Tnx Relentless. Chris have low dbr cos 2 of his last 20 matches was `cm` mode. Atm site have bugs with other mods to calculate well. Just ap mode works well for now,but will fix soon. Also he has been stacked with some lower players than him alot in last 20 matches thats main reason why his dsr atm is not accurate and total unreal. Atm DSR system calculate your 20 last games how you played and with who and vs who you played ofc. So is more like average skill lvl of your last 20 games. If you stack with better ppl mmr than yourself your dsr will grow also if you stack with lower dsr will drop. I stacked with noob friends 30 games ago etc. And last 20 just solo queue and my dsr went from 1200 to 1500+. Also i tryharded my best heroes etc.
If DSR is so volitile it will be interesting to see how things change over time. But these players have a strong stack nearly every game and are really very good, nearly pro level players... yet they have "low" DSR, if you call top 1% low.
I just watched most of this group beat Virtus Pro in a pub game this morning. It was close, VP made some critical errors... but still they are good enough to make VP lose a pub game. Here are some example DSR however.
These are extremely talented dota players, most significantly more skilled than me. But they all have far lower DSR. They do stack very heavily, while I do only a little stacking with plenty of solo que games as well. Perhaps stacking many games leads to a lower DSR score. It is interesting that Bogi noticed his DSR going up so much while in solo que. Maybe solo que games are worth a lot more points.
You realise that I don't play on this account right?
As for those players, they are all stackers that play in LPQ...look at their most recent match: http://dotabuff.com/matches/409879008 That is why they have such low ratings...they are good at the game, but the DSR system is punishing them appropriately for playing in lower level matches (not consistently, but in some, I know there are other virgin tryhards that smurf stack and dodge everyone in LPQ)
Relentless hm..a bit weird yes and not accurate but main reason is I think they stack and face lower players and not strong stacks like they are so they are getting easy games cos they are better alot. They getting easy enemies for their skill lvl so they are not getting too much dsr for wins seems like. I can explain you now on chat 10-15 min. now if you want conciliation between each other. I made some experiments with my last 20 games solo queue and wanna share with you,also I knew a bit more how my dsr was going.
Player number 117998063 DSR is: 1022
This player has a Pro DSR score (top 0,5%).
Hmm...something's fishy about this. I certainly don't consider myself a top 0.5% player, specially considering that I don't use any single high-skill cap hero frequently, except arguably Rubick, wether you consider him a high-skill cap hero.
LK Zano: which hero you play doesn't matter; how well you do with that hero is what matters. Also are you saying that 4/5 support professional players are low skill then? Puppey, Aui, Fluff are very skilled but often play heroes like Lich...
jesus christ cunts, top 1% and top 0.5% is piss easy to achieve (not calling you guys bad, just read the rest of this post), this is because there are so many bad players, and I understand that its difficult to accept that there is a large percentage of the player base that suck so much.
You can see this on dotabuff too if you want, click the matches tab, and open some random matches. 3-5 carries on a single team? Everywhere. Retarded item builds, CHECK. Ridiculously low CS (without any indication of them doing much else like ganking/pushing, etc, and I'm talking more about carries here obviously) check.
It was the exact same story with DBR which some people have so easily forgotten, an 1800 rating (remember lowest was ~1000, top was ~1700) put you in the top 1% of the player base, so no, you being put in top 0.5 or 1% isn't a sign of an inaccurate system, its a sign that you aren't actually aware of the skill level that the majority of the dota 2 player base possesses.
According to their Dbuffs Puppey plays very effective LD, Tinker, Storm Spirit and Morph.
Aui plays a lot of micro intensive heroes
Fluff too, plus a bunch a of skillshot heroes.
:P I'm just critizicing myself a bit, I'm not a patient enough man to practice those heroes, I prefer simple heroes which require more balls than skills to play effectively :D
^ I guess he's furious because a lot of people don't seem to be reading what he posts.
True that I m just checking some games,at matches as you said. There are some weavers with mana boots and Void`s with shadow blades so yes..alot of noobs too many. Alot of ppl top 1% here and at dsr cos usually ppl who play good and well care about their stats and sites/forums like this etc. 90% players play dota few games a week etc. for weekends or on vacations. So they even don`t know what dotabuff is or any other site for dota stats etc.
you have/had haters because the kind of shit that you said on PD was so stupid, that people couldn't figure out whether or not you were a troll or a retard, not because you are/were bad or good at the game.
^
496 is low for this forum and probably average for the total game population.
