I guarantee 65% chance to predict the winner on average.
Do you understand? Or it's still too hard?"
You just don't understand that it's not the same thing, read again what I wrote. And second sentence here is wrong.
On average (if their system is perfect) your real odds to win a game is 50%, it mean some games it's less and some games it's more. What you do is your prediction of the winner team is right in 65% of the games, that is simply not the same thing.
I saw uper that you asked advice from math/stats guys but when you have one you refuse to listen. I'm pretty sure that most of people here could predict the winner more than 65% of the times just via pure intuition if they had the same informations than you on both team (i.e just not their name but their total games, win ratio....).
If you think your system is perfect just don't listen what others person say and assume your always right but there is one thing certain is that there is some big swing in most of the players rankng, so maybe a player can become from shit to pro in 3 days, or maybe your system scoring isn't just that great.
"Be a team player, as a carry - help your teammates with ganks and don't die.
As a support - make sure your carries are safe and have space to farm.
Provide vision for your team and help with ganks.
Playing well WILL result in a higher DSR score.
Playing selfish will decrease it.
"
Now about what you write on your website it's not that clear to me do you have any stats that show how much supports buy wards in their game or it's just a few sentences without link to your rating. Also how you measure someone playing selfish ?
Will you only ever be able to measure non-ranked MMR? I see your system still being useful even with Ranked MMR coming. Because valve will continue to hide non-ranked MMR.
I don't know why so many people have trouble understanding the 65% prediction argument. Maybe if you will get to the probability/combinatorics in math ( that seem to be taught in high-school ).
Elo can predict around 50% of wins. As you try to balance the matches around that - giving both teams, on average, 50% chance of winning.
With other approaches, there is more information extracted than simple Elo rating and this supposedly can give higher predictive power as who will win the game. Since you could balance the game around that or maybe around that AND Elo rating, the games could be more balanced.
I don't believe that this Elo system makes matches when one team has 65% chance of winning and the other 35% - In that case, simple exchanging two players between the two teams should balance the game more to 50% ( 52% - 48%, 55% - 45% win chance in extreme cases ) If it was, this is how forced 50% win-rate would look like and I am sure Valve doesn't use that.
Yes, I will still keep the system operational to give ratings to every player. Additionally, I will tweak the formula to increase accuracy ( by picking winners more often )
@fearohnum I though that one one only applies to MDS users. The entire player base peaks at 400, where as MDS users top at 700. Given that a very small minority of the player base use MDS... etc
either way, If you are going create named brackets, you should provide the percentage for said brackets.
From the top of my head, I could use those parameters:
Kills
Deaths
Assists
KD
KA
DA
KDA
GPM
XPM
Last hits
Denies
Tower damage
with each hero vs the averages and checking where you play with each hero ( so people from 1500 dsr wont stack with lower rated friends to play at 300 dsr and get high stats - I will actively look at each game to make sure you've played that hero in high bracket ).
I'm not sure if you're trolling on purpose in this post or not ?
http://i.imgur.com/HPBoork.png it looks quite oriented to the right for a gaussian distribution to me, and I won't comment the rest of your first sentence.
When I read that you used some SVM tools I thought you had some statistics classes in some university as it's some kind of advanced stuff, I learned statistics in a mathematics university for a bachelor (not sure if it's the correct translation basically the diplome before a phd).
I don't remember well what is teach in probability in high school, but if it's the kind of stuff you used it could explain why you don't really understant what you are doing.
"Elo can predict around 50% of wins." I don't know how the Elo system works but you must be a complete retard to predict below 50% of wins, actually it's impossible to predict below 50% of wins.
Let's assume that the Dire side and radiant side give the same amount of chance to win, and you have an infinite sample of dota 2 games where you must predict the winner. Let's also suppose that you have no others information on players in each team. Now explain me how you will predict the winning team below 50% ?
Is there any way to refresh your score? I lost 400 rating when you updated something and since then the quality of my games have been significantly better, I was curious as to how that reflects on my rating. Its been stuck on the same number for a while now, do they ever get refreshed?
It's crazy that you are so bad at statistics, mathematics and probability and still use some of their tools.
Tell me how you can achieve a prediction rate below 50%.
You wrote "Elo can predict around 50% of wins." So tell me how you can predict below 50% in Dota 2 ? Are you talking about Elo in chess ? Which is completly differents as in chess you have high chance to do a draw.
PS/ You never replied to others comments I made before.
