|
On September 22 2017 16:21 TT1 wrote:
Depends what hours you're playing on and who you're facing.
I really don't think it is dependent on your play time, I often start winning a lot in the late night EST which is when Koreans get online, I think the skill level at Korean hours is more predictable, if you're 1700, you're likely to hit a legit 1700. I was D+ back on iCCup and I hover around 1800-1900 (went as high as 2k), but since people have figured out they can disc to avoid the loss, I'm just dropping like crazy. So I stopped playing until it's patched.
|
The points are whack.... I was C on iccup with P. Recently switched to terran and having played more than 250 games atm.
Sometimes when I log in I meet only worse players winning all my games. Other times I dont stand a chance. My rating fluctuates wildly between 1600-1900. I can literally log out at 1700 and when I login again I have gained or lost 200 points for no apparent reason.
|
On September 22 2017 15:45 Qikz wrote:Show nested quote +On September 22 2017 09:39 castleeMg wrote:
heres the thing, iccup ratings were meaningless after the admins changed the system. before the change, i think the highest rated players were around B... after the change there were about 10 or so A+ and 40 players A or A-. but the problem was some losers would pick and choose their opponents and only play vs D in order to achieve an A- rank, while others played the ladder legitimately and were probably more skilled than the abusers but received a much lower rank. in conclusion dont compare this blizzard ladder to iccup ratings because iccup rating was a very poor way to determine actual skill near the end of their modified ranking system imo at least. The thing is man is iCCUP ratings were meaningless BEFORE the change as well for the exact same reason. Using any form of ranking system as a determination of skill is silly. Day9 made a video about this and it was pretty good. People take their rating far too seriously.
what is your skill if not your ability to win at a game? i think ranking system determines your ability to win pretty fairly
|
On September 22 2017 15:45 Qikz wrote:Show nested quote +On September 22 2017 09:39 castleeMg wrote:
heres the thing, iccup ratings were meaningless after the admins changed the system. before the change, i think the highest rated players were around B... after the change there were about 10 or so A+ and 40 players A or A-. but the problem was some losers would pick and choose their opponents and only play vs D in order to achieve an A- rank, while others played the ladder legitimately and were probably more skilled than the abusers but received a much lower rank. in conclusion dont compare this blizzard ladder to iccup ratings because iccup rating was a very poor way to determine actual skill near the end of their modified ranking system imo at least. The thing is man is iCCUP ratings were meaningless BEFORE the change as well for the exact same reason. Using any form of ranking system as a determination of skill is silly. Day9 made a video about this and it was pretty good. People take their rating far too seriously.
I disagree that a ranking system determining skill is silly. It's really not. I can give many examples.
in 2009 on iccup if a B player played a C player, you knew who would win just by looking at the rank (assuming no smurf). There is no doubt in anyone's mind that a B player that is 150-30 or whatever vs a C player with 150-100.
But let's not just count iccup, you could say the same with League or SC2. If a diamond player plays a GM player, you know who's going to win regardless of the match up.
Now at the absolute highest levels of play I can agree that your skill rating doesn't matter. But for 99% of the players? Yes it shows and does show your skill level to another player.
|
there might be a hidden MMR in the new ladder system. purely anecdotal but i was fishing for a PvZ on a smurf and auto-quit like 10 games in a row vT and vP. dropped my visible rating from 1900 to <1500 but i was still playing >1800 players when i finally got a Z matchup.
|
Norway28702 Posts
On September 23 2017 05:19 blade55555 wrote:Show nested quote +On September 22 2017 15:45 Qikz wrote:On September 22 2017 09:39 castleeMg wrote:
heres the thing, iccup ratings were meaningless after the admins changed the system. before the change, i think the highest rated players were around B... after the change there were about 10 or so A+ and 40 players A or A-. but the problem was some losers would pick and choose their opponents and only play vs D in order to achieve an A- rank, while others played the ladder legitimately and were probably more skilled than the abusers but received a much lower rank. in conclusion dont compare this blizzard ladder to iccup ratings because iccup rating was a very poor way to determine actual skill near the end of their modified ranking system imo at least. The thing is man is iCCUP ratings were meaningless BEFORE the change as well for the exact same reason. Using any form of ranking system as a determination of skill is silly. Day9 made a video about this and it was pretty good. People take their rating far too seriously. I disagree that a ranking system determining skill is silly. It's really not. I can give many examples. in 2009 on iccup if a B player played a C player, you knew who would win just by looking at the rank (assuming no smurf). There is no doubt in anyone's mind that a B player that is 150-30 or whatever vs a C player with 150-100. But let's not just count iccup, you could say the same with League or SC2. If a diamond player plays a GM player, you know who's going to win regardless of the match up. Now at the absolute highest levels of play I can agree that your skill rating doesn't matter. But for 99% of the players? Yes it shows and does show your skill level to another player.
