|
On March 31 2011 13:25 Antedelerium wrote:Show nested quote +On March 31 2011 11:37 Orome wrote:
Only just stumbled on this. Holy cow that's a bad format. I get that they want to have interesting matches from the start, but wow. Now, it's 4am and I hope to god I haven't messed my numbers up, but this is what I got from doing a little late-night math, assuming all players are 50/50 vs. each other. Should be correct.
The probability that one of the 16 seeded players will win the tournament: 79.375%
The probability that someone coming from the open qualifiers will win the tournament: 20.625%
The probability for one seeded player to win the tournament: 4.96%
The probability for one open qualifier player to win the tournament (if 256 qualifiers): 0.08%
For any individual player, the probability to win is 62 times higher if you're seeded than if you're playing the open qualifier. This isn't even taking into account fatigue which is going to affect the qualifiers hugely and the championship match format (where the player coming from the winner's bracket only needs to win one Bo3, which skew the results in the seeded players' favour even more).
Now, obviously 4 qualifiers will get into pool play, which will also give them a 4.96% shot (although getting there is extremely unlikely for any given player). If we therefore count the chances of winning of all the players in pool play together, we get 20x4.96%=99.2%. Now where did those 0.8% go? Well, that's the combined chance of all the players that don't make pool play (ie. drop any series during the qualifiers in a 256 man bracket) to win this tournament.
Seriously MLG?
True, but that's assuming that each Bo3 is like a 50-50 coin flip. Skill plays a lot into it. Yes, it sucks for people in the open qualifier due to the sheer number of matches they have to play, but do well, get some points, and do better in the next round. That's how it goes. It's a season and this is just one event. Arguments aside, thanks for this. I feel like actually understand the whole format for the first time. Awesome graphic too for lazy people who didn't end up reading the whole thing.
Even if you assume someone has an 80% winrate (and I highly doubt anyone is close to that), the chance to win the tournament coming from the open tournament is laughably small.
And yes, I get this is just one event and they can try to get points to get seeded for the next events, but 1. it completely undermines the credibility of each tournament in itself and 2. as long as they keep this format, 16 people are going to be ludicrously favoured every tournament. This makes absolutely no sense, you want people to win a tournament based on how good they play at the tournament, not because they did well in some event a few months back. That Agh probably has about an equal of chance of winning this tournament as Naniwa and Huk is completely ridiculous.
|
On March 31 2011 13:46 Orome wrote:Show nested quote +On March 31 2011 13:25 Antedelerium wrote:On March 31 2011 11:37 Orome wrote:
Only just stumbled on this. Holy cow that's a bad format. I get that they want to have interesting matches from the start, but wow. Now, it's 4am and I hope to god I haven't messed my numbers up, but this is what I got from doing a little late-night math, assuming all players are 50/50 vs. each other. Should be correct.
The probability that one of the 16 seeded players will win the tournament: 79.375%
The probability that someone coming from the open qualifiers will win the tournament: 20.625%
The probability for one seeded player to win the tournament: 4.96%
The probability for one open qualifier player to win the tournament (if 256 qualifiers): 0.08%
For any individual player, the probability to win is 62 times higher if you're seeded than if you're playing the open qualifier. This isn't even taking into account fatigue which is going to affect the qualifiers hugely and the championship match format (where the player coming from the winner's bracket only needs to win one Bo3, which skew the results in the seeded players' favour even more).
Now, obviously 4 qualifiers will get into pool play, which will also give them a 4.96% shot (although getting there is extremely unlikely for any given player). If we therefore count the chances of winning of all the players in pool play together, we get 20x4.96%=99.2%. Now where did those 0.8% go? Well, that's the combined chance of all the players that don't make pool play (ie. drop any series during the qualifiers in a 256 man bracket) to win this tournament.
Seriously MLG?