I've never claimed prediction rate below 50%. Since all you do is take things out of context, this is my last response to you. People who actually want to improve this thing ( Like the PhD from the beginning of the thread ) give useful input. All you do is write about some unimportant bullshit. Is this because you've got a score of 400? It isn't really that bad. You can always improve.
I could not care less about my score except if I would have made this system and I was in the bot 90%. The idea to make a dota scoring is quite ambitious to say the least anyway.
Your whole gibberish started when I said here than a 65% prediction rate isn't that crazy when it is just a bit better than the worst case scenario when you predict 50% of the times the winning team. Then you started to reply agressively with insulting posts and wrong facts. I think I proved my points several times and tried to help you to correct several of your lacks of understanding.
I never wrote unimportant bullshit but tried to point some errors and you took it personnally, the more you wrote the more you showed some of these lacks of understanding.
It's ok to do some mistake as long as you correct them later, now if you think that pointing mistake isn't an useful input then keep your system as wrong as now but don't except anyone to take it seriously but naifs.
@Fearohnum // MyDotaSkill.com
my score updated once after one game: dropping from 1411 to 1350, then stay at 1350 after another 20 games, so how often does it refresh? would u mind telling me my current DSR? thx
@Forza Roma
To my best knowledge, your score didn't update after that one game. The algorithm had changed, so your score changed along with it. It seems impossible for the scores to update on their own. This is the only time the scores have been updated, its being done manually, slowly.
@Revert Death,
I think you need to learn Statistics. 50% prediction is not the worst case, it is the default case. If you can predict more than 50% then it automatically means that the matchmaking system isn't fair/working as it should.
@creator,
I would rather use hero damage instead of KA, as Kills and Assists can pretty easily be faked whereas hero damage can not.
Did you read my previous post ? The 50% numbers they gave is only a theoritical numbers (only one of their criteria), anyway you can play better if you're worry to have bad odds to win a game. Can you define "default case" in statistics, I mean a mathematic definition.
Heroes dammage can be easily faked, just take zeus.
@BE THE LEAF
Don't worry about his bad algorithme several players have proved that his system don't works, most of the "computer science" guys that try to do statistics always do the same shit anyway, they don't understand what they do.
His system is just very bad right now. Basically it just take any number out of his arse that somehow rate some of the pro players above the average players.
@Revert Death
You've posted so much in this thread that I'm starting to think you have an infatuation with the creator. And I see you are a user of the DSR website.. You are what normal people refer to as a "leech". You're also probably muted in DotA 2, as there is NO ONE who would put up with a feminine, whiny attitude such as yours.
The DSR website is what pushed valve to implement RMM, yet RevertDeath contributes nothing, he uses the DSR website, and comes on here to spew his garbage that nobody wants to hear.
Just to let you guys know - the first refresh is under way ( you can see the progress on the main page ) - In the mean time, I am creating an updated algorithm - it will look at 50 variables to determine the bracket that you play in.
So this will be basically an improved version of the bracket estimator that's online now. I think it will be the final version of bracket estimation based DSR calculation.
I will then "merge" or "combine" this bracket estimator with individual hero performance analysis ( the first version of the algorithm that I've shown on Reddit / that was first online ) to account for how well you play any given hero ( in relation to your bracket ).
How funny , yes I used it once, how would have been aware that it exist otherwise ? And I didn't know the author before this thread, sorry for you sherlock. I didn't even talked to him before he started to say nonsense.
Your just delusional what pushed valve to implement RMM is the two years complaining from players.
And sorry for you but i'm not muted in Dota 2, I won't talk about ur avatar and your top 3 and end this conversation here; if you want to say nonsense here just do it in private likethat I won't reply to you and we won't pollute this thread.
Fact is that it doesn't make sense sometimes: My rating drops from 900 to 400, the reasons he gave me reached from I played in a 5 man stack when I obviously did not for 45 games (deleted that part of the comment later) to "the players you played with had low winrates (48%)" and low KDAs. I hope I don't have to lead the discussion again about whether winrate matters a hole lot or not... KDA might matter a bit more.
He also spoke of a kinda low teammatchmaking-rating, although I never ever played a single game in team-matchmaking. I don't know what he sees in his data but he either seems to be unable to read it or just lies.
As I said in Fadis thread a few days ago however, he has a 72% winrate with high KDA and is at 6xx points. He also apparently played against pro players. Low priority dodging cannot answer this rating at all imo, especially when he points out low winrates as reason for my rating.
