2500 mmr guy talk about balances and unfair? Really? Cmon you are litle kid still at this game..you need years of playing to be decent.You are stucked there cos that`s your REAL mmr at the moment,stop cry and improve yourself tnx.
2.5k mmr means you are better than about 60% of dota 2 players. Now that might seem like a good number but it is not because there are a lot of new players. I'd say an average player has 3-3.5k mmr.
The best way to improve yourself is to watch your own replays and learn from the mistakes you made. I don't play a lot myself but I spend time watching competitive a lot and I'm sitting at 4.2k whereas my friends who have been playing for almost 1.5k hrs only have 3.2k.
I totally agree, I'm in the same situation,Today I lost four matches in a row because of disconnected idiots,probably rage quit,i try everything,playing support,carries,nukers,and drop from 2800 to 2500 because of feeding idiots,mostly Russians....
The situation is as follows, the first three picks are heavy carriers, of course,i let the idiots to chose first, they naturally don't look to counter picks, they chose what they like, and then you ask your self, playing support or just behave like others .... Anyway you lose ....
1. I still don't understand why would anyone want to play all pick in ranked?! Scratch that, I don't understand why would anyone want to play all pick even in non-ranked, except for practicing a specific hero...
2. Life is kinda unfair. You'll have to learn to deal with it.
3. I don't really have anything else to say, but having only 2 points seemed to not be enough to justify using this sort of format in the first place, so...
This system is made so with time ( probably a lot ) if you are a good player you WILL win more games against better opponentes ( it doesnt matter how you win those games, the only thing that matters is if you win, even if you have to 1 vs 5 ). Now this is probably true if we follow a statistic.
Remenber matchmaking is still the same.
Personally i dont think that the problem its matchmaking, i think the problem is how they give points to everyone, it doesnt matter if you got carried ( space created my ass guys, sometimes when you feed for 15º time you cant do this excuse anymore ) you will get the same amount of points, this is probably unfair but ..... how will they know who did better ? KDA ? it can be exploited, gpm? it can be exploited, amount of wards ? it can be exploited, almost every in-game indication can be exploited.
If they make a general ranting for some heroes people is going to play X hero like a carry and some will play him like a support.
In other words you cant rank people by gpm, kda, xpm, etc ....
So if that doesnt work the only option is to rank people by win or losses like a normal elo system, this is probably unfair, but there is no other way to know someone perfomance on a game unless valve fucking watch every game and they decide by themselves who did better. ( any other way can be exploited )
This system is probably more accurate with more games, so if you have a low amount of games it will be worst or less accurate.
Well if you dont want to change you account u can just pick a hero who you have a really good winrate, statistically you will win more and get more MMR, you just need to have pacience for this i guess. If you think about it, its kind of unfair, because i will win mmr for playing a hero that pubs dont know how to counter,i will win mmr because they are bad not because i am good. For example i have 70 % winrate with OD, that its not because i am a good player, that its because people doesnt know how to play against that hero.
I wish there was a way of looking every game and knowing who did good and who did bad, but there isnt, so this is the best we got.
If you think you are way better than all your team mates just go and create a new account and solo queue try hard, your initial MMR will be bigger after the 10 first game.
TL:DR pick what you are really good with, with time and pacience your mmr will increase. Or creat a new account.
Edit: Bogi there is probably people who complain when they are the worst player on they teams, but i am sure that some of the ones who complain are actually good players, the thing is that you move reallllllly slow when the anti-smurf system decides if you are or if you arent a smurf. ( unless you get a fucking 10 win streak or some shit, and you need a little bit of luck for this )
poorly balanced heroes and match maker that allows parties outside of 5man cm is bound to make shit happens day in day out ie look at this autist for example he reached 6k rating by playing slark every game http://dotabuff.com/players/49382246
^Yeah like i was saying if that guy played any other hero he probably would be at 5k right ?
I dont think there will never be a good invidual score for dota 2, i think they should focus on improving Team MMR, probably the only score that can be true.
play Captains Mode and Captains Draft. Lock nobrain teamfight combinations (Lich warlock Dark Seer QoP) and win. Thats pretty much how it works. Once you are in the 3.3k MMR region you will win 70% of the games if you win mid and or offlane so try to focus on those roles.
@Dorkly, i don't how to quote here, so excuse for that. But wouldn't a rank based on hero statistics be better?