The thing is that at least for every ladder before this one, it has been possible to inflate your ladder score through various means. On iccup, you'd have some people who would only play one matchup on one particular map (like say, only pvt longinus), most ladders had more points for winning than losing, so that going 200-100 would give way more points than going 100-50 against the same group of players, people could exclusively go for strategies that are great in a best of 1 (like zergs perfecting 3 hatch ling allins) and not play regames.. a LOT of people would look at the stats of other players and dodge playeers with great win/loss ratios. (iccup removed the stats option ingame for this reason, but people would still look at it when asking for games in channels) For people who did none these, you could totally say that an a- player was better than a b+ was better than a b etc. But these types of rating inflations were extremely common, so common that there were certainly a lot of players who would have like, b+ max rating but where their actual skill level would be more like, b-.
|
I've also noticed that US West seems significantly harder than US East. I'm 1455 on US West, where I've played most of my games (around 100), but I started playing a few games on US East and noticed that I'm winning almost every game. I'm currently 1715 on US East, and rising every night.
The people at 1500 on US West actually seem quite good, with real builds and 200+APM. At 1700 US East, I'm still seeing people that have no idea what they're doing, dying to my first few units and such. I think what's happening is that I'm hitting a bunch of Koreans on US West; I see a bunch of Korean names and it seems like a number of people I play don't understand English.
|
Norway28702 Posts
I think you should run into the same opponents regardless of the server you are playing on.. It makes sense that your rank is higher on useast than uswest, but that the ratings differ should be random, unless I'm getting something wrong.
|
|
|
On September 23 2017 06:15 Liquid`Drone wrote:Show nested quote +On September 23 2017 05:19 blade55555 wrote:On September 22 2017 15:45 Qikz wrote:On September 22 2017 09:39 castleeMg wrote:
heres the thing, iccup ratings were meaningless after the admins changed the system. before the change, i think the highest rated players were around B... after the change there were about 10 or so A+ and 40 players A or A-. but the problem was some losers would pick and choose their opponents and only play vs D in order to achieve an A- rank, while others played the ladder legitimately and were probably more skilled than the abusers but received a much lower rank. in conclusion dont compare this blizzard ladder to iccup ratings because iccup rating was a very poor way to determine actual skill near the end of their modified ranking system imo at least. The thing is man is iCCUP ratings were meaningless BEFORE the change as well for the exact same reason. Using any form of ranking system as a determination of skill is silly. Day9 made a video about this and it was pretty good. People take their rating far too seriously. I disagree that a ranking system determining skill is silly. It's really not. I can give many examples. in 2009 on iccup if a B player played a C player, you knew who would win just by looking at the rank (assuming no smurf). There is no doubt in anyone's mind that a B player that is 150-30 or whatever vs a C player with 150-100. But let's not just count iccup, you could say the same with League or SC2. If a diamond player plays a GM player, you know who's going to win regardless of the match up. Now at the absolute highest levels of play I can agree that your skill rating doesn't matter. But for 99% of the players? Yes it shows and does show your skill level to another player. The thing is that at least for every ladder before this one, it has been possible to inflate your ladder score through various means. On iccup, you'd have some people who would only play one matchup on one particular map (like say, only pvt longinus), most ladders had more points for winning than losing, so that going 200-100 would give way more points than going 100-50 against the same group of players, people could exclusively go for strategies that are great in a best of 1 (like zergs perfecting 3 hatch ling allins) and not play regames.. a LOT of people would look at the stats of other players and dodge playeers with great win/loss ratios. (iccup removed the stats option ingame for this reason, but people would still look at it when asking for games in channels) For people who did none these, you could totally say that an a- player was better than a b+ was better than a b etc. But these types of rating inflations were extremely common, so common that there were certainly a lot of players who would have like, b+ max rating but where their actual skill level would be more like, b-.
While true there was abuse, I don't think the majority did this. Like anything else there are the abusers/cheaters, but for the most part I would say it's pretty accurate. I do know I dodged ZvZ hardcore back then and would today if that was how I was playing.
This one would be more accurate I think if it worked lol. Points are so sporadic and it is annoying in the sense that I don't know what my MMR should be. I fluctuate from 1950 to 2160. I'm pretty sure that's also because games are getting counted twice. I'll see after a game I go from 2160 to 2130, but after the next game I play it shows 2100 or 2080 or something like that, same with wins. I will update to 2000 and end up being at 2030 or higher sometimes.
