well zenoth, i did say "almost everything Relentless said" is true. i won't comment on the part about MM creating inherently lopsided games before picks, but i don't think that's true. i didn't perceive a major disagreement on any issue other than the one you just brought up, so me stating that "Relentless is accurate on almost everything" doesn't mean I think you are wrong on most things; quite the opposite.
the only case i can think of is: if there's a stack of 2000 + 5000 + 5000 + 5000 + 5000 players on one team and the other team has 4 "5000" rated players on a stack + 1 "2000" solo queuer, then that's a different story, since if the 2000 solo goes mid/carry, the 4 stack is fucked, but i haven't run into that problem much (maybe 1 out of every 30-40 games), and such a rare occurrence doesn't bother me.
^
I agreed about what u said that ap is not pretty balanced because of the different power of each hero.
Such example, from what i see in this current patch is (6.80) people always picking invoker in the first time, and team who have invoker as their mid teammate have more advantage to win due to the imbalance of the invoker itself.
And what u said mostly is true, its so much easier to me, climbing from 4,5k-5k using lycan. Then climbing 4k-4,5k using many heroes, mirana,clinkz,lifestealer,timbersaw etc.
Because as what zenoth said, in ap it so much easier win with pushing style( Independent), split push then snowballing style (dependent on your teammates)
Anyway, thanks for all of ur discussion. it is so awesome to read
Correct, in order to win consistently, you need to be good with at least 2-3 of the strong heroes. Maybe 1 if you're a slark/lycan picker.
If you want to play heroes not currently seen in the pub meta (such as phantom lancer, a hero I love playing), I advise -cd instead. Ban the strong heroes so you can maintain a more "slow paced/farming" tempo.
Non-support strong heroes to grind MMR with if you really want:
Almost all of the 4.5-5k games have one or more of these in 90% of games, and I'm sure the same applies to 4.5 - 6k games. The only hero in that list I play often is Terrorblade (65% win rate in 29 games...so I gained 225 mmr just from playing him), so if I wanted to grind more MMR I would either have to learn one of the other heroes and play them well or just keep picking Terrorblade.
I don't place much weight on MMR anymore and I think I'm just gonna "play for fun" with friends. Even beesa (6k rated) told me he doesn't think the number means much. When I asked him how to reach higher MMR he just told me the same thing I'm telling you right now, and every player on this forum who understands the game well enough (not every hero is equal) will agree.
The downside to all of this is that if you ask one of these "same hero pickers" to play something that actually requires a bit of skill to pull off, in a role they aren't used to, (earthshaker as support, kunkka mid, etc.) they will fail and come off as overrated retards.
off topic, but i have a pretty bad-mannered (though hilarious) friend at 4.5k mmr. he plays pudge like 70% of the time in ranked, and he always calls people near or even slightly above 5k rated "trash," "retards," "no skill," etc.
at first i wondered why he would say that. after all, if they're rated more highly than he is, doesn't that mean they are better? (if i asked him that he would probably flame me)
now that i think about it, his argument isn't without merit. if he picked one of those heroes i listed above instead of pudge mid (let's face it, 90% of pudge pickers are garbage) every single game i'm sure he would be > 5k as well in less than a month, but he plays the game for fun. which is perfectly fine - to each their own. it's not like people above 5k queue into a fairy tale land where all the sub 5k peasants suck their dicks and support them zealously; quite the opposite actually.
kinda agree about that, although talking trash to other people dont make us look better.
In my personal opinian it will much better if valve decided that ranked mmr only exist on cm/cd. which mean, i guess valve wont do that.
So yeah, people will still trap in their bracket if they dont want to climb using op hero.
Probably not all games that end up in blowouts are matching failures. And some matching failures no doubt end in a "surprise" win for the team that was very likely to lose.