Like, ofcourse there are variations, so do with the actual system. But there are heroes that are pretty straight forward. For example, I never seen an support Anti-Mage (well, there must be, but ofc it's not a relevant percentage), so you could compare KDA, GPM, Tower Dmg, Hero dmg, etc with the month statistics for that particular hero. Would that be more fair? Like I said, ofc there is variations, like Naga can be played as carry or support, so the KDA would be very different. But for most heroes (i guess) they are played in the way of the majority of players play it: Void is usually a carry (although I saw a guy who picked it as a support and did well), Crystal Maiden is usually a support, etc, etc.
So why not to compare Crystal Maiden player KDA, GPM, Dmg done with the hero overall? Lets say CM has an average of 2000 hero dmg per game. If a player gets like 300 hero dmg in a game, it's probably because he didn't play well, as CM is an agressive support. If the overall GPM of CM is 400 and a guy does it 900, probably he killed/assisted a lot (or farmed a lot, but if he spent the whole game farming, he wouldn't have much hero dmg, so he wouldnt have had played well). Wouldn't that be possible to do?
If the player is below the average rate, then he probably didnt play well, so if his team managed to win the game anyway, let's say he would get 20 points for the winning, but would loose 5 because he didnt play well. At the same time, if a guy did 50/2 but his team were all like 0/29, it's not fair that this guy loose the same amount of points that the guy who did 0/30. So we hould loose, lets say, 20 points for the loss, but would win 10 for playing extremely well. Wouldn't that be more fair?
I think your issue is you have 51% wr the game force the team to have 50% chance to win/lose, so what it mean let's say your a player with 60%win it mean you have high chance to win so they must put you with bad players at the same mmr than you, usually player with 48% but twice more game than you.
So because of that you won't improve ur mmr if you play solo. You could try to get 100 lose streak to get a bad % and then put back ur win rate back to 51% and u'll be like 4k mmr.
slow dont get me wrong, personally even with the posibility of exploitiing i think that comparing your perfomance with the overall perfomance with the hero is a good ( or at least aceptable ) way of telling your rank, but i think they did that on hon and it didnt work out great, people played really .... i mean really arrogant just to improve they kda, or people spent all the game farming just to have better gpm . At least that is what i have been told.
I agree with you about that not everyone should lose/win the same amout of points, but i think that would be dangerous because still could make a bad or arrogant behavior on a player that only wants to protect they stats.
Lol. You are rubbish so you get matched with rubbish players. It's as simple as that
Are you sure a system that counts anything but wins would be better? For example I like playing Venge, but with a system taking into account KDA, why would I EVER swap-suicide for my carry? To win the game? No waaay, Id got more points, if we lose but I use my stun for KS and swap someone else inside the teamfight to die instead of me. Also, why deny creeps when I can steel the lasthits of the carry instead? Stacking jungle? No way, Id just kill them myself, more GPM more points.
Also, everyone should lose rating, who plays rat doto? TI3 proved that dota is about destroying the ancient, not about any other numbers. Personally I dont like rat doto, BUT still better than KDA-doto without the goal to win.
TL;DR: my opinion: system is good as it is, you should try to win not increase any other number.
@Vaeld, stupid argument.
You literally have no idea what you are talking about. You are looking at individual factors in isolation, which is wrong. Just because a particular act improves one signal, doesn't mean it doesn't have an effect on others..
ie. Winning a match and doing suicide swaps is probably going to get your rating higher than if you didn't and lost
My point is, I think you cant really say who did better just by looking at stats. Just the most common case, gank, no tp from team, someone dies. How can you tell, whos fault it was? The one who died? And was yelling on voice for 5-10 seconds for a TP? Or the one who didnt tp? You can count tps used, but is tping into lane after death/regen is that good? I dont think any system could really see it, not only from stats. Using more factors and giving scores to decisions in-game by some trained AI, it could work.. But the game is already lagging a lot and it would take way too much work and computation...
Even if the system is good and hard to exploit, people would try... Im happy with the system right now. It will be better with time when everyone gets to the rank where they belong.
I agree with Vaeldiithia. Since I predominantly play support (CM/Veno), my overall stats can be improved by kill stealing[the irresistible urge with CM's Q], getting last hits in the lane instead of zoning out the opponents, choosing to buy items that improve my survivability rather than more wards, not dying for my carry late game. I dont do any of that (not always lol) because all I care is that my team wins.
All individual factors in isolation, for the support, makes up for all the factors that go against that role! Other than the goal of winning, there is not a single reason why supports should not play selfishly.
I like Valve's idea: the outgame matchmaking counts mainly win/loss to assign skill points, but ingame encourages playing as support by making support items cheaper, and giving assist gold, ward gold and so on. So in essence, Valve encourages supports to play the role correctly by making them useful without needing to play selfishly to farm items. In other words, your skill as a support and your impact is measured by how well you play your designated role non-selfishly. Sure, you might get a 'bad carry' or someone who made bad decisions every now and then, but on an average your rating will get better this way. IMO, this is better than comparing ingame stats to assign skill, since in the latter method, a support will most likely lose skill points all the time. And that is unfair.
