This rating still seems to be pretty random as far as I can see, I've read the 10 calibration matches are just to influence the score because it also takes into account your background history but still I cannot understand exactly what's taken into account when scoring in a game.
If anyone can shed some light? For example I am ranked a bit lower than some friends with lower winrates and lower average KDAs as well. If so I imagine GPM are XPM are highly important variables but even so going from game to game the spikes in rating according to that game's performance don't make much sense.
I wasn't expecting much of a rating. I won 7 of my 10 ranking games. Ended up with 2085. I'm wondering if playing support 95% of the time has an influence over the MMR. In any case, it doesn't make any sense to me after my next game, where I played well and had very good stats for a Jakiro in comparison to everyone else in the game. But my ranking went down 25 points.
I would say that its probably best to reserve judgement of the ranking system until at least 100 solo ranked games re played and even then, that might not be nearly enough of a sample for your ranking to be accurate. But as far as I'm concerned my ranking should have improved after that game.
got enough of it... just had to ruin the game for those motherfuckers (actually just made it harder by destroying my items... as they didnt understand though i ended up destroying them 3 times :p).
Mine is 3200 however I think as more games are in the books the MMR will adjust accordingly because there's no way some of the players I'm paired up with/against are the same MMR according to Valves scale. Theres just too many players that are unranked still. I am not quite sure how the old non viewable MMR is factored into Ranked Matchmaking before the 10 placement matches are completed. I guess time will only tell :D
i have 11 wins and only 4 losses and i have 3368 un solo, why? i cant understand
i see people with more losses than win and has more than 4k, how this rank works? :(
i agree i went 9wins 1 loss (last game was stupid close D: lol) in ranked solo and i have a rating of 3388, i am assuming that it uses your public MMR when you start ranked
Got a rating of 2736 after 10 Games, won 5 . My last game was pretty cool, captained my team to victory ( http://dotabuff.com/matches/428463854 ) :) I think the algorythm seems to work.
That is an assumption. We don't have numbers for ranked yet, thus we don't know if the numbers will be higher or not. We also don't know if ranked mmr will slowly decrease over time for all players.
3181 Solo with a record of 10-5, 3215 group with a record of 6-4... the thing about MMR I've noticed you only get slightly better players... and a whole lotta prima donnas who make selfish picks (in AP for solo queues... I hate AP but my CD/CM queues were taking like 12 minutes).
Still seeing people who don't know how to use the courier, understand lane configuration, last hits, tps, dust, etc... so it's hard to believe 3200 is where you're better than 90% of the dota players. Having access to dotabuff gives some insight into how the teams are constructed. It almost always seems you have one "scrub" guy (46-48% win rate), an average joe, and a couple of decent players.
Per the stats... IDK, I've had some pretty awesome doom games (700-800 XP/Gold per min, top number of last hits, 27:1 kda) and only seen my stats go up 25-35 per game. I can't imagine they would matter too much or great support players would end up suffering in the MMR.
I'm guessing if I stick to characters I know (in the 60-70% win range) I'll end up much higher in MMR. My own short comings on a ton of characters have held me back from great overall win %'s.. but it makes people better players imo to play more rolls.
Anyways... I can't imagine how good you must be to have 5XXX. It seems unattainable to me ;)
Can someone link me to a good reference for how they calculate MMR?
When I read about stats as 'percentiles' normally, the lower the percentage, the better/higher a person is? For example, someone with a very high IQ would be in a lower percentile, because only a small number of the population represent that. So when I read :-
This would indicate that a lower MMR is better? Does anybody have any solid information on this, as I find it a bit confusing, even after reading the valve post.