True, but that's assuming that each Bo3 is like a 50-50 coin flip. Skill plays a lot into it. Yes, it sucks for people in the open qualifier due to the sheer number of matches they have to play, but do well, get some points, and do better in the next round. That's how it goes. It's a season and this is just one event. Arguments aside, thanks for this. I feel like actually understand the whole format for the first time. Awesome graphic too for lazy people who didn't end up reading the whole thing. Even if you assume someone has an 80% winrate (and I highly doubt anyone is close to that), the chance to win the tournament coming from the open tournament is laughably small. And yes, I get this is just one event and they can try to get points to get seeded for the next events, but 1. it completely undermines the credibility of each tournament in itself and 2. as long as they keep this format, 16 people are going to be ludicrously favoured every tournament. This makes absolutely no sense, you want people to win a tournament based on how good they play at the tournament, not because they did well in some event a few months back. That Agh probably has about an equal of chance of winning this tournament as Naniwa and Huk is completely ridiculous.
Is it really that much worse then Qualifiers -> Code A -> Code S in GSL? Does the fact that a code S player has a 100% chance to win the tournament undermine that tournaments credibility? I'd argue that it's better then that because you can actually win it in the first event you get into.
Think of this as just a hyper-accelerated version of that, even if you're in the 16 seeded players, coming last in your group means you need to win 2 best of 3's to retain your seeding, coming 2nd last means you need 1 best of 3 to retain your seeding, coming in 2nd means you retain your seed, that is A LOT like GSL, where the lowest 2 in group play go into the up and down matches and where the lowest gets 1 chance, the 2nd lowest gets 2 chances.
|
On March 31 2011 14:08 Dingobloo wrote:Show nested quote +On March 31 2011 13:46 Orome wrote:On March 31 2011 13:25 Antedelerium wrote:On March 31 2011 11:37 Orome wrote:
Only just stumbled on this. Holy cow that's a bad format. I get that they want to have interesting matches from the start, but wow. Now, it's 4am and I hope to god I haven't messed my numbers up, but this is what I got from doing a little late-night math, assuming all players are 50/50 vs. each other. Should be correct.
The probability that one of the 16 seeded players will win the tournament: 79.375%
The probability that someone coming from the open qualifiers will win the tournament: 20.625%
The probability for one seeded player to win the tournament: 4.96%
The probability for one open qualifier player to win the tournament (if 256 qualifiers): 0.08%
For any individual player, the probability to win is 62 times higher if you're seeded than if you're playing the open qualifier. This isn't even taking into account fatigue which is going to affect the qualifiers hugely and the championship match format (where the player coming from the winner's bracket only needs to win one Bo3, which skew the results in the seeded players' favour even more).
Now, obviously 4 qualifiers will get into pool play, which will also give them a 4.96% shot (although getting there is extremely unlikely for any given player). If we therefore count the chances of winning of all the players in pool play together, we get 20x4.96%=99.2%. Now where did those 0.8% go? Well, that's the combined chance of all the players that don't make pool play (ie. drop any series during the qualifiers in a 256 man bracket) to win this tournament.
Seriously MLG?
True, but that's assuming that each Bo3 is like a 50-50 coin flip. Skill plays a lot into it. Yes, it sucks for people in the open qualifier due to the sheer number of matches they have to play, but do well, get some points, and do better in the next round. That's how it goes. It's a season and this is just one event. Arguments aside, thanks for this. I feel like actually understand the whole format for the first time. Awesome graphic too for lazy people who didn't end up reading the whole thing. Even if you assume someone has an 80% winrate (and I highly doubt anyone is close to that), the chance to win the tournament coming from the open tournament is laughably small. And yes, I get this is just one event and they can try to get points to get seeded for the next events, but 1. it completely undermines the credibility of each tournament in itself and 2. as long as they keep this format, 16 people are going to be ludicrously favoured every tournament. This makes absolutely no sense, you want people to win a tournament based on how good they play at the tournament, not because they did well in some event a few months back. That Agh probably has about an equal of chance of winning this tournament as Naniwa and Huk is completely ridiculous. Is it really that much worse then Qualifiers -> Code A -> Code S in GSL? Does the fact that a code S player has a 100% chance to win the tournament undermine that tournaments credibility? I'd argue that it's better then that because you can actually win it in the first event you get into. Think of this as just a hyper-accelerated version of that, even if you're in the 16 seeded players, coming last in last in your group means you need to win 2 best of 3's to retain your seeding, coming 2nd last means you need 1 best of 3 to retain your seeding, coming in 2nd means you retain your seed, that is A LOT like GSL, where the lowest 2 in group play go into the up and down matches and where the lowest gets 1 chance, the 2nd lowest gets 2 chances.