How it comes he says I stacked 5 man when I obviously solo queued? Then delete this part later.
How it comes the owner himself had a stupid 5500 or something like this rating and then suddenly it dropped to 600 on his own website? And hey, now it reached 260, hard times for him.
How Fadi and his friends are mostly kinda low rated when they play vs. very good players?
How random smiley face guys other account is pro (1100) now (btw rising from a horrible rating) with apparently worse farming abilities than me (400) as he said one page before? http://dotabuff.com/players/134235086 Does it mean matchmaking wants to make me carry every feeder and noob team atm because it feels like it? ...
Believe what you want, I don't give a lot on this DSR-thing at the moment and I even said that before my rating dropped from 9xx to 4xx http://dotabuff.com/topics/2013-12-07-my-account-got-disabled-from-ranking--admin-please-read And on what the owner says not more. He can predict 65% which team wins? I've seen not any kind of evidence so far. For some reason he doesn't seem to want to accept that he has to work on it quite a bit, instead points out things that show nothing or are untrue.
@Captain, you've obviously made very little effort. Fadi and his band of russian virgins play in LPQ, a lot of their matches are against extremely low skilled players, and occasionally they queue against another bunch of russian nerds stacking in LPQ without dodging..
Looking at your matches, there are a whole lot of anonymous accounts, much more so than you would expect from someone with a higher rating, and for those that I can see, none are particularly impressive..you might be performing well, but its in pretty low level matches (High bracket), which while decent compared to the whole player base, is actually a lot lower than what other people are playing at.
It seems that you are just another butthurt scrub that over estimated their own ability by using win rate to determine their skill level, which pretty much everyone knows can be incredibly deceiving, and is now desperately flailing around trying to make sense of the situation.
Look at the other guy called community toxin, his win rate/kda are all impressive, but if you took a look at his matches, you would realise he is playing in quite low level matches, and hence he got a lower rating.
What you do in matches is one thing, who its against is another. If Messi scores twice against Bayern, and Aguero slots 2 past Accrington with an identical squad and performance, the only difference being the opposition, should they both get the same rating?
Oh, and, uhh, cos it seems to be the fashion on this thread: UR SYSTM SUKS GO DIE I BETTA THAN 168 I MAD FOR NO REASON UR DISTRIBUTION NOT GAUSSIAN (even though, looking at it, it clearly is, just missing the left tail...) 2 MUCH MISS DATA DUN DU IT UNLESS 100% AKURATE SYSTM
@Captain
The only reason my ranking was so low on that account, was because i stacked with lower leveled players (irl friends) while i was in lpq. You think you are a better farmer than me? Go dream somewhere else... And btw, I get matched up with players like this guy http://dotabuff.com/players/88282736 while solo queuing, so you're probably not in a position to question my skill Level.
@ Fearohnum // MyDotaSkill.com
"Ok, they guarantee 50% chance to win on average.
I guarantee 65% chance to predict the winner on average.
Do you understand? Or it's still too hard?"
You just don't understand that it's not the same thing, read again what I wrote. And second sentence here is wrong.
On average (if their system is perfect) your real odds to win a game is 50%, it mean some games it's less and some games it's more. What you do is your prediction of the winner team is right in 65% of the games, that is simply not the same thing.
I saw uper that you asked advice from math/stats guys but when you have one you refuse to listen. I'm pretty sure that most of people here could predict the winner more than 65% of the times just via pure intuition if they had the same informations than you on both team (i.e just not their name but their total games, win ratio....).
If you think your system is perfect just don't listen what others person say and assume your always right but there is one thing certain is that there is some big swing in most of the players rankng, so maybe a player can become from shit to pro in 3 days, or maybe your system scoring isn't just that great.
PS/ http://i.imgur.com/HPBoork.png I wonder why you chosed to make such a distributionn for a scoring system ?
"Be a team player, as a carry - help your teammates with ganks and don't die.
As a support - make sure your carries are safe and have space to farm.
Provide vision for your team and help with ganks.
Playing well WILL result in a higher DSR score.
Playing selfish will decrease it.
"
Now about what you write on your website it's not that clear to me do you have any stats that show how much supports buy wards in their game or it's just a few sentences without link to your rating. Also how you measure someone playing selfish ?
