On December 09 2011 00:26 Twistacles wrote: Ohhh it was bo1s? That explains idra/sen not qualifying. Kinda lame, though.
The statistical difference between Bo1s, Bo3s and Bo5s is actually pretty small IIRC. If you are strong enough to 2-0 or 2-1 a player, chances are you are strong enough to 1-0 them, 3-0, 3-1 or 3-2 them.
Sure, there is a little bit more variance, but not much.
The statistical difference doesn't show the real difference though. Of course the person who wins the first game has a greater chance to take a best of three, or even a best of 5. That should be obvious to anyone that knows how to do math. But that doesn't mean that best of 1s are therefore equally fair to best of 3s. Even if it is a small difference, that difference can create big changes in a tournament. Look at DRG's recent interview about his Coda A match against Taeja. DRG said that he expected to lose the first map because it was Antiga Shipyard and he expected to win the other two, which he did. If it was a best of 1 then he would have straight up lost.
That would probably be true if one bo1 decided whether your qualify or not, but 6 bo1s should be enough to remove the effect of random variables, like map selection and so on. But frankly, map selection shouldn't be a factor in mirror match ups, so your argument fails there, coz Idra lost two of them. Also, given how zergs are dominating the groups, I'd argue that if anything, maps have been fairly favorable to the swarm.
On December 09 2011 00:44 SC2NeCro wrote: Uh.... what? 1-0 relies on luck way more. Your logic makes no sense.
I mean that statistically, a player which won a Bo3 against an opponent is still overwhelmingly likely to win a Bo1 as well. Sure, people always say it's far more random, but the statistical difference is actually fairly small, again, IIRC. There was a thread about it a while ago.
In the GSL Mvp has about 67% game-winrate, but hes has 73% match-winrate. So though the difference isn't huge, it makes a difference.
For those still wondering about the DT viability against Zerg in BW, bear in mind that you do NOT get an alert when a unit gets oneshot, e.g if you don't pay attention, then it is entirely possible that your overlord/spore can see the unit np but that you won't notice until you hear the slicing and dicing in your base.
It is also possible that you don't hear it at all and turn to your base only to see it devoid of drones and a DT with 15 kills
On December 08 2011 23:41 Tommylew wrote: i wouldnt really say suprising Sen was in by far the haredest group in that the was 5 players all off similar ability who on their day can deliver.
group b had 3 players and welmu so not a shock there as one of them had to miss out.
group C: MVP was always going to walk it as noone in his group were going to come close. After that any of Goody, Aristeo or Cyto would of gone through as all lower european and about same abiility. Group D: Idra would be the only one i would say was a shock. ALways thought ti would be Idra followed by eithier killer or Titan.
Going on groups left ahead Group E: sorry if HuK doesnt get through this group soemthing is wrong and same for morrow both the two best players in this group. HuK by a long way. Group F: darkforce or beasty for 2nd palce and MKP will come first. Group G: Hard group between ToD and Strelok but on their day watch out for any of Kiwi, Cloud or even Bling as an outsider all good players. Id say one of ToD and Strelok will qualify if not both but it wouldnt be a shock if any of the toher 3 other then Bling(biased uk player :D) Group H: Supernova will win this group followed by Macsed with Fenix perhaps if he has his A game.
So in my predictions All Koreans and Chinese will go through.
Also its only top two so the is bound to be good players going out this early :D
Regarding Group E. and do note, I am slightly biased as Snute is norwegian, I wouldn't count the norwegian out, he has posted decent results and just recently took out viOLet in the GameCreds final 3-0, so he's by no means a walkover :D
On December 09 2011 00:26 Twistacles wrote: Ohhh it was bo1s? That explains idra/sen not qualifying. Kinda lame, though.
The statistical difference between Bo1s, Bo3s and Bo5s is actually pretty small IIRC. If you are strong enough to 2-0 or 2-1 a player, chances are you are strong enough to 1-0 them, 3-0, 3-1 or 3-2 them.
Sure, there is a little bit more variance, but not much.
The statistical difference doesn't show the real difference though. Of course the person who wins the first game has a greater chance to take a best of three, or even a best of 5. That should be obvious to anyone that knows how to do math. But that doesn't mean that best of 1s are therefore equally fair to best of 3s. Even if it is a small difference, that difference can create big changes in a tournament. Look at DRG's recent interview about his Coda A match against Taeja. DRG said that he expected to lose the first map because it was Antiga Shipyard and he expected to win the other two, which he did. If it was a best of 1 then he would have straight up lost.
Of course, but as someone else said, look at the number of upsets in this tournament and compare them to the upsets of Dreamhack. If Bo1s were always more volatile, then surely this tournament would have more upsets than Dreamhack? But it doesn't.
Again, I'm not arguing there is no difference between Bo1s and Bo3s. That would be really stupid. I'm arguing that the difference between Bo1s and Bo3s in terms of upsets/"lucky wins" etc is actually pretty small.
And yeah, the statistics obviously don't take maps into consideration. But then, if there is a map on which one player is always expected to lose, that map is probably a bigger problem than the tournament format and the map should be changed before the format. That said, I still think Bo1s are a bad idea - I just don't think players or fans can/should continually use the format as an excuse for when their favourite players lose.
