|
how would there not be a decent amount of "rags to riches stories"?
it's a brand new game. some people who didn't play bw at all will pick this game up, dedicate an inordinate amount of hours to it and eventually become really good. this happens in every game...
i played cs from the time 1.0 came out and was a cal-m player...when i made the move to cs:s i was cal-i/cevo-p, but some players that never played the original cs still made it to cevo-p and were competitive.
conceptually speaking, why is sc2 all of a sudden the exception to the rule?
if they re-arrange players every season, that makes more sense to me...but if that's the case, there should be a more concretely provided "league" system, and not a ladder style that is constantly changing from top to bottom.
if what everyone is saying is true so far, then the ladder system is literally completely meaningless except to the recreational player. this equates to what you would call a pubber or a pub all-star in cs
and if that's the case, we may all be better off pointing our attention to third party leagues like a cevo or something where your progress is tangible
Edit: typo fixes
|
Are you sure the division "rank" is based upon skill/mmr/points? Seems more logical that this modifier is determined by when the division is created. First diamond divisions = 0 modifier. The ones created today gets more bonus.
|
how come I dont see my division on your chart?
|
^ my division and the divisions of 3 of my friends aren't on the chart either
|
On November 19 2010 07:09 Prissy wrote:
how come I dont see my division on your chart?
Most likely, your division didn't have enough people in the top 200, so Excalibur couldn't calculate the ranking for it.
|
But if the MMR never changes, then shouldn't it change people's division overtime as they improve/their MMR rises/lowers ?
It confuses me a little, but I can see that you've spent a LOT of time working on this, and you obviously know a lot more about it than I do!
It would be cool if blizzard just told us how it all works lol.
I still don't understand why its all split up into leagues instead of a single list of everyone on the server like in wc3.
|
On November 19 2010 07:19 QuothTheRaven wrote:Show nested quote +On November 19 2010 07:09 Prissy wrote:
how come I dont see my division on your chart? Most likely, your division didn't have enough people in the top 200, so Excalibur couldn't calculate the ranking for it.
I believe you are right.... because the #1 ranked player only has about 1900pts.
|
On November 18 2010 12:01 Excalibur_Z wrote:
Remember that leagues are split 20% each among active players, but the required activity period is not known.
Are you guessing that this 20% is a target number used for determining how to distribute newly-placed players?
I was a bit confused by the Blizzcon comment that they intended each league to be 20% considering that the actual numbers from censuses of the armory data are skewed heavily toward the bottom, but if that's a target number of 20% each when one only counts players that exceed some activity threshold, that might make sense. It would also mean that Diamond players, on average, play ENORMOUSLY more than other leagues, which I guess shouldn't surprise me.
|
United States12237 Posts
On November 19 2010 07:05 euroboy wrote:
Are you sure the division "rank" is based upon skill/mmr/points? Seems more logical that this modifier is determined by when the division is created. First diamond divisions = 0 modifier. The ones created today gets more bonus.
That's not the case because Hartley Quest didn't exist 2 weeks ago (or if it did, it had nobody in the top 200) yet it's S-Rank. It also wouldn't be fair if your division got a larger modifier just because it's newer because if you were a top-level player you would have an obscene point advantage. red.venom for example just started playing SC2 a short time ago and he's in Moratun Sierra which is also S-Rank.
|
United States12237 Posts
On November 19 2010 07:27 Lysenko wrote:Show nested quote +On November 18 2010 12:01 Excalibur_Z wrote:
Remember that leagues are split 20% each among active players, but the required activity period is not known. Are you guessing that this 20% is a target number used for determining how to distribute newly-placed players? I was a bit confused by the Blizzcon comment that they intended each league to be 20% considering that the actual numbers from censuses of the armory data are skewed heavily toward the bottom, but if that's a target number of 20% each when one only counts players that exceed some activity threshold, that might make sense. It would also mean that Diamond players, on average, play ENORMOUSLY more than other leagues, which I guess shouldn't surprise me.
This is an excerpt from an email chain I had with Bashiok.
Me:
Greg mentioned in Friday's panel that player distribution was 20% across all leagues. However, that contradicts what we're seeing in SC2Ranks parses which seem to have more of a pyramid structure:
For NA:
Diamond 7.88% (45,394)
Platinum 14.69% (84,605)
Gold 18.28% (105,329)
Silver 18.33% (105,583)
Bronze 40.82% (235,144)
Was he citing internal numbers, meaning league distribution will change after Master/Grandmaster League implementation to a percentile-based structure? Or, was he referring to theoretical distribution where each league has an equal MMR range?
Bashiok:
20% across all leagues can’t be correct. It’s definitely more of a pyramid. I’m not sure what he was intending to say. I also didn’t see that panel personally (meaning to watch it today).
Bashiok (few days later):
Hello again. I was completely wrong. Diamond is 20% of all players. Each league represents roughly 20% of active players. So, Diamond is rather huge. It attempts to keep 20% at all times although it could very well fluctuate, BUT, not enough to scale all the way down to 7% as sc2ranks shows. It’s likely that they’re so far off because they’re not accounting for active players (or rather, inactive), which we do. I’ll count this as the last time I open my big yapper before investigating. 
