|
United States15275 Posts
On March 16 2012 04:19 Komentaja wrote:Show nested quote +On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. Just because it's used heavily does not mean it's a good system. Obvious parallels apply. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players. The more logical way to do a tiebreaker is to consider each person in the group as player X, and whoever has the best win percentage (i.e. whoever is the best player that day) among them against X (instead of against a person you happened to be tied with) is determined to be the best player, and advances. This may require more games. It is the same system that Gom itself uses until people happen to become tied in score.
No one ever wins or loses by luck. That would imply that something outside of their control contributed to the record. HuK straightup lost to Fin and Virus, he did not happen to stumble into a pit that made him lose. Your example is a strawman.
Your theoretical situation is impossible. The player that wins the group is never included in the tiebreaker because he will always have the best score.
They are not valued as better players; they are valued as the players with the better records.
Win percentage is irrelevant. You could make everyone play the same amount of games and run into the exact same problem. Then how would you decide who gets to advance?
Impractical. Do you realize how many games they would have to play to rank the players without doubt?
|
On March 16 2012 04:35 CosmicSpiral wrote:Show nested quote +On March 16 2012 04:19 Komentaja wrote:On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. Just because it's used heavily does not mean it's a good system. Obvious parallels apply. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players. The more logical way to do a tiebreaker is to consider each person in the group as player X, and whoever has the best win percentage (i.e. whoever is the best player that day) among them against X (instead of against a person you happened to be tied with) is determined to be the best player, and advances. This may require more games. It is the same system that Gom itself uses until people happen to become tied in score. No one ever wins or loses by luck. That would imply that something outside of their control contributed to the record. HuK straightup lost to Fin and Virus, he did not happen to stumble into a pit that made him lose. Your example is a strawman. Your theoretical situation is impossible. The player that wins the group is never included in the tiebreaker because he will always have the best score. They are not valued as better players; they are valued as the players with the better records. Win percentage is irrelevant. You could make everyone play the same amount of games and run into the exact same problem. Then how would you decide who gets to advance? Impractical. Do you realize how many games they would have to play to rank the players without doubt?
Apparently some people want every qualifier/group/round to be a Bo7 with a Bo5 for tiebreaks, and a Bo3 for tiebreaks within tiebreaks.
Just so it's fair...
(That's clearly sarcastic; the current system is fine)
|
For the people saying its unfair or a bad system.
If Huk simply had won against Fin or Virus he would have avanced.
|
whoa, I'm watching the vods now, and I can't believe Lucky failed that all in against Huk.
All he needed to do was have his lings attack the forge instead of humping the forcefielded pylon.
With a single cannon and 3 sentries, there's not enough DPS to kill lucky's army before he completely dismantles the wall. You attack the pylons and he has to forcefield to save them, then you kill the forge and there is just too much area for him to cover.
Instead, all of his lings did absolutely nothing because they were trying to attack a forcefielded pylon.
Huk did the best he could, but it still looked like he should have lost.
|
On March 16 2012 04:28 Kuskinator wrote:Show nested quote +On March 16 2012 04:19 Komentaja wrote:On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players.. No, no, they have the same W/L ratio - that's why they're tied. Within that group of people who are tied - Virus and Fin had the higher "W/L" out of that group of people in the tiebreak because they both beat HuK.
There is no such thing as "W/L" (in quotes) -- that is, W/L is a hard number. The tied people are tied in W/L, so they should play each other to "untie" themselves. Instead of just picking one of the tied people by a roll of the dice.
|
On March 16 2012 04:44 Oboeman wrote:
All he needed to do was have his lings attack the forge instead of humping the forcefielded pylon.
Pylons are the Zergling equivalent of fire hydrants. It's hard to convince them to hump anything else...
|
On March 16 2012 04:48 mikedebo wrote:Show nested quote +On March 16 2012 04:44 Oboeman wrote:
All he needed to do was have his lings attack the forge instead of humping the forcefielded pylon.
