|
On August 12 2010 01:07 Icks wrote:Show nested quote +On August 12 2010 00:58 floor exercise wrote:On August 12 2010 00:41 Pyrthas wrote:On August 12 2010 00:33 kajeus wrote:On August 12 2010 00:32 Takkara wrote:On August 12 2010 00:30 kajeus wrote:On August 12 2010 00:25 JoshSuth wrote:On August 12 2010 00:23 Pyrthas wrote:On August 12 2010 00:17 Takkara wrote:On August 12 2010 00:10 Pyrthas wrote:
Bonus pool doesn't explain why non-diamond and people playing 15 games are in Blizzard's top 200. (I'm taking mrdx's word here; I haven't checked myself.) There's three explanations: 1) The person in question is not the same person as on the list but another person with a different character code (Blizzard doesn't release the codes) 2) The person's MMR was high even though he was in Platinum. There's talk about having to lose to get promoted. So if someone was 26-0 in Platinum they could still have a skyhigh MMR because they're playing Diamond player in Platinum. 3) There's an error in Blizzard's Top 200 reporting tool. This guy is currently rank 200: http://www.sc2ranks.com./char/us/403486/InSTinKEdit: For posterity, he is currently 7-1 in 1v1. He's got 100 wins (across 1v1, 2v2, 3v3, etc.) and an 83% win percentage... perhaps this shows that Blizzard looks at all divisions for MMR, not just 1v1? 83-17 looks pretty damn good, but so many of them were placements for this fellow. LOL, they looked at 2v2 and 3v3 rankings to determine a "best player"? Hahaha. This is completely bizarre. This company has absolutely no idea how seriously they're taken, and hence they're content to completely fly in the face of other sites' rankings without even a suggestion of an explanation of how the hell they slapped their ranking together. Ahh, e-sports. How little you've come all these years. :D They specifically said they only looked at 1v1 ratings. Ok, then a dude with a 7-1 history is in the top 200 in the USA. The best part here, imo, is that five of those eight games were, of course, placement matches. So this is someone who has played three games after placement, and Blizzard's algorithm decides it has enough information to say he's only worse than 199 other people on the NA server. Actually, and I don't know why or how this happens, but he has never faced anyone under 400 points in diamond As of the compilation of this list he was 6-1 with his best win coming from Tozar, an 800 point protoss, and his one loss coming from Idra It's all very confusing Why is it confusing? If he started later than most of us, it could make sense. We made our placement matches nearly after release, very few people ranked, we play against anyone from good players to bad players. Imagine you start 1v1 now. (I dont know if the system tries to match you against a low lvl player at the begining but anyway...) If the first placement match is against a Diamond and you win, the system will keep on trying to find you better players.
Nah I explained how it happened. if it has no prior information about you it will start you at the lowest skill level, but he had been playing 2v2 at a diamond level, so it started him off at diamond
|
infinity21
Canada6683 Posts
On August 12 2010 01:05 SharkSpider wrote:
In short, skill doesn't necessarily mean win in a game, and the ladder system only has enough players to make accurate rankings in regions below the top few hundred players, because at the high level, being the best player online effectively lets you rack up infinite points. (even though the best 400 players may be offline and you're 401) This will be remedied as time progresses and as more really good players emerge.
You're contradicting yourself. If skill doesn't necessarily mean win, then even if you're the best player on ladder, you will still lose games to worse players and lose a lot of points for the loss.
|
On August 12 2010 00:41 Jibba wrote:Show nested quote +On August 12 2010 00:28 BondGamer wrote:
Wouldn't it be possible for someone like IdrA who has an insane win percentage to just play a couple games a month to remain "active". He will always be in the top 200 then. Just have a second account to play to your hearts content. Which is why competition based purely on ladders are frivolous. I'm not sure why this is a surprise for anyone. The only benefit of ladders over leagues is accessibility, but when we're talking about competitive play, that really doesn't matter. Ladder systems, even better ones like ICCUP, will always be flawed for determining ranking, their main use (like someone said earlier) is to convey a sense of progress and to obtain a general idea of who the best players are. I would argue that competition based on ladders is actually detrimental to competitive play, as we've seen in online WoW. Teams do exactly as you described, which means they aren't even practicing sufficiently. So yeah, ladders kind of suck. Once the major leagues get going (not just tournaments, but also season play), that's when rankings will truly start to develop. Basically, people need to stop worrying about it. The ladder's not going away, and it's not getting any more accurate.