I dislike GSL's format as well since I feel like Ro16 is too little of an achievement to get seeded into the next code S, but there are differences. The biggest qualm I have with MLG's format is that they have one tournament in which they try to crown the champion of this MLG, but with their format, the players have wildly different chances of actually winning the tournament. In the end, they'll claim to have figured out who the best player was that weekend, but seriously, if a qualifier makes it through 17 rounds only to be defeated in the finals by a seeded player, isn't his accomplishment much bigger than the winner's? In GSL, the tournaments are separate. Yes, it's hard to get to code S and comparatively easy to stay in code S, which is something I definitely don't like about the GSL, but at least all the participants in their respective tournaments have the same chance of winning. Every qualifier has the same shot at code A, every code A player the same chances at winning code A and so on. It's a slow system, probably too slow, but it's fair. MLG's is horribly lopsided.
|
the one thing I will add:
they've been doing a similar bracket setup for quite some time in halo (though without the pool play) and it's always been that a team who loses round 1 of open bracket would have to play an insane amount of bo3's in order to win the tournament. This is why sundance has always said (on cast) that MLG is a test of endurance. You don't just have to play well on day 1. You need to play well from start to finish. If you slip up early on and wind up in the losers bracket, then you've earned yourself several more rounds of play.
|
If points from the 3 events last year were not getting reset a new player could place in 3rd or maybe 2nd this weekend and still not get seeded next time. That won't happen this weekend but it could in the 4th or 5th MLG this year after the points build up again.
The top few places in Dallas can suddenly become the worst players in the world and will get to stay seeded for a at least a few events. In a couple MLGs it'll be very hard for a new guy to place highly consistently enough to break into the top 16 seeds. The system is designed to have much less change then the Code S system where a champion like MVP can be knocked out in the next month and in a few months you'll probably have people seeded based on past performance that should not be there anymore but stick around anyways.
Actually, the way GSL got the initial Code S and A roster by assigning points from the 3 2010 GSLs is similar to the way MLG will be ranking players. And the number of players that got into code S from past performance that way but now looked bad is why I think MLG will have a similar problem. Except then imagine they gave the top 16 or 32 of the first GSL a huge advantage to win and place highly in the next two and imagine how much 'better' you think the Code S/A list would have been.
|
On March 31 2011 14:37 coolcor wrote:
If points from the 3 events last year were not getting reset a new player could place in 3rd or maybe 2nd this weekend and still not get seeded next time. That won't happen this weekend but it could in the 4th or 5th MLG this year after the points build up again.
From MLGPro.com:
* After the Dallas Competition, 2011 Starcraft 2 Pro Circuit Rank Points will replace all 2010 Starcraft 2 Pro Circuit Rank Points. 2011 Starcraft 2 Pro Circuit Rank Points may be used as Rolling Rank Points in 2012.
Basically, 2010 counts for this weekends event, after that they become like "lower tier" points used as a tie-breaker, so top 16 this competition are more then likely the seeded players for Columbus.
|
Format really makes getting into the top 16 so important for future events
|
Ya this weekend you need to be top 16 to get a seed. But in Columbus 16th place won't be enough anymore for a new player or a pro who lost early in the open bracket since the top players of Dallas will now have points to add to(you'll need 12th place at least maybe more), and the cutoff for top 16 seeds will go up every event. (until they reset it again or start dropping old points or something). People will have to go through the open bracket into top 16 multiple times to break in and that will be very hard to do.
|
lol i'm not reading all 14 pages but i agree with TC
hate that the open bracket players have to play about a million games but meh. it's alright
i like this format the most because it's like an actual sports bracket. I hate double elimination because there's no such thing as a loser's bracket in the NFL, MLB, NBA playoffs
this format in that aspect is awesome
EDIT: LOL there is a loser's bracket
well then IMO we should do it like real sports and have group play that decides seeding and just 1 winenrs bracket w/ bo7s
|
On March 31 2011 15:19 pbjsandwich wrote:
lol i'm not reading all 14 pages but i agree with TC
hate that the open bracket players have to play about a million games but meh. it's alright
i like this format the most because it's like an actual sports bracket. I hate double elimination because there's no such thing as a loser's bracket in the NFL, MLB, NBA playoffs
this format in that aspect is awesome
EDIT: LOL there is a loser's bracket
well then IMO we should do it like real sports and have group play that decides seeding and just 1 winenrs bracket w/ bo7s
Football and basketball have seasons that last months, with lots of games. MLG is just a handful of events every year. There's way too many people, and MLG wants to reward returning players who have ranked in prior events.