But setting the MMR of the teams to be somewhat equal as Valve does is not going to consistently create games where both teams have an equal chance to win. It's sort of close... but the odds really only average out eventually to be a 50% chance to win over the course of lots of games. Individual games are not 50% chance to win. Some are 40% some are 60%... some are really way off the mark.
http://dotabuff.com/matches/537705693 This game IS a forced win. My team was way way better than the other team. Watch the replay. It's not about picks. Our players are ALL significantly higher skill level and more dota knowledge than ALL of the other teams players. I don't know why this particular game was so unfair. But it was unfair. It was not picks. It was not early mistakes snowballing. We would have won this game 9 times out of 10 with these teams.
They don't lose because medusa, morphling, luna is 3 carries and that requires to much farm.... Yes, they did pick too many carries. But they lost dramatically and immediately because they got outplayed everywhere. They were slower, less accurate, and less knowledgable... AND made bad picks. If they made good picks we would still easily beat them. It would have just taken slightly longer to make it obvious that our team was superior.
relentless how would u even go about testing that counterfactual? "even if they picked a better lineup they would still lose?" that MAY be true, but even so, i can't even begin to think about how to test such a claim.
There is no way to test it unless we actually get the players to all play with those teams again for 10 games or whatever. But I am quite certain the Dire team would be happy to play those 10 games and the Radiant team would refuse.
If you watch the perspective of each of the players in the replay you will see it. The Radiant team is getting out played in all 3 lanes. They are losing not because of the picks, but because they are slower, because they are intimidated and confused, because they are not able to play the lane against people significantly better at dota than they are.
medusa and morph are part of a stack. i don't know if there are any other stacks in that game, but there is one point that hasn't been brought up yet, but relentless pointed this out in another thread: party and solos play together. so if i'm rated 5.5k party (which i'm not, but lets say I am) and 4.7k solo, there's a good chance I will drag down the team if I don't play support and stupidly decide to take mid instead when the other players are all more highly rated, because my party rating is inflated via stacking with better players. the party rating and the solo ratings, for most people, will not be significantly different (+/- 500), but for some they can diverge significantly.
in that case, i suppose you can blame matchmaking and the way it "records" people's performances.
You know full well that someone playing on an unfamiliar hero will be slower to react, take longer to make basic decisions, worse at last hitting etc etc. Compound this with massively disadvantageous picks and it is very easy to see why you think that their skill level is far inferior. But this isn't necessarily the case.
Yes, but that is not what is happening here. We have 6 of the players with stats on so we have at least some information.
These are the records on the heroes used before this game.
Radiant has CM, Medusa, Morphling
CM has 18 win, 4 loss
Medusa 1 win, 1 loss
Morphling 1 win, 1 loss
Dire has Lina, Puck, Sven
Lina 15 win, 9 loss
Puck 1 win, 2 loss
Sven never played
There is not an advantage in heros being unfamiliar. Watch the 64% win Lina lane against the 74% win CM. Watch Lina place far better wards. Watch Lina cast spells better, harass better, lasthit better, move around the map better... outclass the CM in every aspect of the game.
Maybe the annoymous Pudge had not played much pudge... but he can't have played much less than the Puck who was against him who came into the game at 1-2 win/loss. Nevertheless the Puck was far better than the Pudge. I would expect a Puck to beat a Pudge mid, but it was a walkover. The best thing the Pudge managed to do was kill himself.
The Morphling may have been unexperienced with morphling... but he dies to the sven, who has never played sven before this game... morphling dies because his positioning is bad and he is slow. The sven may not have played sven in the last 500 games but he knows dota. The morphling is totally unprepared to face an enemy who might do something like actually kill him early in the laning phase. Morphling plays like laning is only for lasthitting and no one would dare actually try to kill you. He doesn't lose because he is morphling... he loses because he plays like a nub vs a player who can recognize when you are out of position and kill you.
So the question is Why is the Lina way better than the Crystal Maiden? Why does the Lina know how to ward so the CM can't deward it? Why can the Lina control a hero well enough to harrass, deny, and lasthit with attacks... but the CM can't? Why is the Sven way better than the morphling? Why is the Puck way better than the Pudge? Why is the bristleback competant, but the Luna spamming 10 level 2 lucent beams in a futile effort to drive a bristleback out of the lane? Why is the Luna unprepared to face a Lina that can actually hit him with a stun? If the Luna was really a similar level player should he not expect the Lina to be dangerous and stay back?