@Fearohnum // MyDotaSkill.com
Will you only ever be able to measure non-ranked MMR? I see your system still being useful even with Ranked MMR coming. Because valve will continue to hide non-ranked MMR.
@Revert death
What do you mean "make distribution for a scoring system"? This is normal distribution(http://en.wikipedia.org/wiki/Normal_distribution), I didn't make it - it's how things look in nature.
I don't know why so many people have trouble understanding the 65% prediction argument. Maybe if you will get to the probability/combinatorics in math ( that seem to be taught in high-school ).
Elo can predict around 50% of wins. As you try to balance the matches around that - giving both teams, on average, 50% chance of winning.
With other approaches, there is more information extracted than simple Elo rating and this supposedly can give higher predictive power as who will win the game. Since you could balance the game around that or maybe around that AND Elo rating, the games could be more balanced.
I don't believe that this Elo system makes matches when one team has 65% chance of winning and the other 35% - In that case, simple exchanging two players between the two teams should balance the game more to 50% ( 52% - 48%, 55% - 45% win chance in extreme cases ) If it was, this is how forced 50% win-rate would look like and I am sure Valve doesn't use that.
@Libertarian
Yes, I will still keep the system operational to give ratings to every player. Additionally, I will tweak the formula to increase accuracy ( by picking winners more often )
Best regards
@Fear
Can you briefly explain how youre gonna calculate hero skill?
@fearohnum I though that one one only applies to MDS users. The entire player base peaks at 400, where as MDS users top at 700. Given that a very small minority of the player base use MDS... etc
either way, If you are going create named brackets, you should provide the percentage for said brackets.
From the top of my head, I could use those parameters:
Kills
Deaths
Assists
KD
KA
DA
KDA
GPM
XPM
Last hits
Denies
Tower damage
with each hero vs the averages and checking where you play with each hero ( so people from 1500 dsr wont stack with lower rated friends to play at 300 dsr and get high stats - I will actively look at each game to make sure you've played that hero in high bracket ).
Best regards
@Fearohnum // MyDotaSkill.com
I'm not sure if you're trolling on purpose in this post or not ?
http://i.imgur.com/HPBoork.png it looks quite oriented to the right for a gaussian distribution to me, and I won't comment the rest of your first sentence.
When I read that you used some SVM tools I thought you had some statistics classes in some university as it's some kind of advanced stuff, I learned statistics in a mathematics university for a bachelor (not sure if it's the correct translation basically the diplome before a phd).
I don't remember well what is teach in probability in high school, but if it's the kind of stuff you used it could explain why you don't really understant what you are doing.
"Elo can predict around 50% of wins." I don't know how the Elo system works but you must be a complete retard to predict below 50% of wins, actually it's impossible to predict below 50% of wins.
Let's assume that the Dire side and radiant side give the same amount of chance to win, and you have an infinite sample of dota 2 games where you must predict the winner. Let's also suppose that you have no others information on players in each team. Now explain me how you will predict the winning team below 50% ?
I hope that this exercice will enlight you.
I think you must redo all your work.
@Fear
Nice.
How is the system choosing whose DSR to update first? Could you tell me my new score?
The update count has been stuck at 20.119 for a long time.
@Fearohnum // MyDotaSkill.com
Is there any way to refresh your score? I lost 400 rating when you updated something and since then the quality of my games have been significantly better, I was curious as to how that reflects on my rating. Its been stuck on the same number for a while now, do they ever get refreshed?
@ Revert death
If you can't see the difference between 50% (random) and 65% prediction rate, maybe you should redo your bachelors?
@Lachesis
Your current score is: 1335
WATTDAFAK
you just made my night
@Fearohnum // MyDotaSkill.com
It's crazy that you are so bad at statistics, mathematics and probability and still use some of their tools.
Tell me how you can achieve a prediction rate below 50%.
You wrote "Elo can predict around 50% of wins." So tell me how you can predict below 50% in Dota 2 ? Are you talking about Elo in chess ? Which is completly differents as in chess you have high chance to do a draw.
PS/ You never replied to others comments I made before.
@Libertarian
Your current score is: 565
Hey I cant view my score. Says my data couldn't be fetched.
@Divinity.Darkfire
Your score is: 1627
@revert death
I've never claimed prediction rate below 50%. Since all you do is take things out of context, this is my last response to you. People who actually want to improve this thing ( Like the PhD from the beginning of the thread ) give useful input. All you do is write about some unimportant bullshit. Is this because you've got a score of 400? It isn't really that bad. You can always improve.