Also I think most people who complain that all signals must be included are flashy mid players and carries. Are there predominantly support players who have asked for all signals to be included to gauge skill? I doubt it.
It's not about PLAY CARRY, or Most effectiver hero. In high to very high skill bracket it comes down to say 1 ward or who has the better support player. Not in like map awareness but overall GPM and ally clutch saves.
Game could be 50/50 and just 1 or 2 death on support more than the other could matter if you're winning or not.
well try UNRANKED you'll get shit faces that go 0-15-0 when you are 9-1-7 then talk shit about you saying you only KS
If I'm really that good, I deserve to be matched with better players, not with you noobs who have 1.5k games but still will lose flat out to me 1v1. Use a valid argument Bogi that doesn't involve the number of games, because I'm sure I'll crush you all. Thanks. I don't have my entire life to waste on dota so I get matched with "better players", matchmaking is flawed as it is and whoever says otherwise is just an asshat who's satisfied with his 4.5k mmr e-peen.
I think your main concern shouldn't be who you are matched with. It should be who you are matched up against. What i find most unfair in ranked games is that even when you solo queue for a ranked game, you face opponents who are from incomplete stacks. Maybe parties of 2 and 3 or 4 and 1. It's simply unfair to have this kind of disadvantage in a game that determines your bloody rank. That's what concerns me the most.
@Vael, You need to realise that there is no system that can perfectly measure skill. And that isn't a problem, of course you can't fucking measure how someone has played in every game through stats alone. That said, there is a correlation between what you would consider good performance and these indicators.
If I gave you two different players, both had identical allies and opponents in all their matches. If one player had an extra 2 KDA and 100 higher XPM/GPM (using the same hero), over 1000 matches, who do you think is the better player. I'm just using those 3 factors, obviously they are using a lot more than that, or at least if performance was an indicator they would. Things like number of TPs used is stupid, same with wards bought and shit, honestly this just sounds like something a retarded support player would talk about.
@Iso, you are too stupid to be playing this game, consider snakes and ladders, it is about as easy as it is to feed with CM every game.
You legit have no idea what you are talking about, much like all the degenerate 20 MMR dumbshits on here. Simply getting a high GPM does not mean you get a good personal score, and there are separate weightings assigned to different factors. The weighted win/loss is the primary component, the individual performance tries to adjust that result by estimating your contribution to the win/loss.
Stealing a last hit gets you the same KDA, maybe you don't know what KDA stands for, but (1+x)/x is the same as (x+1)/x, the minor amount of bonus gold that you get from taking the kill is irrelevant over the long term. Taking last hits doesn't do you any favours either because it makes you more likely to lose the game. Its not a difficult concept. While playing in a particular way improves one factor, it can at the same time have an even bigger negative/positive effect somewhere else.
At the end of the day, over the long term, playing well will result in a higher MMR than playing badly.
I think its particularly funny how its the shit tier "support" players that are complaining about personal performance. You know why you play support? And this doesn't apply to everyone that plays support, but you and the dude with the cat picture both suck balls at this game.
its not that dificult to understand. I agree that there are games that are just almost impossible to win, like some games that usually appears here with one guy 30/0 and the rest of his team 0/20, but the point is that u dont need to win that games to reach a higher MMR. Games like that happen once in a blue moon. What keeps u in ur low MMR are the games that u can carry (or contribute hard to win), but u fail to do so.
http://dotabuff.com/matches/441341656 > 30min and 22 LH playing an Anti-Mage
http://dotabuff.com/matches/440315490 > ur radiance omni didn't contribute much to the win, and ur team did not that bad.
http://dotabuff.com/matches/440275079 > in a game with luna, PL and tiny on ur team u shouldn't farm that much as a venomancer. U took farm of the main heros of ur team and didn't manage to carry them.
Just a few examples of games u could have won if u played better.
You don't need to win all the games and carry ur teammates who has 30 deaths to get a higher MMR. Don't blame the system, ur playing on the exact pool that u should be.
"@Iso, you are too stupid to be playing this game, consider snakes and ladders, it is about as easy as it is to feed with CM every game.
You legit have no idea what you are talking about, much like all the degenerate 20 MMR dumbshits on here."
You use adhominems to prove a point, since you lack reason. LOL! I dont know my MMR, but I will tell you immediately when I get it. Oh by the way Jeff Moser, who helped create/implement Trueskill system says he is a bad player. By your 'logic', his opinions on the skill calculation is invalid, since he has a 'low MMR', amirite?