That said, there has been no response from mydotaskill on the issue of players not making their accounts public and if it affects the ratings. The rating system is based on the players you play with. If everyone you play with doesn't share their info, i'm not sure if it is able to accurately determine a skill level.
@親和性 ' FameGG.com
I got banned because I said that you get reported if you ignore your entire team.
@Concede
The system uses machine learning techniques and rates your performance with your heroes one by one compared to the average (I guess that also includes which heroes you win against, else I could not explain my high DSR).
Well, lot's of people want to make a drama about me because they obviously need that hated person, and dota players in general dislike the kinds of me who play the same hero over and over again.
After a very negative prejudge against the system, ( i was very -mad- for being silver to bronze, bronze to copper in one day)
Today i realized it is working accurate actually.
Creator -i guess- of the system, here; told me that problem is with my mmr is my stacking friends (they were lp somehow) and he gave me an advice to play solo or better skilled friends.
I decided to play alone first, and won a game by playing mid, destroying enemy mid hero and doing great ganks etc.
Shortly after, i played a game with a skilled friend who is already in 0,5 -1270 DSR. (he is cm here= http://dotabuff.com/matches/409727609)
Finally, im back to my Silver DSR which i believe i normally belong to, and sometimes in gold maybe - if i can win stack -
Sooo, i cant really say the calculating system %100 true about your skill level but in two matches, i earned 200 points thus i am able to say, it sees the difference between games. (i could be happier with gold dsr : P but come on 2 matches only)
Considering VALVE is not providing us any Ladder, MMR table or such things, we should thank the guy for this work.
Good job bro! You really need to give us your name as a creator, if you are not going to close the site in 1 month after seeing flares, mads and bad comments.
Does anyone know how many games Valve takes into account for the actual MMR? Do they only take into account for recent games or the all games?
Player number 42432107 DSR is: 427 This player has a Gold DSR score (top 15%).
Its probably alot higher than valve's rating, my last 30 or so games were solo queue, no parties but I never get into high bracket (which is the top 20%?) when solo queuing. If Valve accounts for all matches that might be why, since my first ever dota/moba games are all played on this account and my several hundred games are plain awful
i think this percentage system only consider players that are visible for dotabuff. And tons of bad players don't turn their privacy off (mostly because they don't know what is it)
Strongest belief about Valve's bracket trick is "win stacks" ... Any "normal skilled "solo / or not" players can get higher bracket if he/they win over and over again.
I was in the high bracket and then "First Blood" update changed a lot of things and i made a huge mistake by playing with some starters, im not able to join the mighty brackets but i can still feel their absence because whenever i start playing some good, my enemies are getting better while my team mates are somehow acting like high! So i say, yeassjjj doto, i know your plan, i got over it, if i can survive, ill get my bracket back! :)
unsure about the mechanism
checked one game before, it was 1411
then after winning last game, it became 1350
btw, in last game players I queued with mostly have 1000+ wins
@raziel, neither valve nor DSR give a shit about what you did a few hundred matches ago, which is why this system is producing numbers that are closer to your actual skill level, where an ELO system would fail. Realise that the majority of the player base is horrible, and that a reading of top 15% is exactly what you should expect from such a rating. I'll take away 1000 points from DBR so the example is crystal clear, 800+ DBR = top 1%, while the highest players had 2700. 0-400 accounts for 60-80% of the player base.
@forza, win rate or games played =/= skill, there is some correlation between the two, ok, but you will find that there are lots of players that have nice win rates that have very low ratings because its from playing in low level matches. Now obviously your rating is not low, but its definitely possible for it to drop based on the average skill level of the game, win or lose. DSR is in a test phase, the first heavily focused on individual performance, the current one focuses on skill level of a match. When the two are combined, you will see more accurate changes.
edit: and yep, out of the people you can see that have been processed; 834, 1279
@outworldzilla, Do you really want me to dig up some stuff dude? Like legit, if you want more people to laugh at you, I'm more than happy to go for it. There was a reason you were the laughing stock of PD. If you think you were right, and that people on here will agree with some of the stuff that you sad, then its really quite sad, and highlights just how stubborn you are.
@terrible, so in my case what would account for the "over rating" since going by Valve's MM, i'm in normal bracket, the 60-80%. Is it known how close the 2 ratings are and what are the differences between them?
