Srsly that page is trash.
I mean cmon, some random guy who renamed himself DSR in Danger is the best player? LuLz
It wasnt posted here because its basically a random kid making a random page. Dota 2 does not release enough information to gather any form of ranking system so anyone trying to say he has created one is just talking out of his own ass
http://www.reddit.com/r/DotA2/comments/1rnt1z/looks_like_the_new_skill_measuring_service_is_up/
The creator is active in this thread.
The top two are Mason and I think Broodstar from Nadota
Dendi got 1445 cos he don`t stack and have just 56% win rate. Almost all other pros stack and usually have 65% + and better this rating thats why.
BurNIng only has 1698 and he has about the same win rate as Loda. I wonder how his "algorithm" works.
the creator prob just put too much focus on the wrong stats (winrate) which gave him odd results. There just is not enough info given out.......
I know but this system is not mmr is dsr . Just volvo know mmr. Maybe he have some better kda ,xpm,gpm etc. than Burning.
Btw someone know after how much game this rating reset again? I checked and was 1044,than again checked after 2 games and same.
1399 (main is a little bit higher), get good nubs. Of course you are going to complain when you get a low rating haha.
Look, first off, its not an accurate measurement, and the creator knows that. He has tried to create a system which is the most accurate with the information available, unfortunately with so many hidden profiles and a rubbish match making system (ie. making teams which huge gaps in ratings together), its not going to be that accurate. What it does do though, its give you a rough idea of where you stand, obviously if you mess around very often, ie. Dendi, your rating isn't going to be as high as others. And while he did take into account stacking, its important to understand that they will come out on top in the long term where they are winning consistently.
Win rates are a much worse indicator of skill than this. Its who your wins and losses are against that is important - and this is where volvo's crap MM comes into play. You losing those ridiculously unbalanced matches will not have a significant impact on your rating, while winning those unbalanced in your favour matches, won't either. Its winning the games where you shouldn't be, or its even that give you the most points. I see noobs with 53-55% win rates every day, and I've also played with some ~50% players that are significantly better than them. I remember in HoN, this guy was able to control his rating points by stacking and losing intentionally, he was able to stay at 1550 (starting point is 1500) with a 70%+ win rate over 300 matches...
edit: lmfao this thread is just win rate win rate win rate, its NOT. Yes there will be correlation between a higher win rate and your rating, but what is important is who your wins and losses are against, and with volvo's shitty match making system, win rate for the most part, is useless
A majority of Terrible's games are played in normal/high a few in very high if someone in his party had a high enough MM. There is a serious flaw if the system puts Terrible at the level of a pro when the highest games on that account are in the very bottom of very high(mostly high-normal)
Just because someone makes a algorithm does not mean it is correct.
ROLF, i have an account with a really bad winrate and shit kda´s overall where i played with a friend ,it was not even a scrub account, i was really new to the game when i played with that account: 1100 DSR
With this one: 880 when i have way better winrate and kda overall i am pretty sure that i play against harder oponnents.
My friend who plays this game really casual and plays a lot limited hero pool ( because he is actually new ) and his winrate is less than 50 % got a 1200 :DDD
This is so ramdon its not even funny.
All my matches were solo queue, and pubstats says otherwise about all my matches being in the normal bracket lmfao: http://pubstats.me/130472516/
oh wait, its the NOBODY CAN BE BETTER THAN ME thread
http://dotabuff.com/matches/188693133
http://dotabuff.com/matches/188722807
http://dotabuff.com/matches/188732351
http://dotabuff.com/matches/188767190
this is where you jump all the way to high
http://dotabuff.com/matches/188800656
http://dotabuff.com/matches/188964183
http://dotabuff.com/matches/189613418
you get the point even on your first page it still seems like you que into high
http://dotabuff.com/matches/225901628
http://dotabuff.com/matches/222887004
Even if as you say all your matches are solo que that means you are a high bracket player who every now and then can get into vhigh if the que is dead
you have never once played against anyone near the 1st page (on that account) yet you are higher ranked then most pro players according to the site. Do you honestly think that is right?