On December 09 2011 00:44 SC2NeCro wrote: Uh.... what? 1-0 relies on luck way more. Your logic makes no sense.
I mean that statistically, a player which won a Bo3 against an opponent is still overwhelmingly likely to win a Bo1 as well. Sure, people always say it's far more random, but the statistical difference is actually fairly small, again, IIRC. There was a thread about it a while ago.
In the GSL Mvp has about 67% game-winrate, but hes has 73% match-winrate. So though the difference isn't huge, it makes a difference.
First of all, GSL caliber players will take games off anyone, it's more competitive and it is expected for series to be won by the minimum magin in many cases. On the other hand, WCG players skills are comparably much lower, it would be expected for MVP to have the same w.r. in both match ups. Second, there is a difference of course under certain conditions of course, but you haven't argued the significance of it yet.
On December 09 2011 00:26 Twistacles wrote: Ohhh it was bo1s? That explains idra/sen not qualifying. Kinda lame, though.
The statistical difference between Bo1s, Bo3s and Bo5s is actually pretty small IIRC. If you are strong enough to 2-0 or 2-1 a player, chances are you are strong enough to 1-0 them, 3-0, 3-1 or 3-2 them.
Sure, there is a little bit more variance, but not much.
The statistical difference doesn't show the real difference though. Of course the person who wins the first game has a greater chance to take a best of three, or even a best of 5. That should be obvious to anyone that knows how to do math. But that doesn't mean that best of 1s are therefore equally fair to best of 3s. Even if it is a small difference, that difference can create big changes in a tournament. Look at DRG's recent interview about his Coda A match against Taeja. DRG said that he expected to lose the first map because it was Antiga Shipyard and he expected to win the other two, which he did. If it was a best of 1 then he would have straight up lost.
That would probably be true if one bo1 decided whether your qualify or not, but 6 bo1s should be enough to remove the effect of random variables, like map selection and so on. But frankly, map selection shouldn't be a factor in mirror match ups, so your argument fails there, coz Idra lost two of them. Also, given how zergs are dominating the groups, I'd argue that if anything, maps have been fairly favorable to the swarm.
I'm not attacking or defending Idra. I don't care about Idra. I am talking about best of 1s. And of course maps affect mirror matchups because you do not play the same way on every single map. Some maps are better for attacking and some maps favor the defender. Some maps are better for a roach infestor style, and some maps favor a ling/baneling/infestor style. Some maps favor mutalisk style (or if you are Nestea, all maps favor Mutalisk style). Are you arguing that a player is going to be good at every single style? Well, MVP probably is, but I would argue most players are not. So yes, maps can affect mirror match ups.
But, using Idra as an example, consider this. Killer beat Orly and Idra, and Orly beat Idra. If these were all best of 3s that turned out the same except that Idra beat Killer, then Killer, Orly, and Idra would be in a 3way tie with a score of 4-2. Maybe Killer goes through anyway, maybe Orly or Idra do. The point I am making is that even in the case of a 4.7% difference in results (1 out of 21 sets of best of threes) the difference in the tournament can be significant if one player goes through and another doesn't.
On December 09 2011 00:26 Twistacles wrote: Ohhh it was bo1s? That explains idra/sen not qualifying. Kinda lame, though.
The statistical difference between Bo1s, Bo3s and Bo5s is actually pretty small IIRC. If you are strong enough to 2-0 or 2-1 a player, chances are you are strong enough to 1-0 them, 3-0, 3-1 or 3-2 them.
Sure, there is a little bit more variance, but not much.
The statistical difference doesn't show the real difference though. Of course the person who wins the first game has a greater chance to take a best of three, or even a best of 5. That should be obvious to anyone that knows how to do math. But that doesn't mean that best of 1s are therefore equally fair to best of 3s. Even if it is a small difference, that difference can create big changes in a tournament. Look at DRG's recent interview about his Coda A match against Taeja. DRG said that he expected to lose the first map because it was Antiga Shipyard and he expected to win the other two, which he did. If it was a best of 1 then he would have straight up lost.
Of course, but as someone else said, look at the number of upsets in this tournament and compare them to the upsets of Dreamhack. If Bo1s were always more volatile, then surely this tournament would have more upsets than Dreamhack? But it doesn't.
Again, I'm not arguing there is no difference between Bo1s and Bo3s. That would be really stupid. I'm arguing that the difference between Bo1s and Bo3s in terms of upsets/"lucky wins" etc is actually pretty small.
And yeah, the statistics obviously don't take maps into consideration. But then, if there is a map on which one player is always expected to lose, that map is probably a bigger problem than the tournament format and the map should be changed before the format. That said, I still think Bo1s are a bad idea - I just don't think players or fans can/should continually use the format as an excuse for when their favourite players lose.
I agree to an extent. The tournament you sign up for has certain rules and those are the rules the player plays by. If they lose in this format, then they lose and that is that. I just don't like the Best of 1 format, and I don't find the statistical defense of it compelling which was what I was replying to. And I also do not agree that the format will necessarily have more upsets if best of 1s are always more volatile. It just has to have some upsets. The more games you play, the harder it is to defeat a better player than you. I would also argue that longer series favor players that can adapt better.