Agreed, despite kespa problems, the format of tournement ELO seems the most legit way to determine this. and ofc some courag-esque tourney wouldnt hurt the "who is viable for ranking" dillemma. Id say take all the top 4 of any fairly large monetary event, and consider them for a pass into the 'league'
Has that 7-1 guy played in some tourny?
|
LOL
7-1 in top 200 and people are DEFENDING IT
ROFLMAO
GOD, UNBELIEVABLE LOL
|
probably includes custom games
|
Who cares, you earn your rep from LAN tournaments. Ladder is merely practice.
|
Battlecraft games should be included in rankings imo
|
On August 12 2010 01:33 travis wrote:
LOL
7-1 in top 200 and people are DEFENDING IT
ROFLMAO
GOD, UNBELIEVABLE LOL
Makes you wonder if blizzard even reviewed the list before releasing it. I don't care if the 7-1 is the best player in the world, there is no way you can know that from 8 games played.
|
United States12278 Posts
On August 11 2010 21:17 paralleluniverse wrote:Show nested quote +The top 200 players are determined across divisions by comparing their relative rankings and skill, while meeting certain requirements, such as ensuring that they’re active. We can all see that the top 200 is NOT sorted by points, and different from the rankings shown at www.sc2ranks.com. For example Dayvie is ranked 49 in the official top 200, but has always been in the top 10 in terms of points. This shows that the ladder ranks that the game uses based on points is nonsense. Either whatever method was used to calculate this top 200 is correct, or ranking based on points is. They can't both be right. If points are not the optimal way to rank players, why is Blizzard using it to rank in the game? Why not use this new method to rank? Or make points converge to the results given by this new method? Basically, Blizzard is admitting their points system for ranking is wrong, making the ladder rankings in the game a charade.
Yeah you're right. Their ranking has to be based on MMR (because it's clearly not points), but may be influenced by ladder activity as well (they mentioned that has an impact, but to what degree we're not sure). Points need to be the sole ranking factor, or at least something else that's equally transparent. Lists like this one only serve to create confusion. My guess would be that this linear ranking is one that uses MMR - sigma*3, but then that wouldn't explain the 7-1 guy at the bottom whose sigma would probably be enormous (unless sigma rapidly changes after each match, far more than we estimated).
Points are also absolutely comparable across divisions because the players all share competition, for those who are saying they aren't.
|
On August 12 2010 01:09 tofucake wrote:
Blizzard has a separate 200 for each region. The best way to compare players across regions is to see how they do in cross-region tournaments.
Yeah, all those cross-region tourna—
|
United States12278 Posts
On August 12 2010 00:19 JoshSuth wrote:
Everyone is not inflated in the same way. Some players are more inflated than others because bonus pool has a cap of 200. If I play every single day and burn off my bonus pool, I have earned more inflated points than a person who hits their cap more than once before burning it off.
I don't think anyone's reached the Bonus Pool cap yet. Just yesterday I checked my 3v3 team that I started on 8/2 and played only up to placement and the bonus pool was 271.
|
United States12278 Posts
Expanding on MMR - sigma * 3 for a moment to try and explain the #200 guy who went 7-1, let's say he went 7-0 and his MMR jumped to like 2900 with a sigma of about 500. Then he lost to Idra who has an MMR of 3000, and because the result was expected and he hasn't played many games, his sigma dropped dramatically to about 50. His "ranking value" after factoring in MMR and uncertainly would then be 2750 which might be enough to hit top 200. It's still an anomaly, but something like that would explain it.
|
On August 12 2010 02:08 Excalibur_Z wrote:
Expanding on MMR - sigma * 3 for a moment to try and explain the #200 guy who went 7-1, let's say he went 7-0 and his MMR jumped to like 2900 with a sigma of about 500. Then he lost to Idra who has an MMR of 3000, and because the result was expected and he hasn't played many games, his sigma dropped dramatically to about 50. His "ranking value" after factoring in MMR and uncertainly would then be 2750 which might be enough to hit top 200. It's still an anomaly, but something like that would explain it.
Someone who is 7-0 would never get matched up with someone like IdrA.
|
There is no good way to really determine a players true skill level using ladder points for various reasons already mentioned. All these rankings are more or less for entertainment purposes and you can't read too much into it.
At http//:ps.sc2pf.com I added an additional sorting method. MScore (modified) is an experimental sorting option that uses a combination of points, win percentage and server to rank the players. While not a perfect way to rank players, it does give a different view and seems to do a better job of ranking players IMO then just using points. Once again, this is just a reflection on how well a player has done in ladder games and not true skill level.
This is from Tues data.
Notable changes. dayvie is dropped out of the top 10, Idra is put back into the top 10.
Sort by MScore (modified)
1 IdrA.US
2 HuK.US
3 roxkisBratOK.EU
4 KiWiKaKi.US
5 ClouD.EU
6 qxc.US
7 JunwiPrime.KR
8 oGsEnsnare.KR
9 TTOne.US
10 KilluaPrime.KR
Sort by Points.
