Bump these teams to SuperTitan please!!!

GreatGonzales · 02-18-2013, 05:25 PM

Quote:GG are you sure it's only the top .1%? I don't think there is any sort of percentile they're going by, but rather your hidden skill past a certain threshold. And like you said, dividing into meaningful groups, I'll use a comparison here. Lets pretend everyone is a bird, with say hummingbirds at the bottom, then sparrows, ravens, I don't know, vultures, then eagles representing the ST. As of right now, there's 2 "eagles", would you put the vultures in an arena with eagles, knowing the eagle will win the majority of the time? No right? Of course, this is not a 100% accurate comparison but hopefully it makes sense.

Again, since the system is a black box, we can't know if its actually the top .1%. But I think it's right? It seems right. An exercise someone could do is take a count of all players on the OSN, but that is still going to be missing some divisions I think.

I disagree with the premise of your example. Yes, if indeed there were only two eagles, then there should be only two entities in the eagle group. But I don't think there are only 2 eagles.

Mr.Banger · (This post was last modified: 02-18-2013 05:31 PM by Mr.Banger.)

(02-18-2013 05:25 PM)GreatGonzales Wrote:
Quote:GG are you sure it's only the top .1%? I don't think there is any sort of percentile they're going by, but rather your hidden skill past a certain threshold. And like you said, dividing into meaningful groups, I'll use a comparison here. Lets pretend everyone is a bird, with say hummingbirds at the bottom, then sparrows, ravens, I don't know, vultures, then eagles representing the ST. As of right now, there's 2 "eagles", would you put the vultures in an arena with eagles, knowing the eagle will win the majority of the time? No right? Of course, this is not a 100% accurate comparison but hopefully it makes sense.

Again, since the system is a black box, we can't know if its actually the top .1%. But I think it's right? It seems right. An exercise someone could do is take a count of all players on the OSN, but that is still going to be missing some divisions I think.

I disagree with the premise of your example. Yes, if indeed there were only two eagles, then there should be only two entities in the eagle group. But I don't think there are only 2 eagles.

The eagles got to being eagles after passing a certain hidden skill level no? If there were more people past the hidden skill level, there would be more STs right?

(02-18-2013 05:19 PM)GreatGonzales Wrote: Alexjiang1/mr. Banger, I really have to disagree with you guys that the status quo is acceptable. 7 months is long enough for a population of players to define its elite.

There's the elite, and the elite of the elite, likes Navy's and the Navy Seals. Isn't that what Super Titan is suppose to represent.

By the way, nice words in italics under your name hahaha

(02-18-2013 05:23 PM)Gfich Wrote: Guys, i have the solution: top lists for 2v2. Solves everything.

Not really, unless it displayed the actual hidden skill.

GreatGonzales · (This post was last modified: 02-18-2013 05:53 PM by GreatGonzales.)

Quote:Yes I agree, 2 ST is not very many, but could it be possible that the hidden skill level of the masters isn't high enough? Like i mentioned, it was previously stated that going around 44-2 is enough to be a 1v1 ST, we went 35-1 and became an ST, there's not that big of a difference right? But imo, I always thought leagues were to determine skill level...

I mean, that's just begging the question - what I'm saying is that the hidden skill threshold should be changed. Consider this: there are 2000 teams, of which 50 of them are truly excellent, and maybe 10 are super super excellent. For the sake of argument, lets say OML is using the top .1% threshold for 2v2. So, that leaves 2 ST and 48 excellent teams who are relegated to masters. So, how do they get into ST? Well, they have to have a good winning streak in league play! But they only get matched against their fellow excellent teams, so they struggle to do better than a 50% win ratio. Thus, they remain master league.

This is what I think is probably going on. But just hypothetically, don't you agree that the threshold should be changed to account for a smaller population of players than 1v1?

You guys should divorce yourselves from the notion that because the promotion threshold has always been this way, then it must be the most ideal set up. OML made an attempt when they set up the leagues, but we know more now and clearly its not working. 7 months and only 2 teams in the highest league = system broken. Yes, promotion is based on hidden skill rating but the hidden skill required for promotion was determined by some percentile of the population, probably .1%. Whatever it is, though, it is too restricting considering the size of the 2v2 population.