Me:
Wow, okay, that's huge. I guess it would depend on the inactivity period? I wonder how strictly it adheres to those adjusted percentages. Thanks for the info!
Bashiok:
To remain ‘active’ is, I’ve heard, based on a number of games played per week.
There is some slush there as well to ensure that, say come patch time, there isn’t some mass promotion/demotion to keep the percentages. And to that end, they’re loose percentages, so Diamond could even be a few percent over or under at any one time, which isn’t an unsubstantial amount of people.
As a result of this, Shadowed added an Activity filter to SC2Ranks, and looking at that (try the filter at 7 days and looking at the Stats page) you start to see that most of the percentages are a lot closer to 20% (with the exception of Bronze which is consistently a bit higher... maybe it's double-counting it because Copper was removed?).
|
I doubt they ever removed copper code-wise, just moved it all down a notch and added Diamond. Copper might have a higher percent because it tries to keep ~20% per league, and Copper is the fallback so it doesn't force absolutely 20% of the teams into Diamond if only 15% deserve it. Random guess.
|
That 20% of active players figure is interesting.
They said Masters will be the top 2%, which people were thinking was going to be the top 1/3rd of diamond based on that 6% figure. However, if it's the top 2% of active players then it would only be the top 10% of diamond players.
|
KELERCHIAN X-RAY REPRESENT! YEAAAAA BOYYYEEEEE. Interesting read though =)
|
United States12237 Posts
So going back to that Taiwanese Plat player for a moment, I'm just going to throw out a few complete guesses:
- His point value of 1117 Plat is between 466 Diamond and 579 Diamond.
- In order to be ranked among all these S-Rank Diamond divisions, he is probably in an S-Rank Platinum division.
- IF the ladder is a contiguous ladder and points are offset by league in addition to division, then adjusting that value of 1117 by increments of 63 means his rating is either 550 or 487.
- IF it is 487 adjusted rating, that would mean a modifier of +630 from S-Rank Plat to S-Rank Diamond which means 10 Diamond tiers.
Again, I want to see more data on this, and I think this Plat guy was a great data point. I'd love to see these guesses backed up.
|
Very convincing data, thank you for putting in the work.
However, I don't understand why the system is like this. Do people from different divisons play different people? no. So, why am I penalized from being in a lesser division??? If I had placed into Diamond a few days earlier than I did, I'd be #70 on the NA top 200 list but instead I am not even on it because I'm in some cruddy division I placed into after owning the game for a few hours.
I liked the idea of it being based on elo but now that I see its points +/- division I kind of don't care about the list any more
|
Hey excalibur_Z, I just want to point you towards my situation as it might help.
I only lost ~150 points when I was promoted to diamond (went from 600 to 450).
My division is new and consists of only recently promoted high win/loss ratio players with low amounts of games played. "Scourge Rho".
http://us.battle.net/sc2/en/profile/2168758/1/Prada/ladder/25685#current-rank
(It's not updated on sc2ranks.com yet)
|
United States12237 Posts
On November 19 2010 09:31 SovSov wrote:Hey excalibur_Z, I just want to point you towards my situation as it might help. I only lost ~150 points when I was promoted to diamond (went from 600 to 450). My division is new and consists of only recently promoted high win/loss ratio players with low amounts of games played. "Scourge Rho". http://us.battle.net/sc2/en/profile/2168758/1/Prada/ladder/25685#current-rank(It's not updated on sc2ranks.com yet)
Do you have the exact numbers? Was it exactly 600 to exactly 450?
|
On November 19 2010 09:14 Alejandrisha wrote:
Very convincing data, thank you for putting in the work.
However, I don't understand why the system is like this. Do people from different divisons play different people? no. So, why am I penalized from being in a lesser division??? If I had placed into Diamond a few days earlier than I did, I'd be #70 on the NA top 200 list but instead I am not even on it because I'm in some cruddy division I placed into after owning the game for a few hours.
I liked the idea of it being based on elo but now that I see its points +/- division I kind of don't care about the list any more
if you had joined a higher divisionn, you'd have less points, and be in the same spot. It's just a modifier that makes your points total look higher tha it is.
|
On November 19 2010 09:46 Excalibur_Z wrote:Do you have the exact numbers? Was it exactly 600 to exactly 450?
I'ma do the math.
So I had 477 platinum (I uploaded a replay which showed my current score) at a time, I added and subtracted all the points I've made since then (only 16 games). After that, I should have 686, but I have 478 diamond (my promotion was in between those 16 games). So apparently I was off, and I have 208 points missing. I can confidently say that number is accurate unless there is something wrong with my logic of getting there.
With what you were saying, does it mean that my division wasn't just a coincidence of recently promoted diamonds? I mean, it would make sense, since everyone in my division likely has high MMR with their win/loss. SURELY there would be someone who has like 100+ games played and got promoted by now into my division if divisions were solely based on date of promotion.
One more thing to note, there is one outlier in my division (NYCdaniel), but checking his matches he is on a HUGE losing streak (probably after he was promoted to diamond?).
|
Very interesting, good job Excalibur_Z! No wonder I wasn't getting on the top 200. Viking Romeo is definitely not S-Rank, if anyone was wondering...
|
|
|
|
|
|
EPT