@loverman, if win stacks is indeed the key, that might explain it. But i dont seem get better opponents when solo queueing, just worse team mates. Only time i seem get better opponents is when stacking with better players, and there its quite clear i'm out of my league
@Terrible
go ahead, you won't find anything anyway :)
Because there is nothing. Except you try to use my replies to people raging at me maybe.
The reason you hate me is obviously because when someone told me that I am bad, I replied with "no, I'm not"?
All forum threads stay on topic by positive comments bringing them back to the topic and ignoring the distractions of trolls and personal fights that develop.
So far my score is still 1468, either my game had no impact or it just doesn't update that frequently. Its been 2 hours since the game ended now.
@Forza Roma
Haha, dream on. Well, maybe that's true. I had 1700 games with Phoenix in DotA 1.
But I really am not proud of having many games with one hero. I envy the other players who can play all different heroes so well. I could if I would learn it, but I'm just not having fun playing most other heroes :(
So I stick with what's most fun. :)
btw my skill didn't change at all after my last 2 games ... :<
Tnx Relentless. Chris have low dbr cos 2 of his last 20 matches was `cm` mode. Atm site have bugs with other mods to calculate well. Just ap mode works well for now,but will fix soon. Also he has been stacked with some lower players than him alot in last 20 matches thats main reason why his dsr atm is not accurate and total unreal. Atm DSR system calculate your 20 last games how you played and with who and vs who you played ofc. So is more like average skill lvl of your last 20 games. If you stack with better ppl mmr than yourself your dsr will grow also if you stack with lower dsr will drop. I stacked with noob friends 30 games ago etc. And last 20 just solo queue and my dsr went from 1200 to 1500+. Also i tryharded my best heroes etc.
If DSR is so volitile it will be interesting to see how things change over time. But these players have a strong stack nearly every game and are really very good, nearly pro level players... yet they have "low" DSR, if you call top 1% low.
I just watched most of this group beat Virtus Pro in a pub game this morning. It was close, VP made some critical errors... but still they are good enough to make VP lose a pub game. Here are some example DSR however.
Vroksnak 816 DSR http://dotabuff.com/players/86738494
Fadi 665 DSR http://dotabuff.com/players/146122532
Baitz 556 DSR http://dotabuff.com/players/32878348
Chris 950 DSR http://dotabuff.com/players/59625555
Vaikiss 1167 DSR http://dotabuff.com/players/86777307
Van Art 1149 DSR http://dotabuff.com/players/68180198
These are extremely talented dota players, most significantly more skilled than me. But they all have far lower DSR. They do stack very heavily, while I do only a little stacking with plenty of solo que games as well. Perhaps stacking many games leads to a lower DSR score. It is interesting that Bogi noticed his DSR going up so much while in solo que. Maybe solo que games are worth a lot more points.
1525 omg, feel so pro lol
it should consider more than 20 games... players that are on the same level as me or even quite above are underrated
You realise that I don't play on this account right?
As for those players, they are all stackers that play in LPQ...look at their most recent match: http://dotabuff.com/matches/409879008
That is why they have such low ratings...they are good at the game, but the DSR system is punishing them appropriately for playing in lower level matches (not consistently, but in some, I know there are other virgin tryhards that smurf stack and dodge everyone in LPQ)
Relentless hm..a bit weird yes and not accurate but main reason is I think they stack and face lower players and not strong stacks like they are so they are getting easy games cos they are better alot. They getting easy enemies for their skill lvl so they are not getting too much dsr for wins seems like. I can explain you now on chat 10-15 min. now if you want conciliation between each other. I made some experiments with my last 20 games solo queue and wanna share with you,also I knew a bit more how my dsr was going.
Terrible is right. Also I think there is some stack protection,so tryhard stacks wont get big dsr.
Can you check my Friend's accounts?
http://dotabuff.com/players/86728079
http://dotabuff.com/players/3436704
could u check me and my friend's? thanks in advance!
http://dotabuff.com/players/104210326
http://dotabuff.com/players/93760207
i dropped by 700 points since the start of this thread
so sad ;_;
Player number 117998063 DSR is: 1022
This player has a Pro DSR score (top 0,5%).
Hmm...something's fishy about this. I certainly don't consider myself a top 0.5% player, specially considering that I don't use any single high-skill cap hero frequently, except arguably Rubick, wether you consider him a high-skill cap hero.
LK Zano: which hero you play doesn't matter; how well you do with that hero is what matters. Also are you saying that 4/5 support professional players are low skill then? Puppey, Aui, Fluff are very skilled but often play heroes like Lich...
jesus christ cunts, top 1% and top 0.5% is piss easy to achieve (not calling you guys bad, just read the rest of this post), this is because there are so many bad players, and I understand that its difficult to accept that there is a large percentage of the player base that suck so much.