OK cheers for that. It's a bit of a muddy explanation from Valve, but I suppose it would make sense as the more experienced players tend to have a higher MMR now that I've looked at some DOTA profiles. Cheers :)
1000
9 wins 1 lose = 3600 rating gg wp volvo
yesterday 5550 today 5450) 3 games vs 5 full man stack and stupid LC pikers
brood picker calling someone "stupid lc pickers" and complaining about mmr LMAFOAOFAOFAO
Wasn't this thread supposed to be about "POST YOUR SOLO RANKED MATCH RATING"?
9 wins, 1 loss, 3175. Am I a scrub?
4350, though it keeps dropping because my teammates don't want to work as a team. :(
Necro, yes :D
@Quad
no it keeps dropping because they overestimated how good you might be :P
This rating still seems to be pretty random as far as I can see, I've read the 10 calibration matches are just to influence the score because it also takes into account your background history but still I cannot understand exactly what's taken into account when scoring in a game.
If anyone can shed some light? For example I am ranked a bit lower than some friends with lower winrates and lower average KDAs as well. If so I imagine GPM are XPM are highly important variables but even so going from game to game the spikes in rating according to that game's performance don't make much sense.
5082 solo
@ nihihi#AR
it means that they are matched against better opponents
I wasn't expecting much of a rating. I won 7 of my 10 ranking games. Ended up with 2085. I'm wondering if playing support 95% of the time has an influence over the MMR. In any case, it doesn't make any sense to me after my next game, where I played well and had very good stats for a Jakiro in comparison to everyone else in the game. But my ranking went down 25 points.
http://dotabuff.com/matches/428329562
I would say that its probably best to reserve judgement of the ranking system until at least 100 solo ranked games re played and even then, that might not be nearly enough of a sample for your ranking to be accurate. But as far as I'm concerned my ranking should have improved after that game.
wins: 153
losses: 168
xp: 748
level: 19
initial_skill: 0
leaver_count: 5
low_priority_until_date: 1381088018
prevent_text_chat_until_date: 0
prevent_voice_until_date: 0
last_abandoned_game_date: 1386132578
leaver_penalty_count: 3
completed_game_streak: 52
teaching: 1
leadership: 5
friendly: 7
forgiving: 4
successful_report_count: 0
notify_successful_report: 0
account_disabled_until_date: 0
account_disabled_count: 0
showcase_hero_id: 23
match_disabled_until_date: 1382066263
match_disabled_count: 2
partner_account_type: PARTNER_NONE
low_priority_games_remaining: 0
calibration_games_remaining: 9
solo_competitive_rank: 2100
solo_calibration_games_remaining: 0
recruitment_level: 0
secondary_leaver_count: 0
last_secondary_abandoned_game_date: 0
why would you score go up when you played average on a game that you lost?
3004 solo trench tier reporting in.
@karmababy
desolator venge?
desolator crit wd?
linkens scepter gyro?
LMFAO
scrub game, scrub players
Yeah, scrub game I know, but why would my score go down in a game where I play better than everyone else?
because you lost.......
got enough of it... just had to ruin the game for those motherfuckers (actually just made it harder by destroying my items... as they didnt understand though i ended up destroying them 3 times :p).
REMEMBER MY NICK FUCKERS
DONT FUCKING FUCK WITH ME
Wow! People actually bragging about ruining games?
Mine is 3200 however I think as more games are in the books the MMR will adjust accordingly because there's no way some of the players I'm paired up with/against are the same MMR according to Valves scale. Theres just too many players that are unranked still. I am not quite sure how the old non viewable MMR is factored into Ranked Matchmaking before the 10 placement matches are completed. I guess time will only tell :D
3800, 7 - 3 (win loss) meh
but i dont get it, do the ratings that valve put
5% 1100
10% 1500
25% 2000
50% 2250
75% 2731
90% 3200
95% 3900
99% 4100
still count? because many ppl (atleast so ive seen) have 5k+ rating
the avg is higher here because the people with higher scores post them while the people who have horrible scores dont......
i have 11 wins and only 4 losses and i have 3368 un solo, why? i cant understand
i see people with more losses than win and has more than 4k, how this rank works? :(
does exist a website to see this ranking?
i agree i went 9wins 1 loss (last game was stupid close D: lol) in ranked solo and i have a rating of 3388, i am assuming that it uses your public MMR when you start ranked
And continuing from Sampson's point, some people exaggerate their scores (a certain notorious forum troll comes to mind).
