EPTWith regards to the queue, it jumped from 550 profiles/hour to 6000 profiles/hour which is pretty neat.
Not claiming it's accurate at all but at least it's improving.
General Discussion - Page 4453
| Forum Index > Dota 2 General |
IRC chatter should remain in IRC - http://webchat.quakenet.org/?channels=tl.dota2 Posts that relate to topics with their own thread, such as in News, Tournaments or Strategy should go in those threads. | ||
|
Laurens
Belgium4558 Posts
With regards to the queue, it jumped from 550 profiles/hour to 6000 profiles/hour which is pretty neat. Not claiming it's accurate at all but at least it's improving. | ||
|
solidbebe
Netherlands4921 Posts
| ||
|
-Celestial-
United Kingdom3867 Posts
On December 04 2013 20:50 Laurens wrote: Not claiming it's accurate at all but at least it's improving. Its definitely nowhere near accurate. I have a few accounts and out of interest I checked two of them. One has about eight times the DSR rating of the other. | ||
|
Staboteur
Canada1873 Posts
My buddies I play with have a range from ~550 to ~1300 rating, which suggests that I'm twice as good at what I do as compared to half my team. That's -absolutely ridiculous- | ||
|
TheKefka
Croatia11752 Posts
On December 04 2013 20:34 Jetaap wrote: Yeah about that, this DSR thing is such a joke but i have a better score than the terrible friends i stack with so i'm going to say it's totally legit for now. I have 1200 dsr,is that like average?I see the top people having like 1.9 k or something | ||
|
Jetaap
France4814 Posts
On December 04 2013 21:37 Staboteur wrote: Yeah no not accurate. My buddies I play with have a range from ~550 to ~1300 rating, which suggests that I'm twice as good at what I do as compared to half my team. That's -absolutely ridiculous- While I agree with you that DSR is useless and doesn't represent skill accurately, having someone with twice as many points doesn't mean he is twice as good. For example in starcraft 2 you wouldn't say that a 1000 points master player is twice as good as a 500 point master player. edit: We cannot say what's a good rankign/average etc without having an idea of the distribution of the data, but the owner is not providing that so it's just a meaningless number. I have a 750 rating, does it mean i'm in the top 25%? or bottom 25? It could be anything so it's impossible to say if a 1200 score is good or not | ||
|
Asha
United Kingdom38257 Posts
On December 04 2013 20:50 Laurens wrote: It seems to be improving with every update though. Initially supports got shafted and there was obviously a stupid way of calculating the DSR that put supports at a disadvantage, now that is no longer the case. I got trenched after the last update and I play more support than anything else lol! well, even if I don't need a rating to tell me I'm terrible, it's fun to see someone try to develop one and then actually work on improving their system rather than abandoning it when it has issues. | ||
|
Jetaap
France4814 Posts
| ||
|
Staboteur
Canada1873 Posts
On December 04 2013 21:54 Jetaap wrote: While I agree with you that DSR is useless and doesn't represent skill accurately, having someone with twice as many points doesn't mean he is twice as good. For example in starcraft 2 you wouldn't say that a 1000 points master player is twice as good as a 500 point master player. edit: We cannot say what's a good rankign/average etc without having an idea of the distribution of the data, but the owner is not providing that so it's just a meaningless number. I have a 750 rating, does it mean i'm in the top 25%? or bottom 25? It could be anything so it's impossible to say if a 1200 score is good or not True! According to this rating system, I am closer to the relative skill level of Mushi (albeit 10 months ago Mushi) at a difference of 50 points than I am to the next best player I play with, at a difference of ~300 points. The difference in skill between me and Mushi is absolutely massive, whereas the difference in skill between me and my buddy is tiny, and I honestly can't say if the difference goes my way or his. Also useful data is that the caster, GoDz (1889), is ranked higher than most of the players he will cast. You're right that "Twice as good" doesn't apply if it isn't a linear metric, but the system is ridiculous even if the slope isn't as steep as I think it is, but it quickly fades into complete irrelevance if even a 500 rating cannot be said to be unquestionably inferior to a 1000 rating. | ||
|
Jetaap
France4814 Posts
On December 04 2013 22:09 Staboteur wrote: True! According to this rating system, I am closer to the relative skill level of Mushi (albeit 10 months ago Mushi) at a difference of 50 points than I am to the next best player I play with, at a difference of ~300 points. The difference in skill between me and Mushi is absolutely massive, whereas the difference in skill between me and my buddy is tiny, and I honestly can't say if the difference goes my way or his. Also useful data is that the caster, GoDz (1889), is ranked higher than most of the players he will cast. You're right that "Twice as good" doesn't apply if it isn't a linear metric, but the system is ridiculous even if the slope isn't as steep as I think it is, but it quickly fades into complete irrelevance if even a 500 rating cannot be said to be unquestionably inferior to a 1000 rating. GoDz is an ex pro HON player if i'm not mistaken, so that's not that outrageous that he would have a very high DSR. I don't think pro's rating is really that relevant since a lot of them don't play a lot of pubs, or don't play them seriously at all. If a pro wants to play serious he will scrim, whereas someone like godz will probably play a pub. | ||
|
TimoS1703
Germany250 Posts
but it quickly fades into complete irrelevance if even a 500 rating cannot be said to be unquestionably inferior to a 1000 rating. Right, got a pretty high DSR and most of my friends have a score that is 300-400 points lower even though I'm sure that some of them are better players then I am, that means more consistent and if we'd play 100 1v1's they would win a majority of the games (ok you can't compare 1v1 to the performance in a 5v5 but anyway...). I don't get why this page doesn't give us more information on how this rating ist computed. The community could help to improve things. But in the meantime you really can't take this rating serious... | ||
|
teddyoojo
Germany22369 Posts
| ||
|
Asha
United Kingdom38257 Posts
| ||
|
BurningSera
Ireland19621 Posts
On December 04 2013 22:44 teddyoojo wrote: its funny that VG is the best team in china now that they have sylar #ti2carryoftheyear forever top3 carry in my heart. | ||
|
teddyoojo
Germany22369 Posts
| ||
|
Gosi
Sweden9072 Posts
siQ | ||
|
hariooo
Canada2830 Posts
On December 04 2013 21:54 Jetaap wrote: While I agree with you that DSR is useless and doesn't represent skill accurately, having someone with twice as many points doesn't mean he is twice as good. For example in starcraft 2 you wouldn't say that a 1000 points master player is twice as good as a 500 point master player. edit: We cannot say what's a good rankign/average etc without having an idea of the distribution of the data, but the owner is not providing that so it's just a meaningless number. I have a 750 rating, does it mean i'm in the top 25%? or bottom 25? It could be anything so it's impossible to say if a 1200 score is good or not 750 means you're in roughly the top 20-25% according to the reddit post about this. Different sites are having different reactions to DSR, so it's kind of telling that no one has posted their DSR on TL yet. http://mydotaskill.com/?id=64299123 Though it's kinda useless until I can see what my friends have. Edit: nvm it's being recalculated lol | ||
|
Steveling
Greece10806 Posts
| ||
|
Excludos
Norway8279 Posts
On December 04 2013 23:35 Steveling wrote: I was away for some months and what's DSR? Its a system that takes the 3 last digits of the number plate on the last car you drove in and multiplies it by two. And apparently thats how good you are in dota 2. | ||
|
PassiveAce
United States18076 Posts
| ||
| ||