Quote:^^ What he said! You are a smarter man than I. Will you accept the position as my spokesperson?

Haha, with pleasure. I am all fired up.

Mr.Banger · 02-18-2013, 05:53 PM

(02-18-2013 05:32 PM)GreatGonzales Wrote:
Quote:Yes I agree, 2 ST is not very many, but could it be possible that the hidden skill level of the masters isn't high enough? Like i mentioned, it was previously stated that going around 44-2 is enough to be a 1v1 ST, we went 35-1 and became an ST, there's not that big of a difference right? But imo, I always thought leagues were to determine skill level...

I mean, that's just begging the question - what I'm saying is that the hidden skill threshold should be changed. Consider this: there are 2000 teams, of which 50 of them are truly excellent, and maybe 10 are super super excellent. For the sake of argument, lets say OML is using the top .1% threshold for 2v2. So, that leaves 2 ST and 48 excellent teams who are relegated to masters. So, how do they get into ST? Well, they have to have a good winning streak in league play! But they only get matched against their fellow excellent teams, so they struggle to do better than a 50% win ratio. Thus, they remain master league.

This is what I think is probably going on. But just hypothetically, don't you agree that the threshold should be changed to account for a smaller population of players than 1v1?
You guys should divorce yourselves from the notion that because the promotion threshold has always been this way, then it must be the most ideal set up. OML made an attempt when they set up the leagues, but we know more now and clearly its not working. 7 months and only 2 teams in the highest league = system broken. Yes, promotion is based on hidden skill rating but the hidden skill required for promotion was determined by some percentile of the population, probably .1%. Whatever it is, though, it is too restricting considering the size of the 2v2 population.

Okay. Lets go with the top .1% for the sake of argument. These 2 teams will have a higher hidden skill then the others. If this top .1% applied, getting to ST is simple, either wait for more people to come around or outskill one of the 2 teams. But honestly though, I strongly believe they didn't make it the top .1% but rather a hidden skill threshold which I don't feel like should be changed. Why? Because, it IS possible to get past that threshold. If they changed the threshold, being a 2v2 ST wouldn't have as much meaning as 1v1 because 2v2 has a lower hidden skill requirement.

GreatGonzales · (This post was last modified: 02-18-2013 06:07 PM by GreatGonzales.)

Quote:Okay. Lets go with the top .1% for the sake of argument. These 2 teams will have a higher hidden skill then the others. If this top .1% applied, getting to ST is simple, either wait for more people to come around or outskill one of the 2 teams. But honestly though, I strongly believe they didn't make it the top .1% but rather a hidden skill threshold which I don't feel like should be changed. Why? Because, it IS possible to get past that threshold. If they changed the threshold, being a 2v2 ST wouldn't have as much meaning as 1v1 because 2v2 has a lower hidden skill requirement.

This is your error: the number of players that can exceed a certain skill threshold is probably related to the size of the population of players. I say probably because again, it's a black box. But ELO, for example, is related to the size of the population. You need people below you in order to prop yourself up to a higher standing, is one way to think about it.

Again, you're right that OML isn't using percentiles but a hidden skill threshold. But they decided the skill threshold based on some delineation of the player population, right? This is what I mean when I discuss percentiles.

Lets agree that there IS an unreasonable threshold out there. Suppose the threshold was the top .001%, virtually ensuring that no more than 1 team would ever have the league of ST. That would not have much utility, right? There is an ideal here and I think it is frankly obvious that we are not there.

I mean, what if there were only 5 1v1 supertitans? Wouldn't that seem wrong to you? Yes the 2 teams making it into ST would have the highest hidden skill rating, just as poweewee has a much higher skill rating than harvanold. The idea is we open the league up so that meaningful distinctions can be made.

Mr.Banger · 02-18-2013, 06:13 PM

(02-18-2013 06:05 PM)GreatGonzales Wrote:
Quote:Okay. Lets go with the top .1% for the sake of argument. These 2 teams will have a higher hidden skill then the others. If this top .1% applied, getting to ST is simple, either wait for more people to come around or outskill one of the 2 teams. But honestly though, I strongly believe they didn't make it the top .1% but rather a hidden skill threshold which I don't feel like should be changed. Why? Because, it IS possible to get past that threshold. If they changed the threshold, being a 2v2 ST wouldn't have as much meaning as 1v1 because 2v2 has a lower hidden skill requirement.

