriots? World cup has pool play followed by the final's games, but those games are also contrained by time and budget. If you knew what you were talking about you would know almost all professional football leagues have a best of series at home and away and how much you win by and score matters. So that example is shit honestly, because a team could win 3-1 at home then lose 2-0 away and they would not progress. The reason they can do that is because there aren't as many constraints as in the world cup. so, let me ask you this (because this is what my point boils down to), what do you think the majority of people would enjoy more as a spectacle, a grand final featuring two players who have to play a Bo7 to decide the winner, where the score starts out at 0-0 or a grand final featuring two players where one player goes into the match with a two game advantage? I know which one would be more exciting for me personally. as far as football goes, yes, you're correct that many round robin/knockout tournaments use an aggregate scoring format fairly similar to the one used at mlg, however, this is never ever used in the final match, not in the uefa cup, champions league or any other major tournament of that nature. This is because it would risk severely diluting the spectacle of the grand final, where the two best teams are supposed to play off in an epic match to decide the ultimate winner. Having this aggregate rule in place dampens the luster of the final match. see: idra vs select, ttone vs jinro for evidence of this. Again, there is reasoning behind them not playing a final over multiple days in the champions league or w/e, and that reason is not applicable to starcraft. In Starcraft the extended series can be played in one location at one time, in a reasonable amount of time. Why don't they though? ask yourself that question. Why do they play the final at a neutral ground and only play one match? they could just as easily play the final over two legs at each of the team's home stadiums like they do in the Ro16, quarterfinals, semifinals,etc, but they don't. Why?
They want to maximize their profit on one game that can jack their sales to an obscene amount with less planning and money spent. In mlg, this does not matter because everything is there and they do not have to split games because of physical endurance. They can play it all on that day, their isn't a home field advantage or the like.
|
Listening to the extended series debate, an argument against extended series was that a tournament does not determine player rankings well anyway and should just determine the winner.
Going by this argument alone, then we should just run single eliminations instead of double eliminations since they equally determine just the winner, ignoring other rankings.
I'm not supporting the extended series format as I personally find it confusing for spectators, just wanted to highlight that the point argued in that perspective doesn't really sell the point.
|
On November 11 2010 04:59 Thoramas wrote:
Going by this argument alone, then we should just run single eliminations instead of double eliminations since they equally determine just the winner, ignoring other rankings.
A double elimination tournament has a better likelihood of having a stronger player as the winner. It's not just used for determining the other places better, and having more games to play.
|
Does Tyler live in Texas or NC? I'm confused : /
|
On November 11 2010 04:45 Liquid`Tyler wrote:Show nested quote +On November 11 2010 04:34 Jibba wrote:
Their point is that even if it may be best, the results are so far from the truth that it shouldn't be packaged as a ranking. It's comparing 60% accuracy to 40% accuracy. And yet MLG ranks its player 1-16 and, when combining tournaments, goes beyond even that for seeding. Prize money is significant for ranks 1-8. I don't know what to call it other than ranking. My point, which I guess they missed, and which randplaty just picked up on a bit, is that the obvious and ostentatious purpose of any competition is to rank the performers as best as possible. When choosing format and rules, other considerations come into play like time, resources, what the spectators want, what the players want, etc, but you are always trying to maximize the accuracy of rankings within the restraints of all those other things. What would suck is if there's a rule that increases accuracy of rankings but gets removed because players and spectators think it decreases accuracy of rankings (or because they don't care about the increased accuracy and they have an unjustified bias against the rule).
Are you satisfied then that jinro and TTOne are the absolute two best players in North America? Because I'm not, and I don't think many others would be either. It is so unbelievably unreasonable and unrealistic to attempt to infer perfect or even approximate rankings from a tournament that took place over three days, regardless of whether the extended rule is in place or not. This fact is, the tournament that took place in dallas over the weekend was an isolated incident, the circumstances of which will never be able to be perfectly replicated again, therefore you cannot reasonably suggest that you would get the same results if you held another 128 man tournament with the same players. This is a logical fact which something so immaterial as having an extended series rule in place could never under any circumstances hope to overcome. Therefore, by that fact alone, any attempts to strive at some universal settlement regarding a precise hierarchy of the skill levels of a group of players is futile and pointless, especially when pursuing such a thing involves sacrificing the entertainment value of the competition, which is THE REAL 'purpose' of a tournament such as this.
|
On November 11 2010 04:27 randplaty wrote:
If it is not to determine the best player, we might as well play the lottery to determine the winner.
This is really pretty silly. There have been quite a few big tournaments now (GSL, IEM, MLG), none of these had repeat winners even though they've had repeat competitors? Are they failing to determine the best player? Is this not even measurable? I think the extended series makes sense for the finals, but even there I'm uneasy:
-Player C gets knocked down to the losers bracket by Player B 2-0.
-Player B then gets knocked down to the losers bracket by player A 2-0.
-Player A loses and will face the winner of player B and player C in the loser's bracket.
-In the losers player B and player C play each other in the losers bracket.
If player B wins, player A faces player B in a Bo7 with a 2-0 lead.
If player C wins, player A faces player C in a Bo3 with no lead.
Why should player A care who he faces?
If the only point of a tournament is to figure out who the best player is, why have double elimination at all? I personally like the format and I actually think it's BETTER at determining the best player than a single elimination, but they could run a single elimination Bo5 with fewer total games played in most cases (see note at bottom). I think the idea of an extended series isn't trying to somehow enforce single elimination logic onto a double elimination format, they're different, you can't just take the logic from one and put it onto the other.
Note:
Math behind rounds:
in a tournament with N players
Bo5 single elimination takes at least (N-1)*3 at most (N-1)*5 games (all players are eliminated in 5 games series)
Bo3 double elim takes at least 2(N-1)*2 and at most (2N-1)*3 games to finish
so: 3N-3 <= 1xBo5 <= 5N-5
4N-4 <= 2xBo3 <= 6N-3 (btw, extended series actually adds potential games to this)
|
On November 11 2010 04:40 aNDRoM wrote:
@NoXious90 why bring the whole "best two players in North America" thing into it? MLG doesn't have that power, nor do I think they have that lofty a goal. They only need to determine who were the "best" x players for THIS TOURNAMENT so they can figure out who to write checks to and in what amount.
im in complete agreement with you?
|
On November 11 2010 05:03 vaahto wrote:Show nested quote +On November 11 2010 04:59 Thoramas wrote:
Going by this argument alone, then we should just run single eliminations instead of double eliminations since they equally determine just the winner, ignoring other rankings.
A double elimination tournament has a better likelihood of having a stronger player as the winner. It's not just used for determining the other places better, and having more games to play.
Then by that point you could say that the extended series does that better than double elimination, and there would be no end to the debate. I'm not trying to pick a fight, just saying that it might have been better to debate from a different perspective.
|
If the purpose of a tournament was to determine the best player, then we would just give the prize money to flash every time, but its not, tournaments are just an event, and it needs to determine the winner of the event based on a fair system, which the extended series takes away from. Sure, if player A goes 3-4 against player B and advances, most people would argue player B is better. But that doesn't mean player B should advance, being better is not the same thing as winning a series to stay in the tournament.
|
|
|
|
|
|
EPT