Value judgments on people because they have low mmr is quite ridiculous. I will ignore your useless opinion on low mmr. If you have a point, put it here. Take your useless, baseless opinion elsewhere.
"Simply getting a high GPM does not mean you get a good personal score, and there are separate weightings assigned to different factors. The weighted win/loss is the primary component, the individual performance tries to adjust that result by estimating your contribution to the win/loss. "
I didnt say 'simply' getting high gpm. I said if you consider the joint distribution of all random variables (signals as you call it), supports will have it lesser perhaps pointwise. I understand your point that individual factors to team *should* be accounted. But I am merely conjecturing it is impossible to account for all the actions that led to a win.
"Stealing a last hit gets you the same KDA, maybe you don't know what KDA stands for, but (1+x)/x is the same as (x+1)/x, the minor amount of bonus gold that you get from taking the kill is irrelevant over the long term. Taking last hits doesn't do you any favours either because it makes you more likely to lose the game. Its not a difficult concept. While playing in a particular way improves one factor, it can at the same time have an even bigger negative/positive effect somewhere else."
I understand both the kda concept and last hits. The point is simple: Show me the linear combination co-efficients (or weights as you call it) of the equation used to update MMR and we *will* find a way to play that will net us mmr points even when we dont win. In other words, any objective that is not winning can be exploited. See Jeff Mosers article.
--------------------
"I think its particularly funny how its the shit tier "support" players that are complaining about personal performance. You know why you play support? And this doesn't apply to everyone that plays support, but you and the dude with the cat picture both suck balls at this game. "
First of all, I am supporting the MMR's win/loss method of calculation of skill and I am NOT complaining. So I am not complaining about personal performance of Supports. I am merely claiming that supports in general will tend to have lesser 'impact signals' in the game. If you think it is wrong, present statistics not opinions.
lmfao 2500 mmr.
youre 2500 because you belong there.
ask any skilled player with over 5k rating and theyll tell you.
anything under 4k is a joke and anybody that is truely skilled at dota can steamroll those games even with a single support.
@Coutunes, ofcourse there are games that i could have been better, or something. I'm not complaining about those games that are even. I'm complaining about games where you clearly can't win, whatever you do. You can do 90/0 but your team sucks so hard that you will loose either way. Sure, the game you did point out as Anti-Mage was terrible for beggining to end. Surely I could have done better, but I was paired with an Earth Spirit (in the lane phase) that didnt know anything about his hero, he didnt know how to combo or place remnants; didnt know ANYTHING. We got completely dominated by Magnus and Lina. And tbh, I don't see how could anyone (better players than me) could do better with a teammate like that. Really, he stole farm, he died early... everything he did was a disaster. Later on, I couldn't farm cause there was no wards, I tried jungling a bit, but their team contested it easily stunlocking me. What could I do, for real? I know i didn't do well as AM (i'm terrible with it), but I didn't have any conditions to play better than that. Omniknight sure that was my fault, I should have done other items, but the game was so easy that I thought it wouldn't last long, and that was my fault. The game was even, no complaints about this one. And the one you pointed out about Venomancer, thank you, I did not see that, I'll try to improove that. I have a problem which is I don't trust my teammates after I see them doing stupid things... It's not that I don't play coletively, but I tend to try to do shit on my own. Like, if i'm playing a versatile hero (like Naga or Veno) and I see that my carriers are failing hard, I try to set into a carry role, the same goes with supporting. I know that this is not right, but I don't think we were going to win anyways even if I didnt steal their farm (which I didn't realize I was doing, as I always try to focus on denying and leaving the carry free farm {atleast i try it}; what pisses me off is that even free farming they still miss silly LH, and i'm not talking about 3 or 4 'cus shit happens, i'm talking about like every single one... that makes me not to rely on that player, I know it's wrong but I can't help it.}). Anyway, thank you very much for pointing out my mistakes, surely it'll make a better player out of me. If people here were more like you instead of going "ROFL 2500 MMR", I think this community would be a better place. Again, ty!