I think that the best explanation of these things is that the players on Dire are underated and the players on Radient overated so that while their MMRs are similar, they are NOT similar in actual skill at dota.
Your analysis is plausible. I haven't looked at the replay personally and it will be difficult for me to objectively discern the skill differences since I typically analyze by comparing with what I would do, so as a player in the game I'll take your word for it.
However, the point still stands. Your explanation is a good one, and probably the most likely one. But that then is a problem with the rating system itself. Everyone plays in a range of values, but matchmaking is only capable of using one value, and a possibly inflated/deflated one - yes, when everything lines up and you have a team of underrated players against a team of overrated players, one team is far more likely to win before the picks even start. This is something that will happen even in the Elo system for chess, where one player might be really a 1400 player but rated at 1300 after losing a series of games while inebriated; if he plays against a 1300 player who just scraped his way there after a few lucky wins, of course the first player will be much more likely to win. But this is an inherent fault of the rating system itself, which can only hope to capture the approximate skill level of the player.
If your case is that matchmaking is flawed because of these games, I concede that such games will happen now and then within a large sample size, but that it isn't attributable to matchmaking. Furthermore many games that were possibly balanced before picks but ended up in a one-sided stomp will be mistaken as such games.
Guys MMR is just a number, you cant measure skill in a game with so many variables like dota with just a simple number, the system can get closer maybe, but not 100 % exactly ( hell i would say its not even 80 % ), i mean in most of the situations a 5k player will be better than a 3k player, but lets say, for example that someone only plays ranked once and when that person gets his mmr ( lets say 4k ) and he never plays ranked again, in some months the guy ( if he actually tried to improve ) will probably play like a 5 k player, and when he is back to ranked he is going to stomp all of his games, and this system will not work because it will not give every player a balanced match ( the enemies of this player will play at the 4k lvl when he is playing at a 5 k lvl ).
What i am trying to say is that people should stop giving matchmaking so much importance, they should focus more on their game and trying to improve, you dont need a number to tell you that you are better now than before.
i think the other problem is that dota is a game with many factors. MMR is supposed to account for all of these indirectly, but it's a pretty poor predictor of how well someone plays in any particular game. i have a friend rated 5.3k and he got a bunch of teammates who got outpicked in lane phase, and then got raped by a fed riki. He told me none of his allies (who should be rated 4.5-5k) got a single sentry/dust, and he had to buy all of it as a mid QoP. I mean, shouldn't 4k (and even 3.5k/3k players if they are knowledgeable enough) players know some basics of this game? If there's at least 1 invis hero get detection...lol.
"Skill" in the abstract sense may not even be a linear concept. Someone might be "skilled" in x area but severely deficient in y area. They can still reach a certain MMR just by emphasizing their x/capitalizing on its advantages, but it doesn't mean they're well-rounded.
Lack knowledge of how to use consumables is extremely common in pub games.
Many people do not buy obs wards because the only ward spot they have a little knowledge about is rune warding. And even if they do know ward spots they don't know when and how to place the wards.
Many people do not buy sentries because the have zero idea of how to deward... placing a sentry correctly requires not only knowledge of ward spots generally but watching the enemy supports enough to know which ones they are currently using. Placing a sentry to defend a location means you have to predict where an invisible hero will be having never seen them in the first place. Using dust is easy, really at least 2 and probably 3 heroes should be carrying dust mid through late game against invis dependent teams. But since pubs don't use it much they often fail when they do use it. They dust late and can't stun or slow or don't even see where the invis hero went to escape or they dust late after the invis hero has already attacked. Because they have failed in using dust before, they don't rely on it.
TP Scrolls are not purchased and not used because people try to use them and fail porting out to late (get stunned die anyway), porting in too late (team already died now you add another death), porting uselessly to a tower they can't defend anyway, porting too late base already died... pubs have a lot of failures of tp scrolls in their mind. Though using it may seem simple, its also easy to make a big mistake with slightly wrong timing.