Could you calculate my score?
@Valar Morghulis
Your DSR is: 1288
Could you calculate my score?
@ Fearohnum // MyDotaSkill.com
I could not care less about my score except if I would have made this system and I was in the bot 90%. The idea to make a dota scoring is quite ambitious to say the least anyway.
Your whole gibberish started when I said here than a 65% prediction rate isn't that crazy when it is just a bit better than the worst case scenario when you predict 50% of the times the winning team. Then you started to reply agressively with insulting posts and wrong facts. I think I proved my points several times and tried to help you to correct several of your lacks of understanding.
I never wrote unimportant bullshit but tried to point some errors and you took it personnally, the more you wrote the more you showed some of these lacks of understanding.
It's ok to do some mistake as long as you correct them later, now if you think that pointing mistake isn't an useful input then keep your system as wrong as now but don't except anyone to take it seriously but naifs.
You should completly redo it.
Good luck,
Cheers.
Could you calculate mine please? Thanks.
@Fearohnum // MyDotaSkill.com
my score updated once after one game: dropping from 1411 to 1350, then stay at 1350 after another 20 games, so how often does it refresh? would u mind telling me my current DSR? thx
@Forza Roma
To my best knowledge, your score didn't update after that one game. The algorithm had changed, so your score changed along with it. It seems impossible for the scores to update on their own. This is the only time the scores have been updated, its being done manually, slowly.
@a.joller
DSR: 649
@King Leonaidas
DSR: 1088
@Forza Roma
DSR: 1139
@Revert Death,
I think you need to learn Statistics. 50% prediction is not the worst case, it is the default case. If you can predict more than 50% then it automatically means that the matchmaking system isn't fair/working as it should.
@creator,
I would rather use hero damage instead of KA, as Kills and Assists can pretty easily be faked whereas hero damage can not.
@outworldzilla
Fair point. Generally I will extract any variables I can.
Has a new algorithm been implemented? My DSR went from 1527 to 445, which is a ridiculously huge change.
Could you kindly calculate my score?
--> http://mydotaskill.com/index.php?id=http%3A%2F%2Fdotabuff.com%2Fplayers%2F78410961
http://i.imgur.com/HPBoork.png
@Outworldzilla
Did you read my previous post ? The 50% numbers they gave is only a theoritical numbers (only one of their criteria), anyway you can play better if you're worry to have bad odds to win a game. Can you define "default case" in statistics, I mean a mathematic definition.
Heroes dammage can be easily faked, just take zeus.
@BE THE LEAF
Don't worry about his bad algorithme several players have proved that his system don't works, most of the "computer science" guys that try to do statistics always do the same shit anyway, they don't understand what they do.
His system is just very bad right now. Basically it just take any number out of his arse that somehow rate some of the pro players above the average players.
@Revert Death
You've posted so much in this thread that I'm starting to think you have an infatuation with the creator. And I see you are a user of the DSR website.. You are what normal people refer to as a "leech". You're also probably muted in DotA 2, as there is NO ONE who would put up with a feminine, whiny attitude such as yours.
The DSR website is what pushed valve to implement RMM, yet RevertDeath contributes nothing, he uses the DSR website, and comes on here to spew his garbage that nobody wants to hear.
revert death is just full of crap
this shit works and wp
Just to let you guys know - the first refresh is under way ( you can see the progress on the main page ) - In the mean time, I am creating an updated algorithm - it will look at 50 variables to determine the bracket that you play in.
So this will be basically an improved version of the bracket estimator that's online now. I think it will be the final version of bracket estimation based DSR calculation.
I will then "merge" or "combine" this bracket estimator with individual hero performance analysis ( the first version of the algorithm that I've shown on Reddit / that was first online ) to account for how well you play any given hero ( in relation to your bracket ).
Best regards
@Libertarian
How funny , yes I used it once, how would have been aware that it exist otherwise ? And I didn't know the author before this thread, sorry for you sherlock. I didn't even talked to him before he started to say nonsense.
Your just delusional what pushed valve to implement RMM is the two years complaining from players.
And sorry for you but i'm not muted in Dota 2, I won't talk about ur avatar and your top 3 and end this conversation here; if you want to say nonsense here just do it in private likethat I won't reply to you and we won't pollute this thread.
Could you calculate my score ? gaven't played in a while... Thanks
Could ya check mine please.