Pylons are the Zergling equivalent of fire hydrants. It's hard to convince them to hump anything else...
haha good analogy
|
Wow the last game was so so epic TvT mech of ForGG and Virus. One of the highest level of mech play for sure, so back and forth, who says mech is turtle anyway.
|
On March 16 2012 04:47 Komentaja wrote:Show nested quote +On March 16 2012 04:28 Kuskinator wrote:On March 16 2012 04:19 Komentaja wrote:On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players.. No, no, they have the same W/L ratio - that's why they're tied. Within that group of people who are tied - Virus and Fin had the higher "W/L" out of that group of people in the tiebreak because they both beat HuK. There is no such thing as "W/L" (in quotes) -- that is, W/L is a hard number. The tied people are tied in W/L, so they should play each other to "untie" themselves. Instead of just picking one of the tied people by a roll of the dice.
They were tied 3-2.
Move them into a separate group, and they are 2-0, 1-1, and 0-2.
IF they have to play again, we can end up in a situation where HuK wins 2-0 (moving him 2-2) Fin goes 1-1 (moving him 2-2) and Virus loses 0-2 (moving him 2-2). Now, do you use the second round of games (i.e. the tiebreaker just played) to determine who goes though? In this case it would be HuK and Fin.
But wait! We can't do that because they're all actually 2-2 between themselves. Now we're in a situation that MLG is in regularly with the loser bracket and dreaded "extended series" rule. Why should you be punished for losing to somebody in a certain order?
|
Also: virus and fin proved that they can lose to average players in all matchups.
Huk only lost to terran. He could have made a great run out of that.
|
United Kingdom38255 Posts
On March 16 2012 05:16 Heimatloser wrote:
Also: virus and fin proved that they can lose to average players in all matchups.
Huk only lost to terran. He could have made a great run out of that.
Virus only lost to protoss...
(while winning once in each match up)
|
As soon as I saw the results I prepared for the the flak.
This one seems pretty clean cut tho, Huk, Virus and Fin all had 3-2, but huk lost to both virus and fin.
Seems ok. Huk was damn close tho. Next time he will make it
|
On March 16 2012 04:35 CosmicSpiral wrote:Show nested quote +On March 16 2012 04:19 Komentaja wrote:On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. Just because it's used heavily does not mean it's a good system. Obvious parallels apply. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players. The more logical way to do a tiebreaker is to consider each person in the group as player X, and whoever has the best win percentage (i.e. whoever is the best player that day) among them against X (instead of against a person you happened to be tied with) is determined to be the best player, and advances. This may require more games. It is the same system that Gom itself uses until people happen to become tied in score. No one ever wins or loses by luck. That would imply that something outside of their control contributed to the record. HuK straightup lost to Fin and Virus, he did not happen to stumble into a pit that made him lose. Your example is a strawman. Your theoretical situation is impossible. The player that wins the group is never included in the tiebreaker because he will always have the best score. They are not valued as better players; they are valued as the players with the better records. Win percentage is irrelevant. You could make everyone play the same amount of games and run into the exact same problem. Then how would you decide who gets to advance? Impractical. Do you realize how many games they would have to play to rank the players without doubt?
You misunderstand my post on all counts. First, the tied player that advances from the tie without another game played won the roll of the dice that determined he had beat the wrong people -- a.k.a., he was lucky. Second, nobody has the "better record" if they have a tied W/L, if we take all players to be equal. Third, you contradict yourself by asking me how you would "decide who gets to advance" and then saying that my decision you are asking for would be "impractical".
I'll assume that you are saying only that my system is impractical. Well, my response to that is, to break a two-player tie, it would take only one extra game. So for 90% of the cases, I don't think that is impractical.
To break a three-player tie, it would take a significant number of games, however. In that case, I would have them play 5 games: 1 to figure out the winner of players A and B, 1 to figure out the winner of players B and C, and if in both cases it's B, B advances. If it's B in the first case and C in the second, they play each other 1 game and A plays C 1 game. The winner of B and C faces the winner of A and C, and that game's winner is the one who advances. Complicated, and 5 extra games, but I would still say 5 games that were needed. Obviously the 3-player tie-breaker is not mathematically fair, but it's pretty close. And the three-player tie is very unlikely to happen, even though it happened here.
Example with players A, B, and C. A>B, B>C. A>B, B>C. A>B. Player A advances with a maximum of 5 games played.