Smokes are often used very badly. People smoke at no targets because ward vision is poor and waste time. People smoke into a surprise fight with the enemy team they can't win. People smoke in enemy ward vision (even creep vision) and then waste time. People smoke 5 so its super obvious that they are doing rosh or ganking as they all disappear from the map. Because of failed smokes, people are afraid to use them and don't buy them.
Salves... who uses salves to heal their team so they can keep farming or pushing? That just never happens unless I do it. It has huge impact, but people think of their items as their personal property. They never think of an outside the box use like, keeping your carry in a lane to farm and push and follow up an advantage after a fight.
Most 4k to 5k MMR players got where they are by massing games of their favorite snowball hero. Do they deserve the score? Yes... for that hero. If they try to play something else they instantly become a liability.
That's a factor that make some matches much less fair than Valve might have set it up. But picks are not separate from MMR. Picking a hero that will win the game is also a skill that has a big impact on MMR. As I always tell people if you pick heroes that you know how to use to win your MMR will go up. If you pick heroes that often lose for you, then you lose MMR.
Valve can try to make the game "fair before picks" but picks are also part of the game. A higher MMR player should pick better. If they don't that mistake might be enough to lose the game. I think people should be expected to make tryhard picks in ranked games... but there is no way to force them to do it.
If you are in a game that was lost by picks you just have to re-focus on your execution. Make playing your hero well for practice the objective. If that does not win the game because of picks... then acknowledging that some games really are not fair will help to ease the frustration of the certain impending loss. You may be helpless to impact the ultimate outcome of a game, but you can still outplay enemy heroes in individual fights.
It's more or less because playing those snowball heroes allows u to control the tempo of the game and allows you to lift really heavy games whereas playing as a support you're pretty much worthless and contribute far lesser
cliff jungling prophet does nothing all game but feed and make the team rage.
look at his games history/skill builds/etc. beyond a shred of reasonable doubt - he's obviously throwing in as many games as possible but Valve hasn't banned him.
sounds like a forced loss when people get paired with throwers/utter shitcans (i don't mean just people playing badly, i mean literally the worst scum you can find in a dota game).
Relentless while I agree with you I think a lot of people don't take ranked seriously enough to just pick the heroes they are the best with or the most broken right now. Almost everyone I know just randoms. Sure they might repick if they get some weird hero like Meepo or if our lineup is horrible but in the end we play this game for fun and just playing TB/Phoenix/old ES is not fun for everyone. The MMR system should be simple (as it is). I don't think anyone claims it's amazing or anything because it's really nothing special.
this discussion is really really great, keep it up.
So insightful
well zenoth, i did say "almost everything Relentless said" is true. i won't comment on the part about MM creating inherently lopsided games before picks, but i don't think that's true. i didn't perceive a major disagreement on any issue other than the one you just brought up, so me stating that "Relentless is accurate on almost everything" doesn't mean I think you are wrong on most things; quite the opposite.
the only case i can think of is: if there's a stack of 2000 + 5000 + 5000 + 5000 + 5000 players on one team and the other team has 4 "5000" rated players on a stack + 1 "2000" solo queuer, then that's a different story, since if the 2000 solo goes mid/carry, the 4 stack is fucked, but i haven't run into that problem much (maybe 1 out of every 30-40 games), and such a rare occurrence doesn't bother me.
^
I agreed about what u said that ap is not pretty balanced because of the different power of each hero.
Such example, from what i see in this current patch is (6.80) people always picking invoker in the first time, and team who have invoker as their mid teammate have more advantage to win due to the imbalance of the invoker itself.
And what u said mostly is true, its so much easier to me, climbing from 4,5k-5k using lycan. Then climbing 4k-4,5k using many heroes, mirana,clinkz,lifestealer,timbersaw etc.
Because as what zenoth said, in ap it so much easier win with pushing style( Independent), split push then snowballing style (dependent on your teammates)
Anyway, thanks for all of ur discussion. it is so awesome to read
Correct, in order to win consistently, you need to be good with at least 2-3 of the strong heroes. Maybe 1 if you're a slark/lycan picker.