I am a bit surprised the match making hasn't been fixed yet...
|
Norway28702 Posts
I think a very large portion of people did some degree of it. Like everyone would host the motw they liked the most rather than randomly distribute between them. Most people would choose the guy with 70-60 stats rather than the guy with 38-2 if both were asking for game in the channel. Having random match making with random maps makes it very different, and I totally agree that if the remastered ladder becomes good, then these ratings will be meaningful. I'm not saying that iccup ratings weren't meaningful either.. The way I see it is, they used to be kinda like how BMI is for telling whether someone is in good shape or not. Certainly on average a guy with 22 bmi is in better shape than someone with 30 - for determining the average of a large group, it's a fairly good measurement. But it's not reliable when comparing two particular individuals. 
|
Just to give personal data, I was C- on iCCup and I am currently around 1750-1800 mmr with 12-4, so I expect to go a little higher since i'm not anywhere close to 50% winrate atm. I'm surprised there is still so many bugs related to the stats and pts system though.
|
On September 23 2017 05:19 blade55555 wrote:Show nested quote +On September 22 2017 15:45 Qikz wrote:On September 22 2017 09:39 castleeMg wrote:
heres the thing, iccup ratings were meaningless after the admins changed the system. before the change, i think the highest rated players were around B... after the change there were about 10 or so A+ and 40 players A or A-. but the problem was some losers would pick and choose their opponents and only play vs D in order to achieve an A- rank, while others played the ladder legitimately and were probably more skilled than the abusers but received a much lower rank. in conclusion dont compare this blizzard ladder to iccup ratings because iccup rating was a very poor way to determine actual skill near the end of their modified ranking system imo at least. The thing is man is iCCUP ratings were meaningless BEFORE the change as well for the exact same reason. Using any form of ranking system as a determination of skill is silly. Day9 made a video about this and it was pretty good. People take their rating far too seriously. I disagree that a ranking system determining skill is silly. It's really not. I can give many examples. in 2009 on iccup if a B player played a C player, you knew who would win just by looking at the rank (assuming no smurf). There is no doubt in anyone's mind that a B player that is 150-30 or whatever vs a C player with 150-100.
But what if that B player will never play any C player, except ones he carefully chooses? That's how he got that inflated ranking. And now he can boast around how he's B. Which is silly.
Or what if that B player only does one all-in per matchup, but knows those really well. In rematches he doesn't stand a chance, but on this relatively anonymous 1v1 ladder, he scores high. He would never do a rematch. Still he can boast around how he's B. Which is silly.
It's not bad, it's not cheating, but it is silly.
(The right way to approach a ladder is as a means to practice. With auto-matchmaking, the rating number could even be hidden. So no body knows. So there's no reason to do silly things to get on top.)
|
The points don't matter. Right now, it's like asking who has the biggest rock collection. And I don't mean shiny rocks. I mean like just regular, small, uninteresting rocks. B- players are at 1800, D+ players are at 2000... it's just completely unreliable as a measure of skill.
|
On September 23 2017 07:48 Liquid`Drone wrote:
I think you should run into the same opponents regardless of the server you are playing on.. It makes sense that your rank is higher on useast than uswest, but that the ratings differ should be random, unless I'm getting something wrong.
Well maybe I'm wrong but I believe, you get matched most often against people who are close to you if a match can quickly be found near your MMR with a nearby player. So especially while you are not near the top percents of the MMR ranks, you would get matched against ppl of your server a lot, and the ratings may scale different per server. Until they get adjusted differently by playing a large proportion of games against players of other servers at some threshold.
|
United Kingdom12023 Posts
On September 23 2017 16:40 ninazerg wrote:
The points don't matter. Right now, it's like asking who has the biggest rock collection. And I don't mean shiny rocks. I mean like just regular, small, uninteresting rocks. B- players are at 1800, D+ players are at 2000... it's just completely unreliable as a measure of skill.
Even the old system was. I find it hilarious that it's taken until now for people to realise this.
|
Norway28702 Posts
On September 23 2017 19:10 ProMeTheus112 wrote:Show nested quote +On September 23 2017 07:48 Liquid`Drone wrote:
I think you should run into the same opponents regardless of the server you are playing on.. It makes sense that your rank is higher on useast than uswest, but that the ratings differ should be random, unless I'm getting something wrong. Well maybe I'm wrong but I believe, you get matched most often against people who are close to you if a match can quickly be found near your MMR with a nearby player. So especially while you are not near the top percents of the MMR ranks, you would get matched against ppl of your server a lot, and the ratings may scale different per server. Until they get adjusted differently by playing a large proportion of games against players of other servers at some threshold.
But whether they are close to you should be irrelevant of server if you yourself is changing servers, right? If someone lives on the west coast of US they should get the same opponents regardless of whether they themselves choose USEast or USWest to play on. I live in Europe, I should be matched first against european opponents regardless of whether I play on west or europe.
|
Netherlands4511 Posts
Your opponents are determined by your geo-ip location and what server you launch the game with from the b.net app.
US East or US West doesn't make a difference.
I think this ladder system, once it's fixed, is a good way to determine skill. But there really isnt any need for win streak bonus points, since it just blows people up temporarily who are on a hot streak. It does very little, these players will likely just go on a losing streak at another point and reach their actual skill rating.
First it'd be great if the disconnect problem, the double wins, double loss problems could get fixed. Then we can discuss what MMR says about your skill.
|
I was C/C+ one time even B- in last 2 years so for me it looks more like
1900=C 2000=C+ 2100=B-/B 2200B+ 2300A-
|
There is no correlation at the moment.
|
|
|
|
|
|
EPT
