|
This is certainly not the case since some players must post their profile in that site for it to show up.
|
United States12235 Posts
Added more information about the Bonus Pool today, specifically that everyone gets the same bonus pool and it builds at a rate of 1 per 2 hours.
|
On August 10 2010 07:43 carwashguy wrote:This is certainly not the case since some players must post their profile in that site for it to show up.
You're quoting a 9 day old post (granted I'm quoting a 2 day old one!) and saying it's inaccurate, which isn't surprising. The 'accepted' distribution right now is about 5%/15%/20%/20%/35%, which has held true regardless of having 300k or 900k characters.
|
On August 12 2010 04:15 Shadowed wrote:Show nested quote +On August 10 2010 07:43 carwashguy wrote:This is certainly not the case since some players must post their profile in that site for it to show up. You're quoting a 9 day old post (granted I'm quoting a 2 day old one!) and saying it's inaccurate, which isn't surprising. The 'accepted' distribution right now is about 5%/15%/20%/20%/35%, which has held true regardless of having 300k or 900k characters.
If no one has addressed the issue, why does it matter if the post is 9 days old? Besides, it's not as if the website has changed its method of ripping data in the last week. I'm not sure what you mean by "characters," but I'd be interested in knowing how you draw a conclusion as to what is "accepted."
|
On August 12 2010 08:05 carwashguy wrote:Show nested quote +On August 12 2010 04:15 Shadowed wrote:On August 10 2010 07:43 carwashguy wrote:This is certainly not the case since some players must post their profile in that site for it to show up. You're quoting a 9 day old post (granted I'm quoting a 2 day old one!) and saying it's inaccurate, which isn't surprising. The 'accepted' distribution right now is about 5%/15%/20%/20%/35%, which has held true regardless of having 300k or 900k characters. If no one has addressed the issue, why does it matter if the post is 9 days old? Besides, it's not as if the website has changed its method of ripping data in the last week. I'm not sure what you mean by "characters," but I'd be interested in knowing how you draw a conclusion as to what is "accepted."
Characters mean... characters, players, toons, chars, whatever you choose to call them.
It matters because 11 days ago every site had a very limited set of data, RTS was working off of around 70k (maybe a little less?) characters compared to the 700k it has today. The distribution RTS is listing has matched the same one SC2Ranks show, and I'm sure between us we have a pretty good distribution of characters/divisions.
As always, data will shift as more and more characters are added, but that shouldn't shift it by more than a percent or two. There's also the natural changes you will see as the matchmaker moves people between leagues and gets a better idea where people stand.
[edit] Or more accurately, there isn't a hidden cache of 500,000 Bronze players sitting somewhere that will drastically throw it off.
|
Great post, really enjoyed it.
I have been looking for this information for a while now. While one must keep in mind that it is impossible to say definitively how the system works without getting the info directly from the creators, I am comfortable with accepting your explanations based on the evidence you provided.
I am really glad that I don't have to be number 1 in my division to move up. That was starting to stress me out. Logically, one would have to obtain and maintain a rank 1 in their division for a set period of time, or perhaps for a set number of games, or even get to number 1 and not lose for a set number of games, before they would be promoted. Such a system would be quicker than waiting for your scheduled review date, but meh. At least now I know.
Given this info, do you all think it would be a good practice to start saying "gl hf, silver" when you started a game? if the opponent reciprocated, you would both start to get an idea of how your w/l ratio is for that particular league, thus giving you a better idea of if you will be eligible for promotion whenever the time comes.
|
On August 12 2010 08:20 Shadowed wrote:Show nested quote +On August 12 2010 08:05 carwashguy wrote:On August 12 2010 04:15 Shadowed wrote:On August 10 2010 07:43 carwashguy wrote:This is certainly not the case since some players must post their profile in that site for it to show up. You're quoting a 9 day old post (granted I'm quoting a 2 day old one!) and saying it's inaccurate, which isn't surprising. The 'accepted' distribution right now is about 5%/15%/20%/20%/35%, which has held true regardless of having 300k or 900k characters. If no one has addressed the issue, why does it matter if the post is 9 days old? Besides, it's not as if the website has changed its method of ripping data in the last week. I'm not sure what you mean by "characters," but I'd be interested in knowing how you draw a conclusion as to what is "accepted." Characters mean... characters, players, toons, chars, whatever you choose to call them.
So like... a player profile? Okay.
|
On August 12 2010 08:20 Shadowed wrote:
The distribution RTS is listing has matched the same one SC2Ranks show, and I'm sure between us we have a pretty good distribution of characters/divisions.
As always, data will shift as more and more characters are added, but that shouldn't shift it by more than a percent or two. There's also the natural changes you will see as the matchmaker moves people between leagues and gets a better idea where people stand.