I dont think that even the ones in the top 10 in that page thinks that the system is even close to be accurate.
So if the majority of my recent matches are in VH, that means that I am a high bracket player. Well done. And apparently stomping in most of the matches which were in the high bracket, or doing decently even if the team fed makes me a bad player. lol?
I didn't play in either of those slark matches (teaching a friend at LAN), and I left that LD match after 30-35 minutes.
edit: and considering that I said that it was an inaccurate measurement in my first post too. ha. so much butthurt
"He has tried to create a system which is the most accurate with the information available"
your words not mine
the creator took data and used what he thought was correct but was mistaken in how stats can be manipulated so it is not the most accurate if the formula is inherently wrong
The thing is that the information is so little that there is no point on even trying to create a system. How can i be on a 40 % winrate acount on normal braket agaisnt really bad players with 1100 dsr ?
I guess this is just to have a little bit of fun, i think 3 or 4 month early someone tried to create something like this, i dont evenr remenber wtf happened to that system.
Yep, just as I thought. mad cuz bad.
Note the word "tried", I didn't say it was the most accurate, and it doesn't matter, because you can't create an accurate rating system when so many people have their information missing and the match making is this poor (which will heavily influence the weighting of a lot of matches, as opposed to a good match making system, where the skill level of other players in a match could give you a decent idea of what kind of skill level the match is.
"He has tried to create a system which is the most accurate with the information available"
"I didn't say it was the most accurate"
-Terrible
You didn't pass first grade English, did you? Stop playing DotA and start paying attention in school.
" What it does do though, its give you a rough idea of where you stand,"
It doesnt. Only a true system would give you a rough idea of where you stand, you tried to defend it a little bit, and its really, really bad so there is no sense in try to defend this.
Whatever, i really like your ramdon troll posts here dude, waiting for the next one <3
New feature: tier lists w/ weird names
Top 30% you are notified and your score is now a 'gold DSR'
Top 60% = Silver
Top 80 (lol) = Bronze
Top 0.1% = Mithril
Top 3% = Pro
Top 1% = Jadeite
Top 10% = Platinum
here's a guide how to get high dsr:
make a new account
stack with 5 to 30-0 or so score with high kda and gpm
enjoy 1800+
here's a guide how to get high valve elo rating:
make a new account
stack with 5 to 30-0 or so score with high kda and gpm
enjoy 1800+
ps you're dumb as shit ^___^
A random guy makes his own MMR system which he doesn't even reveal the formula of and I am expected to believe it? This makes even less sense then everyone in this forum.
well i mean that's just untrue, you'd see constantly trashcan 5 man stacks who play on russia servers at odd times on the front page then but that doesn't happen now does it?
which means you're dumb as shit, just like this rating system
You are retarded.
Show me one profile of someone who got 30-0 while stacking with 'russian trashcans' that won't appear on the front page when he finally goes to play alone.
I don't know why everyone pretends Valve system is any good when it comes to those imaginated cases. It's not. It's equally retarded and can be gamed easily.
Zaloguj się aby dodawać komentarze.
http://mydotaskill.com/
EDIT: More information from the creator here:http://www.reddit.com/r/DotA2/comments/1rnt1z/looks_like_the_new_skill_measuring_service_is_up/
One notable statement the creator made is that heroes are compared so that supports aren't screwed over by the use of some of his metrics. This results in some seemingly weird-ass ratings like that guy who only plays Broodmother who has a really high rating that's higher than many pros. The thing is, maybe he's really good at Broodmother compared to the few who play BM extensively (including pros) and since there isn't much data on other heroes relative to his Broodmother he has a high rating.
For those of you bring up certain pro players' ratings and using this 'evidence' to dismiss the entire system: some pros don't stack as much in normal matchmaking and some do. A lot of them play in-houses or in leagues such as IXDL; these games aren't added to statistics.
FYI the top rated player is Mason who is well known on NADota and was on Take5 with Arteezy and Demon before it was DQed from MLG qualifiers. He stacks a lot.