EPTThere was a post about Iccup's distribution a while back, and it was determined that at the higher ranks, there are much fewer players.
http://www.teamliquid.net/forum/viewmessage.php?topic_id=99016
[R] Ladder Rank Distribtion
| Forum Index > SC2 General |
|
Commodore
United States97 Posts
There was a post about Iccup's distribution a while back, and it was determined that at the higher ranks, there are much fewer players. http://www.teamliquid.net/forum/viewmessage.php?topic_id=99016 | ||
|
Inschato
Canada1349 Posts
| ||
|
Dionyseus
United States2068 Posts
| ||
|
MMmmmmmmmm
United States36 Posts
| ||
|
ForTheSwarm
United States556 Posts
Go commy! Ha ha. I'm actually interested to find this out as well. | ||
|
Go0g3n
Russian Federation410 Posts
Using the graph above, I'd say everyone above D+ is in platinum. | ||
|
Perseverance
Japan2800 Posts
| ||
|
Irrelevant
United States2364 Posts
| ||
|
Ruthless
United States492 Posts
| ||
|
spinesheath
Germany8679 Posts
On March 17 2010 03:07 Go0g3n wrote: I'd say there are 60 platinum divisions in all 3 realms combined, about 100 players each, - 6k out of 25.000 testers, of course, mostly StarCraft and WarCraft enthusiasts. Using the graph above, I'd say everyone above D+ is in platinum. Maybe around 50-60 on average, but 100 is way too much. I know platinum divisions with as few as 35 players. | ||
|
HalfAmazing
Netherlands402 Posts
| ||
|
Dionyseus
United States2068 Posts
On March 17 2010 03:30 spinesheath wrote: Maybe around 50-60 on average, but 100 is way too much. I know platinum divisions with as few as 35 players. Yea, 100 is the maximum a division can have. Your estimate of around 50-60 on average sounds about right to me. | ||
|
BlasiuS
United States2405 Posts
SC2 B.net is different. Before you play your 1st 10 games, you are unranked, and don't belong to any league. This helps separate the new players from the established players, which is much better (and helps against smurfs). I would be interested to know from your pie chart, how many of those D players legitimately play at D rank level, and how many are new accounts/old accounts/smurfs. edit: I am a C- level BW player, and I reached platinum league easily, and have not moved since I was placed about 3 weeks ago. | ||
Excalibur_Z
United States12235 Posts
On March 17 2010 03:07 Perseverance wrote: Hopefully the distribution ends up being somewhere near ICCUP's as I would like Platinum to actually mean something. I wouldn't count on it. From what we can gather now, B rank and beyond in ICCup amounts to about 5% of the entire playerbase. Estimates are that Platinum will include roughly the top 10% of players. If we assume this (and there is no factual basis for this whatsoever): Pro represents the top 0.1% of players Platinum 0.1-10% Gold 10-25% Silver 25-50% Bronze 50-75% Copper 75-100% Then that will already distribute players more evenly than ICCup, though what the last poster in this thread said about ICCup starting everyone out at D is an important point to remember, as it doesn't factor in inactivity in the way that the SC2 system will (likely by reassigning you as your relative MMR falls compared to others as the overall skill level increases over time). | ||
|
Disastorm
United States922 Posts
| ||
|
Dionyseus
United States2068 Posts
On March 17 2010 07:42 Disastorm wrote: What somoene said earlier about not starting you off at the lowest and work your way up is the biggest flaw in blizzard's system. especially if they plan to make u take the 10 game test every season (not sure if they are gonna do that though). If they do that, that would mean when the new season starts your league would be determined by luck, by who you happen to play in the 10 game placement matches (and since everyone would be the same rank, you have equal chance to fight a good player as a crappy player). I think the biggest flaw in the system is that you don't get a seperate rating for each race. I played only as P for my first 300 games, and now I started testing out Z and am getting placed against people as if I were still playing my favorite race, which doesn't make any sense to me. | ||
|
duckhunt
Canada311 Posts
| ||
|
Sere
158 Posts
I'm guessing he was placed in Platinum because he was one of the first players to join, and they started filling at the top. Likewise, I was placed in Gold because I haven't seen any new Platinum leagues since I started. I've been #1 Gold for a tad over a week, if that matters. | ||
|
semantics
10040 Posts
not quite sure how you can mine the data for your iccup rank and battle.net rank and the data be reliable might be able to find the data for the b.net rank distribution | ||
|
Excalibur_Z
United States12235 Posts
On March 17 2010 07:42 Disastorm wrote: What somoene said earlier about not starting you off at the lowest and work your way up is the biggest flaw in blizzard's system. especially if they plan to make u take the 10 game test every season (not sure if they are gonna do that though). If they do that, that would mean when the new season starts your league would be determined by luck, by who you happen to play in the 10 game placement matches (and since everyone would be the same rank, you have equal chance to fight a good player as a crappy player). Think about how the system operates, and you'll discover why it's vastly superior to the ICCup system (or any legacy system). When you're 0-0 playing your placement matches, your MMR is of course going to be 0 (though this may be influenced by how experienced in RTS games you declare to be when you first start). Let's say the newbiest newb would have an MMR of 0 and the highest end Platinum player would have an MMR of about 3000 (to use WoW arena values). Your first match is going to be against someone who also has a very low MMR, so probably not a very good player, or someone also on his first placement match. Beating this person will probably net you about 300 MMR. Your next match is going to be against a rather low-end player, again with a similar MMR. Your volatility rating is still very high because you have so few games played, so if you beat this guy, you'll probably get another 400 MMR. Game 3 will be against someone who is low-mid range, again similar MMR. You win, and say you get 600 MMR for this. At this point your MMR is about 1300, and game 4 is against someone who's pretty solidly mid-level. You win this one and get 800 MMR. Now you're starting to play against people who are pretty good with your 2100 MMR. Let's say you lose game 5. Your MMR falls by about 400, because the system knows that you're better than the 1300 guy but not quite up to the 2100 guy. Game 6 you win and get 300 MMR, so you go back up to 1900. You still have a pretty high volatility rating, maybe that last game was a fluke. The system still believes you're pretty good, because after all, you've been winning more than you've been losing. Game 7, you win, you're up to 2400. Game 8 you're playing against someone who's very good, but you lose. You go back down to 2000 and the system is starting to think that maybe you're not quite as good. Game 9 you lose again, you go down to 1800. It's starting to get a pretty good idea of where you should be and your volatility drops. Game 10 you win and you go up to 1950. You get seated in Gold, being among the top 25% of players. This MMR will continue to change, sometimes wildly according to your win and loss streaks, as it constantly determines where to place you. If you all of a sudden get your MMR up to 2600 and hovers around there, you may suddenly find yourself promoted to Platinum. This is a way better system because those ten placement games more or less put you where you need to be at first. I don't expect people's MMRs to change every season either (there would be no reason for them to do that). | ||
| ||
![[image loading]](http://img.photobucket.com/albums/v462/3xcommodore/iccuprankspie.gif)