You can see this on dotabuff too if you want, click the matches tab, and open some random matches. 3-5 carries on a single team? Everywhere. Retarded item builds, CHECK. Ridiculously low CS (without any indication of them doing much else like ganking/pushing, etc, and I'm talking more about carries here obviously) check.
It was the exact same story with DBR which some people have so easily forgotten, an 1800 rating (remember lowest was ~1000, top was ~1700) put you in the top 1% of the player base, so no, you being put in top 0.5 or 1% isn't a sign of an inaccurate system, its a sign that you aren't actually aware of the skill level that the majority of the dota 2 player base possesses.
Terrible sama, why so furious?
According to their Dbuffs Puppey plays very effective LD, Tinker, Storm Spirit and Morph.
Aui plays a lot of micro intensive heroes
Fluff too, plus a bunch a of skillshot heroes.
:P I'm just critizicing myself a bit, I'm not a patient enough man to practice those heroes, I prefer simple heroes which require more balls than skills to play effectively :D
^ I guess he's furious because a lot of people don't seem to be reading what he posts.
True that I m just checking some games,at matches as you said. There are some weavers with mana boots and Void`s with shadow blades so yes..alot of noobs too many. Alot of ppl top 1% here and at dsr cos usually ppl who play good and well care about their stats and sites/forums like this etc. 90% players play dota few games a week etc. for weekends or on vacations. So they even don`t know what dotabuff is or any other site for dota stats etc.
http://mydotaskill.com/index.php?id=86717663
Haha, mithril rating, what the hell...
Still I think it won't change anything for all haters who want to see me at 600 points ...
^
Looks accurate.
you have/had haters because the kind of shit that you said on PD was so stupid, that people couldn't figure out whether or not you were a troll or a retard, not because you are/were bad or good at the game.
Went from 950 ish to 1730 after update. Probably coz i solo que.
@Terrible
What did he do?
Tell more about drama! ^^
Finding individual skill with so many unknowns and for a team based game is bound to be flawed.
My DSR is 496.
That result base by the last 20 game ??? well, I win 14 of 20 last game, that score is low or high ?
^
496 is low for this forum and probably average for the total game population.
That said, there has been no response from mydotaskill on the issue of players not making their accounts public and if it affects the ratings. The rating system is based on the players you play with. If everyone you play with doesn't share their info, i'm not sure if it is able to accurately determine a skill level.
@親和性 ' FameGG.com
I got banned because I said that you get reported if you ignore your entire team.
@Concede
The system uses machine learning techniques and rates your performance with your heroes one by one compared to the average (I guess that also includes which heroes you win against, else I could not explain my high DSR).
Ah, and I was expecting more drama!
Well, lot's of people want to make a drama about me because they obviously need that hated person, and dota players in general dislike the kinds of me who play the same hero over and over again.
I played one random game of solo que. So far it has not changed my score. How often does it update?
well, I win 14 of 20 game and still low, then what the hell the top ranks doing in they match to get 1700 scores ??????
After a very negative prejudge against the system, ( i was very -mad- for being silver to bronze, bronze to copper in one day)
Today i realized it is working accurate actually.
Creator -i guess- of the system, here; told me that problem is with my mmr is my stacking friends (they were lp somehow) and he gave me an advice to play solo or better skilled friends.
I decided to play alone first, and won a game by playing mid, destroying enemy mid hero and doing great ganks etc.
Shortly after, i played a game with a skilled friend who is already in 0,5 -1270 DSR. (he is cm here= http://dotabuff.com/matches/409727609)
Finally, im back to my Silver DSR which i believe i normally belong to, and sometimes in gold maybe - if i can win stack -
Sooo, i cant really say the calculating system %100 true about your skill level but in two matches, i earned 200 points thus i am able to say, it sees the difference between games. (i could be happier with gold dsr : P but come on 2 matches only)
Considering VALVE is not providing us any Ladder, MMR table or such things, we should thank the guy for this work.
Good job bro! You really need to give us your name as a creator, if you are not going to close the site in 1 month after seeing flares, mads and bad comments.
Does anyone know how many games Valve takes into account for the actual MMR? Do they only take into account for recent games or the all games?
Player number 42432107 DSR is: 427 This player has a Gold DSR score (top 15%).