This is your error: the number of players that can exceed a certain skill threshold is probably related to the size of the population of players. I say probably because again, it's a black box. But ELO, for example, is related to the size of the population. You need people below you in order to prop yourself up to a higher standing, is one way to think about it.

Again, you're right that OML isn't using percentiles but a hidden skill threshold. But they decided the skill threshold based on some delineation of the player population, right? This is what I mean when I discuss percentiles.

Lets agree that there IS an unreasonable threshold out there. Suppose the threshold was the top .001%, virtually ensuring that no more than 1 team would ever have the league of ST. That would not have much utility, right? There is an ideal here and I think it is frankly obvious that we are not there.
I mean, what if there were only 5 1v1 supertitans? Wouldn't that seem wrong to you? Yes the 2 teams making it into ST would have the highest hidden skill rating, just as poweewee has a much higher skill rating than harvanold. The idea is we open the league up so that meaningful distinctions can be made.

I will assume ELO, as in chess ratings. I'm not sure how it correlates with the size of the population? No matter who you beat, your ELO will go up, and once it goes past say 2500 then you're technically a GM (other requriments involved), or a Super Titan. And yes it does seem wrong to me if there were only 5 1v1 ST. But I would assume OML put the 1v1 and 2v2 ST hidden skill at the same threshold. So there are 2 2v2 teams that are past the ST threshold, and the masters, like 1v1 masters, have not passed the threshold yet.

GreatGonzales · (This post was last modified: 02-18-2013 06:24 PM by GreatGonzales.)

Breaking this down for people who don't want to read my essays:

1. 7 months is long enough for a player population to settle, and the elite to emerge.
2. The current league promotion scheme for 2v2 does not effectively divide the player population into meaningful groups. Estimate: fluffy: 800, clever: 800, gifted: 300, master: 100, supertitan: 2
3. The ST promotion system should be relaxed to being in line with realities of the 2v2 population. Note: NOT to bestow charity to undeserving teams, but to adjust the system in such a way so that supertitan becomes a more meaningful distinction.

Quote:I will assume ELO, as in chess ratings. I'm not sure how it correlates with the size of the population? No matter who you beat, your ELO will go up, and once it goes past say 2500 then you're technically a GM (other requriments involved), or a Super Titan.

No because you get diminishing returns as you increase in rating. The size of population matters, at least for ELO. Getting to 2500 in ELO, to use your example, is basically impossible without a very large population.

Quote:And yes it does seem wrong to me if there were only 5 1v1 ST. But I would assume OML put the 1v1 and 2v2 ST hidden skill at the same threshold. So there are 2 2v2 teams that are past the ST threshold, and the masters, like 1v1 masters, have not passed the threshold yet.

And this is precisely the problem. I agree, probably ONL used the same threshold for 1v1 and 2v2. But because the 2v2 population is smaller, so is the ST population. As we've seen, it is too small to be of use. Therefore it should be adjusted to be in line with differences in the 2v2 population.

Ok I'm going to bed. It's 3:30 AM here Tongue

Necrocat219 · 02-18-2013, 06:36 PM

I think that players who don't know each other in real life should get a chance at getting ST status; at the moment the only two teams up there is a player with two accounts which mean they can synergise easily and brothers who can help each other understand what the other is doing. 2vs2 probably needs slightly more lax boundaries regarding ST as the current ST's are probably much much higher in hidden rating than other teams.

worldfamous · 02-18-2013, 08:32 PM

(02-18-2013 06:36 PM)Necrocat219 Wrote: I think that players who don't know each other in real life should get a chance at getting ST status; at the moment the only two teams up there is a player with two accounts which mean they can synergise easily and brothers who can help each other understand what the other is doing. 2vs2 probably needs slightly more lax boundaries regarding ST as the current ST's are probably much much higher in hidden rating than other teams.

Are you saying Gurleyman and Krogoth is one person?

Necrocat219 · 02-18-2013, 09:20 PM

I can't remember where but somewhere, forgive me if I'm wrong, but I think I recall Alex and Banger talking about playing vs the only other ST pair which was a player who had two gamecenter accounts. That's what I remember from a while ago though so they could have been replaced by another pair or something along those lines.