1 HuK.US
2 TTOne.US
3 dayvie.US
4 roxkisBratOK.EU
5 요츠바랑.KR
6 ClouD.EU
7 mTwDIMAGA.EU
8 KiWiKaKi.US
9 JunwiPrime.KR
10 ajtls.US
|
United States12278 Posts
On August 12 2010 02:13 wuddersup wrote:Show nested quote +On August 12 2010 02:08 Excalibur_Z wrote:
Expanding on MMR - sigma * 3 for a moment to try and explain the #200 guy who went 7-1, let's say he went 7-0 and his MMR jumped to like 2900 with a sigma of about 500. Then he lost to Idra who has an MMR of 3000, and because the result was expected and he hasn't played many games, his sigma dropped dramatically to about 50. His "ranking value" after factoring in MMR and uncertainly would then be 2750 which might be enough to hit top 200. It's still an anomaly, but something like that would explain it. Someone who is 7-0 would never get matched up with someone like IdrA.
floor exercise's post on page 4 says that he won against Tozar and lost to Idra.
|
On August 12 2010 01:00 Dyno. wrote:
We've known since early beta that points are not of equal value across all divisions.
It was made especially apparent when there used to be a bug that allowed you to be "promoted" from one platinum division to another (before diamond existed). Players would often see a drastic rise or fall of their rating.
Source? I thought it was agreed that ranks were not equal across all divisions. Points are not exact across all divisions in the same league, but would more or less be a solid indicator of rank in that league. Since you are not pitted against people only from your division but rather matched based on rating (points) in your division, it can be assumed that points would have to be rather equal for this to work.
The bug you're referring to was, indeed, a bug. I'm pretty sure a blue post was made to address that issue.
|
Nerd rage in this thread is off the charts.
|
On August 12 2010 02:17 Arcalious wrote:
There is no good way to really determine a players true skill level using ladder points for various reasons already mentioned. All these rankings are more or less for entertainment purposes and you can't read too much into it.
At http//:ps.sc2pf.com I added an additional sorting method. MScore (modified) is an experimental sorting option that uses a combination of points, win percentage and server to rank the players. While not a perfect way to rank players, it does give a different view and seems to do a better job of ranking players IMO then just using points. Once again, this is just a reflection on how well a player has done in ladder games and not true skill level.
This is from Tues data.
Notable changes. dayvie is dropped out of the top 10, Idra is put back into the top 10.
Sort by MScore (modified)
1 IdrA.US
2 HuK.US
3 roxkisBratOK.EU
4 KiWiKaKi.US
5 ClouD.EU
6 qxc.US
7 JunwiPrime.KR
8 oGsEnsnare.KR
9 TTOne.US
10 KilluaPrime.KR
Sort by Points.
1 HuK.US
2 TTOne.US
3 dayvie.US
4 roxkisBratOK.EU
5 요츠바랑.KR
6 ClouD.EU
7 mTwDIMAGA.EU
8 KiWiKaKi.US
9 JunwiPrime.KR
10 ajtls.US
How can you rank people across regions? Doesn't seem like a very good idea.
|
(I clicked on this because Charade is my bnet name. Hahah)
I do not understand why they would rank by a different system than the one they use. :\
|
This is so deja vu of exactly the reaction when the new MMR system was released in WoW. Everyone has an idea of how it 'should' work and then when it doesn't work that way they get all pissy instead of focusing on how it actually works.
Ignore your points, ignore your league, ignore your division. None of it means anything, it's all there for casuals to give them the sense of progression, the never-ending carrot on the stick that worked so successfully for them in WoW.
For competitive players, MMR is all that matters. You can tell exactly where you stand by who you are facing. You might be in platinum, but you just queued up and at the loading screen your opponent is HuK.. you are obviously way behind your MMR and the system is still catching you up. This is the exact same way as WoW when you start a new arena team you start at 0 rating but your MMR might be 2900.. because the system knows you are a pro, you will start your first game off playing against the top players in the world instead of playing 100 games smashing newbies until the dirt until you get where you should be.
I'd put my yearly salary on that 7-1 guy playing top players for all 8 of those games. The system saw his 2v2 league, his 3v3 league, etc and saw he was a very good player. He didn't go 5-0 in placement by stomping bronze leaguers. He lost to IdrA chances are all of his wins were against top diamond players as well. Now if he had went 1-7 instead of 7-1 his MMR would drop like a rock and he'd end up where he needs to be.
The system works, give it time.. the game just came out 2 weeks ago. That's a little fact a lot of people seem to be ignoring, any ranking system of any type will get more accurate the more time goes by.
|
|
|
|
|
|
EPT