More than 3 players is almost never going to happen, but if it does, it would be impractical to put in a perfectly fair system. Therefore, a roll of the dice (or Gom's system) could determine which of the 4 are the lucky 3 who play each other in the system I described above, or in some similar system.
Edit: Logical mistake.
|
On March 16 2012 05:22 Komentaja wrote:Show nested quote +On March 16 2012 04:35 CosmicSpiral wrote:On March 16 2012 04:19 Komentaja wrote:On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. Just because it's used heavily does not mean it's a good system. Obvious parallels apply. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players. The more logical way to do a tiebreaker is to consider each person in the group as player X, and whoever has the best win percentage (i.e. whoever is the best player that day) among them against X (instead of against a person you happened to be tied with) is determined to be the best player, and advances. This may require more games. It is the same system that Gom itself uses until people happen to become tied in score. No one ever wins or loses by luck. That would imply that something outside of their control contributed to the record. HuK straightup lost to Fin and Virus, he did not happen to stumble into a pit that made him lose. Your example is a strawman. Your theoretical situation is impossible. The player that wins the group is never included in the tiebreaker because he will always have the best score. They are not valued as better players; they are valued as the players with the better records. Win percentage is irrelevant. You could make everyone play the same amount of games and run into the exact same problem. Then how would you decide who gets to advance? Impractical. Do you realize how many games they would have to play to rank the players without doubt? To break a three-player tie, it would take a significant number of games, however. In that case, I would have them play 5 games: 1 to figure out the winner of players A and B, 1 to figure out the winner of players B and C, and if in both cases it's B, B advances. If it's B in the first case and C in the second, they play each other 1 game and A plays C 1 game. The winner of B and C faces the winner of A and C, and that game's winner is the one who advances. Complicated, and 5 extra games, but I would still say 5 games that were needed. Obviously the 3-player tie-breaker is not mathematically fair, but it's pretty close. And the three-player tie is very unlikely to happen, even though it happened here. .
The current system cuts out the tie breaker though.
A already beat B & C, B already beat C.
|
WHAAAAA? i turned it off yesterday thinking HuK and Forgg would make it. Out of everyone else though...... i never ever expected virus
|
Sad for Lucky. A little bit better execution and he would have been there. :-(
Here's hoping that some more Zergs can make it up through Code A this season!
|
On March 16 2012 04:24 msjakofsky wrote:Show nested quote +On March 16 2012 03:39 mrtomjones wrote:On March 16 2012 03:22 Witten wrote:On March 16 2012 01:55 Alryk wrote:On March 15 2012 06:09 CosmicSpiral wrote:
Virus for first place (I can dream right?)! Hahaha. Yes you can! Poor Huk  Yes poor Huk, dropping into Code B straight from Code S, being gifted an Up/Down seed and then dropping into Code A. Yes, poor poor Huk. poor mc, drg, polt, sen, idra, etc falling out of code S and being "gifted" code S. Stop being fricking hypocrits. MC and DRG and even MMA don't really belong to that list... they've all been seeded to code S because of MLG results with korean competition and they have all been doing well there since then, MMA with 2 major gsl wins, DRG with one, MC smashing his way to round of 8... Idra and Sen were 100% undeserving, Polt is debatable (he belongs there but whether he deserves the seed or not is debatable), Huk simply doesn't deserve it, basically beating oz in his strongest matchup is the only thing he did to "deserve" it.
Huk made the round of 8 in GSL, has had numerous top finishes beating numerous top Koreans in the process. He deserved the up and downs. If you didnt notice he almost made it through them too. He didn't and that sucks but I hope he gets through code A now.
|
On March 16 2012 05:36 Kuskinator wrote:Show nested quote +On March 16 2012 05:22 Komentaja wrote:On March 16 2012 04:35 CosmicSpiral wrote:On March 16 2012 04:19 Komentaja wrote:On March 16 2012 04:02 PureLuckz wrote:On March 16 2012 03:39 Komentaja wrote:
That's a terrible weird tiebreaker system. Why don't they do tiebreakers based on assuming all players are equal? I.E., go by the win/loss ratio.