[edit] Or more accurately, there isn't a hidden cache of 500,000 Bronze players sitting somewhere that will drastically throw it off.
Well how can u claim that? If you only list people that are self reported (or reported by someone else) u certainly have a preselction of a certain player type (The player type that is willing to look up his standings, is interested in the broader SC2 community by visiting community sites etc.). From a statistical standpoint you could never claim to have a normal distribution because of that. Having a big sample doesnt mean having equally distributed data. Actually there is a prominent example at the beginning of polls, where (I hope I'm getting this right) an American newspaper tried to predict the outcome of the presidential election. They had a clear result from their polls but they were blatently wrong just because they carried out their data collection through the phone. And only a certain socio-economic strata could afford phones at that time which skewed the sample pool a lot. Well, and its quite likely that the same applies to our example.
|
i think i dont understand the rating system completely... i was a platin player and got promoted to diamond. now i loose much more games and the opponents are sometimes soooo good, that i dont really have a chance.
is it only a time matter, till i go back to platin due MMR?
ps. Nice post.
|
is it only a time matter, till i go back to platin due MMR?
ps. Nice post.
no, MMR has decided you are better than the average platinum-guy, therefore the promotion - now it's deciding if you are better than the average diamond guy
happened to me too, I also got promoted too fast and lost the first couple of games in diamond; after a while I played lower diamonds and higher platinum-players and did/am doing fine
|
nice to know. thank you for the information! i thought i would stay at diamond loosing all games =)
now i won some games, but still hard! I think i need a training partner =)
|
0-1499: Newb
1500-1799: Average
1800-1999: Fairly skilled
2000-2199: Very skilled
2200-3000: Extremely skilled
sorry but this is very wrong
1-1500: Difficulty with basic motor functions.
1501-2000: Borderline retardation. Qualifies for handicapped parking.
2001-2400: Average Intelligence, realizes they are bad.
2401-2600: Above Average Intelligence. Capable of separating goods from bads.
2601-2800: Highly intelligent
2801-2999: Genius
3000+: Super Genius
|
I am not sure if I agree completely with how the system ranks individuals, but I suppose a tournament style cannot be implemented with so many players. The system does not accurately place players who play fewer games, despite winning or not. It is more based on the amount of time spent playing matches, instead of inherent skill level. For those who may be very skilled, yet have very little time to play...bronze league will be your realm.
Thanks for the explanation. It does make complete since how it all works, I am just not sure I completely agree with mechanics to it.
|
United States12235 Posts
On September 07 2010 03:40 Zebah wrote:
I am not sure if I agree completely with how the system ranks individuals, but I suppose a tournament style cannot be implemented with so many players. The system does not accurately place players who play fewer games, despite winning or not. It is more based on the amount of time spent playing matches, instead of inherent skill level. For those who may be very skilled, yet have very little time to play...bronze league will be your realm.
Thanks for the explanation. It does make complete since how it all works, I am just not sure I completely agree with mechanics to it.
That's not how it works at all. If you are very skilled but only play a few games, you'll still end up in Platinum or Diamond. That's actually the whole point of the system, in fact.
It's older ladders like ICCup that start everyone at the bottom and you move up by playing lots of games. I think you may be confused.
|
Another reason for the bonus pool is that it makes people constantly seem like they're improving, since as long as they run out of bonus pool and are staying constant, they'll be gaining points at a steady rate.
Since the whole division system is here to coddle people and hold their hands also, the current bonus pool method makes perfect sense to encourage people to play more, as well as encourage noobs that theyre getting better because they have +621 points from the bonus pool and are now at 1000+ points.
|
Thanks for the detailed information! Helps understanding the ladder system much more now.
|
Do custom matches impact on your hidden rating?
|
I have a question (don't find the clear answer anywhere)
If i'm in a division with lot of really good player, will i have more troubles having a good rating compared to someone who is in a poor division ?
Does someone with 1000 rating in super division have the same skill level as a 1000 rating player who is in a low level division ?
|
United States12235 Posts
On September 24 2010 21:11 cozzE wrote:
Do custom matches impact on your hidden rating?
Custom matches do not have any impact, no.
|
United States12235 Posts
On September 24 2010 21:39 Fabious wrote:
I have a question (don't find the clear answer anywhere)
If i'm in a division with lot of really good player, will i have more troubles having a good rating compared to someone who is in a poor division ?
Does someone with 1000 rating in super division have the same skill level as a 1000 rating player who is in a low level division ?
You play against people beyond your division, so the answer is no to the first question. 1000 should be equivalent to 1000 from another division because the player pool is shared. The "tough" division just makes it harder to get a high rank within that division because the better players there will have super high ratings, securing the top ranks of the division.
|
|
|
|
|
|
EPT