Its probably alot higher than valve's rating, my last 30 or so games were solo queue, no parties but I never get into high bracket (which is the top 20%?) when solo queuing. If Valve accounts for all matches that might be why, since my first ever dota/moba games are all played on this account and my several hundred games are plain awful
i think this percentage system only consider players that are visible for dotabuff. And tons of bad players don't turn their privacy off (mostly because they don't know what is it)
If Outworldzilla has such rating that indicates something wrong with the system.
Strongest belief about Valve's bracket trick is "win stacks" ... Any "normal skilled "solo / or not" players can get higher bracket if he/they win over and over again.
I was in the high bracket and then "First Blood" update changed a lot of things and i made a huge mistake by playing with some starters, im not able to join the mighty brackets but i can still feel their absence because whenever i start playing some good, my enemies are getting better while my team mates are somehow acting like high! So i say, yeassjjj doto, i know your plan, i got over it, if i can survive, ill get my bracket back! :)
Liked your name btw, Raziel.
unsure about the mechanism
checked one game before, it was 1411
then after winning last game, it became 1350
btw, in last game players I queued with mostly have 1000+ wins
@raziel, neither valve nor DSR give a shit about what you did a few hundred matches ago, which is why this system is producing numbers that are closer to your actual skill level, where an ELO system would fail. Realise that the majority of the player base is horrible, and that a reading of top 15% is exactly what you should expect from such a rating. I'll take away 1000 points from DBR so the example is crystal clear, 800+ DBR = top 1%, while the highest players had 2700. 0-400 accounts for 60-80% of the player base.
@forza, win rate or games played =/= skill, there is some correlation between the two, ok, but you will find that there are lots of players that have nice win rates that have very low ratings because its from playing in low level matches. Now obviously your rating is not low, but its definitely possible for it to drop based on the average skill level of the game, win or lose. DSR is in a test phase, the first heavily focused on individual performance, the current one focuses on skill level of a match. When the two are combined, you will see more accurate changes.
edit: and yep, out of the people you can see that have been processed; 834, 1279
@outworldzilla As i understand it, that is not the case at all.
@outworldzilla, Do you really want me to dig up some stuff dude? Like legit, if you want more people to laugh at you, I'm more than happy to go for it. There was a reason you were the laughing stock of PD. If you think you were right, and that people on here will agree with some of the stuff that you sad, then its really quite sad, and highlights just how stubborn you are.
wow how did i drop from 1800 to 1400 in two games :'(
Wow, did you even look at the page?
@terrible, so in my case what would account for the "over rating" since going by Valve's MM, i'm in normal bracket, the 60-80%. Is it known how close the 2 ratings are and what are the differences between them?
@loverman, if win stacks is indeed the key, that might explain it. But i dont seem get better opponents when solo queueing, just worse team mates. Only time i seem get better opponents is when stacking with better players, and there its quite clear i'm out of my league
@Terrible
go ahead, you won't find anything anyway :)
Because there is nothing. Except you try to use my replies to people raging at me maybe.
The reason you hate me is obviously because when someone told me that I am bad, I replied with "no, I'm not"?
Any chance this can stay on topic?
"As i understand it, that is not the case at all. "
Where did you read that? I was basing my statement on the link provided by the OP (in another thread as it seems :X).
Edit: Here http://www.reddit.com/r/DotA2/comments/1rhfs8/show_rdota2_i_am_developing_a_skill_measuring/
All forum threads stay on topic by positive comments bringing them back to the topic and ignoring the distractions of trolls and personal fights that develop.
So far my score is still 1468, either my game had no impact or it just doesn't update that frequently. Its been 2 hours since the game ended now.
@Outworldzilla
if counting games in dota1, I should have more games playing with clinkz than you playing with od
@Relentless
my score dropped from 1411 to 1350 after last game, in which I did fine
http://dotabuff.com/matches/410059080
@Forza Roma
Haha, dream on. Well, maybe that's true. I had 1700 games with Phoenix in DotA 1.
But I really am not proud of having many games with one hero. I envy the other players who can play all different heroes so well. I could if I would learn it, but I'm just not having fun playing most other heroes :(
So I stick with what's most fun. :)
btw my skill didn't change at all after my last 2 games ... :<
@terrible no i didnt ? why would i read 500random comments only wat i say matters to me :/
@Outworldzilla
dota1 phoenix picker haha, more hated than sb pickers in 6.78
But people in the forums and in the games were heavily underrating him (only played Phoenix between 6.71 and 6.74).
Was a good beginner hero though :)