Not only that but those scores that you are seeing brain are from pub games, ranked will be slightly higher.
Got a rating of 2736 after 10 Games, won 5 . My last game was pretty cool, captained my team to victory ( http://dotabuff.com/matches/428463854 ) :) I think the algorythm seems to work.
rly dont see how you get points after losing, you can get as much 1 million gold advantage and lose and you wont get points.
you can have a high rank with low winrate if the matches you won were played vs well ranked people
I don't get how people get 4000+ with less than 40% winrate tho
I just finished 10 games on solo ranked, I'm 9/2 with 3900 mmr
edit: 8/2 not 9/2
A Hobbit well that is just false.
4651
Seems to be going up, I'm keeping an Excel file of it. Last two matches netted a +100.
6-0 wonder what my actual rmmr is tho.
3.248 4 win 6 losses
does anyone know how this rank is calculated? im not sure i remember it being posted anywhere
@Sampson:
That is an assumption. We don't have numbers for ranked yet, thus we don't know if the numbers will be higher or not. We also don't know if ranked mmr will slowly decrease over time for all players.
3181 Solo with a record of 10-5, 3215 group with a record of 6-4... the thing about MMR I've noticed you only get slightly better players... and a whole lotta prima donnas who make selfish picks (in AP for solo queues... I hate AP but my CD/CM queues were taking like 12 minutes).
Still seeing people who don't know how to use the courier, understand lane configuration, last hits, tps, dust, etc... so it's hard to believe 3200 is where you're better than 90% of the dota players. Having access to dotabuff gives some insight into how the teams are constructed. It almost always seems you have one "scrub" guy (46-48% win rate), an average joe, and a couple of decent players.
Per the stats... IDK, I've had some pretty awesome doom games (700-800 XP/Gold per min, top number of last hits, 27:1 kda) and only seen my stats go up 25-35 per game. I can't imagine they would matter too much or great support players would end up suffering in the MMR.
I'm guessing if I stick to characters I know (in the 60-70% win range) I'll end up much higher in MMR. My own short comings on a ton of characters have held me back from great overall win %'s.. but it makes people better players imo to play more rolls.
Anyways... I can't imagine how good you must be to have 5XXX. It seems unattainable to me ;)
Can someone link me to a good reference for how they calculate MMR?
When I read about stats as 'percentiles' normally, the lower the percentage, the better/higher a person is? For example, someone with a very high IQ would be in a lower percentile, because only a small number of the population represent that. So when I read :-
5% 1100
10% 1500
25% 2000
50% 2250
75% 2731
90% 3200
95% 3900
99% 4100
This would indicate that a lower MMR is better? Does anybody have any solid information on this, as I find it a bit confusing, even after reading the valve post.
first 10 games (4w-6L) i got 3414 then after a few more games i now have 3519
12 ranked matches with 7 wins 5 loses
rank is 3342. is that bad?
seems about right, mine is 3623. I want that 4000
@bad spleen
No it does not indicate lower is better. When it says 4100 is 99%, it means players with that ranking are higher than 99% of other players.
@Concede
OK cheers for that. It's a bit of a muddy explanation from Valve, but I suppose it would make sense as the more experienced players tend to have a higher MMR now that I've looked at some DOTA profiles. Cheers :)
im 2986..... i dont mind this number right now coz i most of time i lag out ..... so far i left 80+ matches ... i cant help with my network
can some1 tell me what my mmr is lol or a ballpark number based on what im being matched with
4k
Playing Earth Spirit can cause all sorts of wonders to one's winrate and rank.
3400 not bad but this is definitely going to go up