Fear, i'm trying to check for a buddy and neither of us can get his to come up. His dotabuff: http://dotabuff.com/players/128636005
Mine is showing up fine, but his is not.
Come on, this guy has more than me:http://mydotaskill.com/index.php?id=http%3A%2F%2Fdotabuff.com%2Fplayers%2F141217222
Fact is that it doesn't make sense sometimes: My rating drops from 900 to 400, the reasons he gave me reached from I played in a 5 man stack when I obviously did not for 45 games (deleted that part of the comment later) to "the players you played with had low winrates (48%)" and low KDAs. I hope I don't have to lead the discussion again about whether winrate matters a hole lot or not... KDA might matter a bit more.
He also spoke of a kinda low teammatchmaking-rating, although I never ever played a single game in team-matchmaking. I don't know what he sees in his data but he either seems to be unable to read it or just lies.
As I said in Fadis thread a few days ago however, he has a 72% winrate with high KDA and is at 6xx points. He also apparently played against pro players. Low priority dodging cannot answer this rating at all imo, especially when he points out low winrates as reason for my rating.
How it comes he says I stacked 5 man when I obviously solo queued? Then delete this part later.
How it comes the owner himself had a stupid 5500 or something like this rating and then suddenly it dropped to 600 on his own website? And hey, now it reached 260, hard times for him.
How Fadi and his friends are mostly kinda low rated when they play vs. very good players?
How random smiley face guys other account is pro (1100) now (btw rising from a horrible rating) with apparently worse farming abilities than me (400) as he said one page before? http://dotabuff.com/players/134235086
Does it mean matchmaking wants to make me carry every feeder and noob team atm because it feels like it? ...
Believe what you want, I don't give a lot on this DSR-thing at the moment and I even said that before my rating dropped from 9xx to 4xx http://dotabuff.com/topics/2013-12-07-my-account-got-disabled-from-ranking--admin-please-read
And on what the owner says not more. He can predict 65% which team wins? I've seen not any kind of evidence so far. For some reason he doesn't seem to want to accept that he has to work on it quite a bit, instead points out things that show nothing or are untrue.
@Captain, you've obviously made very little effort. Fadi and his band of russian virgins play in LPQ, a lot of their matches are against extremely low skilled players, and occasionally they queue against another bunch of russian nerds stacking in LPQ without dodging..
Looking at your matches, there are a whole lot of anonymous accounts, much more so than you would expect from someone with a higher rating, and for those that I can see, none are particularly impressive..you might be performing well, but its in pretty low level matches (High bracket), which while decent compared to the whole player base, is actually a lot lower than what other people are playing at.
It seems that you are just another butthurt scrub that over estimated their own ability by using win rate to determine their skill level, which pretty much everyone knows can be incredibly deceiving, and is now desperately flailing around trying to make sense of the situation.
Look at the other guy called community toxin, his win rate/kda are all impressive, but if you took a look at his matches, you would realise he is playing in quite low level matches, and hence he got a lower rating.
What you do in matches is one thing, who its against is another. If Messi scores twice against Bayern, and Aguero slots 2 past Accrington with an identical squad and performance, the only difference being the opposition, should they both get the same rating?
168 :'('''' Need more Wisp/Ursa.
Oh, and, uhh, cos it seems to be the fashion on this thread: UR SYSTM SUKS GO DIE I BETTA THAN 168 I MAD FOR NO REASON UR DISTRIBUTION NOT GAUSSIAN (even though, looking at it, it clearly is, just missing the left tail...) 2 MUCH MISS DATA DUN DU IT UNLESS 100% AKURATE SYSTM
Accrington Stanley....who are they?
@Captain
The only reason my ranking was so low on that account, was because i stacked with lower leveled players (irl friends) while i was in lpq. You think you are a better farmer than me? Go dream somewhere else... And btw, I get matched up with players like this guy http://dotabuff.com/players/88282736 while solo queuing, so you're probably not in a position to question my skill Level.
does DSR update mine is rly high http://mydotaskill.com/index.php?id=http%3A%2F%2Fdotabuff.com%2Fplayers%2F87107328 1721
lpq isnt good for your rating guys
(-963)
xO
^LPQ players too good
@Fearohnum // MyDotaSkill.com
Can I suggest adding an FAQ to the page? I forgot what order the colours were and their distribution ... and I'm sure I'm not alone :p
^ Agreed
mine didnt change , ive played over 20 matches and my score is still the same