@Wei_, my friend is about (or over, dunno for sure) 5k rating and he said that he is still paired with complete dumbasses that surely dont belong that match. He is kinda skilled and we were playing together and even he could manage to carry most games, there were some games that he couldn't do anything, cause Riki just thought rushing Vladmir without boots, tangos or any other items would be a great idea, and the Pudge thought that maxing Rot and Flesh Heap would be better than put a single point in Hook... the point is that if you are facing players that are supposed to be on the same level as yours, you shouldn't be playing as a 5K MMR PRO (when you are 2.5k) to stand a chance to win. You should be playing like 2.5k (as everyone else in the game should) and still have a chance. But instead, doesn't matter how good you are, there are certain situations where you can't do shit, like your opposite team is so fed that you can't do shit cause you are underleveled, underfarmed and your towers are going off one by one. Ofc you could do something to stop that from hapenning, but you can't be in 3 lanes at the same time (unless you're meepo lol), so even if you could stop some guys from feeding, if you are paired with terrible players, the two others will do their job and feed like hell. I'm not claiming myself to be the best shit out there, clearly i'm not, but it's rare when I feed the other team like if it was my first game ever, I can't get is how can this happen so often if I'm supposed to play with people close to my skill level.
Anyway, thanks for every reply.
Wei_ another mid or feed idiot (i can see by your heroes),btw,problem is starting ranking not after that,you can move up and dawn,but it is really slow,you want to say that SmallStepsCorrupt (see in lines before) who played 3K matches and have win ratio above 50% belongs to noob group who are 20 level and have same MMR like he have ???
System is far far from good but we can hope that will be better.
Anyway,for all players who got feeling that is something wrong ,play CM-CD and forget AP,noobs in AP knows to play 3 heroes perfect and play only that no meter what are lineup demands....
I really want to see SD-RD modes in Ranked matches,hope one day it will be....
Please sign in to post comments.
So, yea. You were supposed to play with people at a close level as yours, weren't you? Ok, sometimes it happens, but sometimes (mainly when you are in a 4 games winning streak or so) you are put together with guys that clearly do not have your level; you can't do anything, really.
I can't carry a game alone (and i don't think i should, dota is a team game they say), so ofcourse I rely on my team. What to do when the team dont have any knowledge about the game, positioning, etc.? I find myself stuck on the 2500 MMR, I get up to 2580, then I loose 3 or 4 games in a row and get back to 2400's.
I saw some ppl saying that if you couldn't carry the game, that it may be your fault. But, wait, what? I saw a topic where a guy told exactly that to a weaver who did something like 17/1, but his teammates were like 0/13, 1/15. I don't think Dendi could carry a game like this, lol... one point: you were not supposed to play like the better player that ever walked down on earth to have a chance to win. Instead, the purpose of ranked matches are not to make teams 50%/50% (or about it)? There is sometimes that you know that you'll loose the game don't matter what you do, i.e your team feeding like crazy.
I'm not here to say that im a pro or something, because i'm clearly not (got back to dota last month or so). But i'm a solid player (in the levels that i'm used to play) and i keep getting mixed with guys that do not have any conditions to be playing there. By 2500 MMR i think i was supposed to play with average players. Instead, I get in a bunch of newbies. Today I played with a Shadow Demon that didnt know what Disruption does (so he didn't use it), he did buy wards but didn't place it, etc. Bought mekansm (i guess) and did'nt use it, things like that. And the other team was full of ok players, who were getting fed easily cause my team didn't stand any chance to resist.
I admit, that happens a lot when I win too. I played only 3 or 4 matches where the teams were even, that my skill could have changed the result of the match. The rest is matches where you know you are going to win/loose in the first 5 minutes played.
And i noticed that those newcommers are in a party, so... wouldn't it be easier to set SOLO QUEUE vs SOLO QUEUE and PARTY vs PARTY? SOLO x SOLO would, atleast, have people that probably never played with each other, so there wouldn't be teamplay advantages like we experience today.
Ofcourse that is a not very good question, but i'll ask it anyway. What are the heroes that have most game impact? I wan't to get to 3000-3500 MMR where I SUPPOSE i won't get those newbies (yea, ofc i will get randomed with bad players, but atleast they should know what their skills do, what wards are and where to place them, etc). I think that those who have most game impact are mostly mid heroes, like QoP, Storm Spirit, Templar Assassin, Dragon Knight, etc. Well, mostly semi-carries.
I don't have a specific role, sometimes I play hard carry, sometimes I go for support when no one else does (and I think I do an ok job, compared to the supports i see around), etc. The problem is that normally gankers/semi carriers are the first pick of every team, so I get afraid of no one going for support or even hard carry.
What do you think is the best way to get your MMR up (ofc its winning lol)? I was thinking on play Io, cause i see it as a very fun and strong hero, but I think it relies way too much on teammates (yea hes not much of a hero on his own), and rarely I get a reliable carry to pair with. I see some guys with 80%+ winrate (with 1000+ games) with Io, on the other hand, Io's general winrate is about 40%.
Well, that's it. Sorry 'bout my english, not my mother language, as I bet you can notice.