|
On March 31 2011 11:37 Orome wrote:
Only just stumbled on this. Holy cow that's a bad format. I get that they want to have interesting matches from the start, but wow. Now, it's 4am and I hope to god I haven't messed my numbers up, but this is what I got from doing a little late-night math, assuming all players are 50/50 vs. each other. Should be correct.
The probability that one of the 16 seeded players will win the tournament: 79.375%
The probability that someone coming from the open qualifiers will win the tournament: 20.625%
The probability for one seeded player to win the tournament: 4.96%
The probability for one open qualifier player to win the tournament (if 256 qualifiers): 0.08%
For any individual player, the probability to win is 62 times higher if you're seeded than if you're playing the open qualifier. This isn't even taking into account fatigue which is going to affect the qualifiers hugely and the championship match format (where the player coming from the winner's bracket only needs to win one Bo3, which skew the results in the seeded players' favour even more).
Now, obviously 4 qualifiers will get into pool play, which will also give them a 4.96% shot (although getting there is extremely unlikely for any given player). If we therefore count the chances of winning of all the players in pool play together, we get 20x4.96%=99.2%. Now where did those 0.8% go? Well, that's the combined chance of all the players that don't make pool play (ie. drop any series during the qualifiers in a 256 man bracket) to win this tournament.
Seriously MLG?
Ouch.
MLG, taking SPORT out of ESPORT.
|
okay my question is...if youre in the open bracket and you lose a game, since that bracket is double elimination, can you possibly get back into the top 4 and go on to group play?
|
On March 31 2011 20:41 bigbeau wrote:
okay my question is...if youre in the open bracket and you lose a game, since that bracket is double elimination, can you possibly get back into the top 4 and go on to group play?
Yes you can.
|
This format favours the top 16 players far too much.
|
On March 31 2011 20:41 bigbeau wrote:
okay my question is...if youre in the open bracket and you lose a game, since that bracket is double elimination, can you possibly get back into the top 4 and go on to group play?
If you lose a match (Bo3) in the open bracket you go to the loser's bracket and you can't get into group play (you can get into the loser's side of the championship bracket, but not group play).
|
On March 31 2011 20:45 zaii wrote:Show nested quote +On March 31 2011 20:41 bigbeau wrote:
okay my question is...if youre in the open bracket and you lose a game, since that bracket is double elimination, can you possibly get back into the top 4 and go on to group play? Yes you can.
No, you can't.
That's pretty obvious, only the top4 in the WB will make it to group play. Basically you have to make WB semifinal to get to group play, so in a 256 man bracket to get top 4 you need to win 6 straight bo3 games.
Cheers to MLG 
|
Good job. motbob > slasher.
|
Man, I am currently building a space shuttle that should fly me to the moon, but they forgot to deliver the construction manual for it.
But nevermind, I think I can use this sheet. It will do the trick:
On February 27 2011 15:13 motbob wrote:And here's the image that ties it all together! ![[image loading]](http://imgur.com/PUwhY.png) click the image for a larger version
|
thanks so much for explaining it. And I don't really like how if you lose in Championship bracket you still get a chance to win it all by advancing your way through the losers bracket.
Other than that I think it's a good thing to try and change it up a bit. It will be nice to see if other tournaments follow this or change their setup to represent something like this.
|
I know a lot have figured out the "worst case scenario" for someone in the open bracket to work their way up though the ranks and the number of games that they would have to play. However I don't think I have seen much on the rare possibility that the open bracket would get filled out with 256 players. Granted, I don't know the turnouts MLG gets for the open crowd and I am just assuming that those matches are played on site in the BO3 format (from what I could read). Thats one monster of a tournament before anything can even start!
|
|
|
|
|
|
EPT![[image loading]](http://i.imgur.com/elIjP.png)