Please don't insult my "loyalty" or "bias". Insult my idea instead, if you must insult. Nothing weird about it. Head-to-head is used in a lot of major sports as one of the higher priority tiebreakers. The NBA and NFL has head-to-head win % pretty high up there in the order of tiebreakers. Unfortunately HuK's only losts came from people he was tied against. Just because it's used heavily does not mean it's a good system. Obvious parallels apply. The reason I think it's weird is because it allows luck to be a factor. "Unfortunately" (let's say, unluckily) HuK lost against certain people. Therefore those people were valued as somehow better players than the others despite the only basis being a few isolated games that very hour. The situation becomes even more ridiculous when, say, two players are tied and are rated higher than the person that won the group, despite having lower W/L ratios and despite the fact that the winner of the group had already done better than the two tied players. The more logical way to do a tiebreaker is to consider each person in the group as player X, and whoever has the best win percentage (i.e. whoever is the best player that day) among them against X (instead of against a person you happened to be tied with) is determined to be the best player, and advances. This may require more games. It is the same system that Gom itself uses until people happen to become tied in score. No one ever wins or loses by luck. That would imply that something outside of their control contributed to the record. HuK straightup lost to Fin and Virus, he did not happen to stumble into a pit that made him lose. Your example is a strawman. Your theoretical situation is impossible. The player that wins the group is never included in the tiebreaker because he will always have the best score. They are not valued as better players; they are valued as the players with the better records. Win percentage is irrelevant. You could make everyone play the same amount of games and run into the exact same problem. Then how would you decide who gets to advance? Impractical. Do you realize how many games they would have to play to rank the players without doubt? To break a three-player tie, it would take a significant number of games, however. In that case, I would have them play 5 games: 1 to figure out the winner of players A and B, 1 to figure out the winner of players B and C, and if in both cases it's B, B advances. If it's B in the first case and C in the second, they play each other 1 game and A plays C 1 game. The winner of B and C faces the winner of A and C, and that game's winner is the one who advances. Complicated, and 5 extra games, but I would still say 5 games that were needed. Obviously the 3-player tie-breaker is not mathematically fair, but it's pretty close. And the three-player tie is very unlikely to happen, even though it happened here. . The current system cuts out the tie breaker though. A already beat B & C, B already beat C.
Sure, but A had also lost to D and E whereas C beat D and E. We're back to the tie. The reason they have to play the games again is so we don't value any certain previous wins over any other specific previous wins.
|
On March 16 2012 05:46 mrtomjones wrote:Show nested quote +On March 16 2012 04:24 msjakofsky wrote:On March 16 2012 03:39 mrtomjones wrote:On March 16 2012 03:22 Witten wrote:On March 16 2012 01:55 Alryk wrote:On March 15 2012 06:09 CosmicSpiral wrote:
Virus for first place (I can dream right?)! Hahaha. Yes you can! Poor Huk Yes poor Huk, dropping into Code B straight from Code S, being gifted an Up/Down seed and then dropping into Code A. Yes, poor poor Huk. poor mc, drg, polt, sen, idra, etc falling out of code S and being "gifted" code S. Stop being fricking hypocrits. MC and DRG and even MMA don't really belong to that list... they've all been seeded to code S because of MLG results with korean competition and they have all been doing well there since then, MMA with 2 major gsl wins, DRG with one, MC smashing his way to round of 8... Idra and Sen were 100% undeserving, Polt is debatable (he belongs there but whether he deserves the seed or not is debatable), Huk simply doesn't deserve it, basically beating oz in his strongest matchup is the only thing he did to "deserve" it. Huk made the round of 8 in GSL, has had numerous top finishes beating numerous top Koreans in the process. He deserved the up and downs. If you didnt notice he almost made it through them too. He didn't and that sucks but I hope he gets through code A now.
So Jinro deserves a seed, too?
So hongun does and coca does, noblesse does, keen does, clide does, byun does. AND SO ON.
Just because Huk reached top8 once does not make him more worthy of a seed than all these players who actually have better overall stats against korean players.
In fact, Huk only got that far because his Code A spot was given to him way back. He never ever qualified for GSL the "real" way and since random results (top8...WOW) do not justify seeds, he does not deserve his spot more ( or just as much ) as other players.
|
is fOrGG wearing and old oGs shirt or does NaDa Mall still sponsor oGs?
|
|
|
|
|
|
EPT