Look, I generally agree with you. The difference isn't going to be gargantuan if you use best of 1s over best of 3s. I am just arguing that the difference is significant enough to change the tournament experience for the players and fans. That is all.
On December 09 2011 01:15 flowSthead wrote: I agree to an extent. The tournament you sign up for has certain rules and those are the rules the player plays by. If they lose in this format, then they lose and that is that. I just don't like the Best of 1 format, and I don't find the statistical defense of it compelling which was what I was replying to. And I also do not agree that the format will necessarily have more upsets if best of 1s are always more volatile. It just has to have some upsets. The more games you play, the harder it is to defeat a better player than you. I would also argue that longer series favor players that can adapt better.
Look, I generally agree with you. The difference isn't going to be gargantuan if you use best of 1s over best of 3s. I am just arguing that the difference is significant enough to change the tournament experience for the players and fans. That is all.
Ah, OK. I think we're both saying pretty much the same thing, just in different ways ^_^
On December 09 2011 00:26 Twistacles wrote: Ohhh it was bo1s? That explains idra/sen not qualifying. Kinda lame, though.
The statistical difference between Bo1s, Bo3s and Bo5s is actually pretty small IIRC. If you are strong enough to 2-0 or 2-1 a player, chances are you are strong enough to 1-0 them, 3-0, 3-1 or 3-2 them.
Sure, there is a little bit more variance, but not much.
The statistical difference doesn't show the real difference though. Of course the person who wins the first game has a greater chance to take a best of three, or even a best of 5. That should be obvious to anyone that knows how to do math. But that doesn't mean that best of 1s are therefore equally fair to best of 3s. Even if it is a small difference, that difference can create big changes in a tournament. Look at DRG's recent interview about his Coda A match against Taeja. DRG said that he expected to lose the first map because it was Antiga Shipyard and he expected to win the other two, which he did. If it was a best of 1 then he would have straight up lost.
That would probably be true if one bo1 decided whether your qualify or not, but 6 bo1s should be enough to remove the effect of random variables, like map selection and so on. But frankly, map selection shouldn't be a factor in mirror match ups, so your argument fails there, coz Idra lost two of them. Also, given how zergs are dominating the groups, I'd argue that if anything, maps have been fairly favorable to the swarm.
I'm not attacking or defending Idra. I don't care about Idra. I am talking about best of 1s. And of course maps affect mirror matchups because you do not play the same way on every single map. Some maps are better for attacking and some maps favor the defender. Some maps are better for a roach infestor style, and some maps favor a ling/baneling/infestor style. Some maps favor mutalisk style (or if you are Nestea, all maps favor Mutalisk style). Are you arguing that a player is going to be good at every single style? Well, MVP probably is, but I would argue most players are not. So yes, maps can affect mirror match ups.
But, using Idra as an example, consider this. Killer beat Orly and Idra, and Orly beat Idra. If these were all best of 3s that turned out the same except that Idra beat Killer, then Killer, Orly, and Idra would be in a 3way tie with a score of 4-2. Maybe Killer goes through anyway, maybe Orly or Idra do. The point I am making is that even in the case of a 4.7% difference in results (1 out of 21 sets of best of threes) the difference in the tournament can be significant if one player goes through and another doesn't.
In a mirror match up, you either adapt and use the best build/style suited for the map in question, or simply you should lose, because your opponent would have shown to be better than you if he's able to use that. I think inflexibility should be punished.. perhaps you don't.
That's only true if the differences are statistically significant, a point which you haven't proven yet.
On December 09 2011 01:19 allen_ami wrote: F91 and xigua qualified??
pure luck due to the format!!
Did you watch the games? Seems like Xigua played well since he went 6-0 against Kas, Welmu and Happy. Not the easiest opponents to start with..
If someone did watch Xiguas games, mind giving me a short summary of his play? it was very disappointing last time i saw him play in a tournament.
Caught a bit of his game with Kas(?) and he just seems more aggressive and his macro has improved a ton. Dunno, last I saw him he was pretty bad as well but now it seems like he understands how to play at a higher level. Interested to see more from him.
On December 09 2011 01:15 flowSthead wrote: I agree to an extent. The tournament you sign up for has certain rules and those are the rules the player plays by. If they lose in this format, then they lose and that is that. I just don't like the Best of 1 format, and I don't find the statistical defense of it compelling which was what I was replying to. And I also do not agree that the format will necessarily have more upsets if best of 1s are always more volatile. It just has to have some upsets. The more games you play, the harder it is to defeat a better player than you. I would also argue that longer series favor players that can adapt better.
Look, I generally agree with you. The difference isn't going to be gargantuan if you use best of 1s over best of 3s. I am just arguing that the difference is significant enough to change the tournament experience for the players and fans. That is all.
Ah, OK. I think we're both saying pretty much the same thing, just in different ways ^_^
Shake hands and make up?
Absolutely. Hugs for everyone! (and by everyone I mean you).
EPT