If you want to play heroes not currently seen in the pub meta (such as phantom lancer, a hero I love playing), I advise -cd instead. Ban the strong heroes so you can maintain a more "slow paced/farming" tempo.
Non-support strong heroes to grind MMR with if you really want:
-Invoker
-Weaver
-Storm
-Slark
-Terrorblade
-Ember
-Timbersaw
-Phoenix
-Mirana
Almost all of the 4.5-5k games have one or more of these in 90% of games, and I'm sure the same applies to 4.5 - 6k games. The only hero in that list I play often is Terrorblade (65% win rate in 29 games...so I gained 225 mmr just from playing him), so if I wanted to grind more MMR I would either have to learn one of the other heroes and play them well or just keep picking Terrorblade.
I don't place much weight on MMR anymore and I think I'm just gonna "play for fun" with friends. Even beesa (6k rated) told me he doesn't think the number means much. When I asked him how to reach higher MMR he just told me the same thing I'm telling you right now, and every player on this forum who understands the game well enough (not every hero is equal) will agree.
The downside to all of this is that if you ask one of these "same hero pickers" to play something that actually requires a bit of skill to pull off, in a role they aren't used to, (earthshaker as support, kunkka mid, etc.) they will fail and come off as overrated retards.
yeah, thats really true. Thats why when i played other heroes (non strong heroes) i didnt play at mmr, but casual pub.
The purpose of playing should be fun, and it will not be fun if we play the same heroes every time.
off topic, but i have a pretty bad-mannered (though hilarious) friend at 4.5k mmr. he plays pudge like 70% of the time in ranked, and he always calls people near or even slightly above 5k rated "trash," "retards," "no skill," etc.
at first i wondered why he would say that. after all, if they're rated more highly than he is, doesn't that mean they are better? (if i asked him that he would probably flame me)
now that i think about it, his argument isn't without merit. if he picked one of those heroes i listed above instead of pudge mid (let's face it, 90% of pudge pickers are garbage) every single game i'm sure he would be > 5k as well in less than a month, but he plays the game for fun. which is perfectly fine - to each their own. it's not like people above 5k queue into a fairy tale land where all the sub 5k peasants suck their dicks and support them zealously; quite the opposite actually.
kinda agree about that, although talking trash to other people dont make us look better.
In my personal opinian it will much better if valve decided that ranked mmr only exist on cm/cd. which mean, i guess valve wont do that.
So yeah, people will still trap in their bracket if they dont want to climb using op hero.
Probably not all games that end up in blowouts are matching failures. And some matching failures no doubt end in a "surprise" win for the team that was very likely to lose.
But setting the MMR of the teams to be somewhat equal as Valve does is not going to consistently create games where both teams have an equal chance to win. It's sort of close... but the odds really only average out eventually to be a 50% chance to win over the course of lots of games. Individual games are not 50% chance to win. Some are 40% some are 60%... some are really way off the mark.
http://dotabuff.com/matches/537705693 This game IS a forced win. My team was way way better than the other team. Watch the replay. It's not about picks. Our players are ALL significantly higher skill level and more dota knowledge than ALL of the other teams players. I don't know why this particular game was so unfair. But it was unfair. It was not picks. It was not early mistakes snowballing. We would have won this game 9 times out of 10 with these teams.
They don't lose because medusa, morphling, luna is 3 carries and that requires to much farm.... Yes, they did pick too many carries. But they lost dramatically and immediately because they got outplayed everywhere. They were slower, less accurate, and less knowledgable... AND made bad picks. If they made good picks we would still easily beat them. It would have just taken slightly longer to make it obvious that our team was superior.
relentless how would u even go about testing that counterfactual? "even if they picked a better lineup they would still lose?" that MAY be true, but even so, i can't even begin to think about how to test such a claim.
There is no way to test it unless we actually get the players to all play with those teams again for 10 games or whatever. But I am quite certain the Dire team would be happy to play those 10 games and the Radiant team would refuse.
If you watch the perspective of each of the players in the replay you will see it. The Radiant team is getting out played in all 3 lanes. They are losing not because of the picks, but because they are slower, because they are intimidated and confused, because they are not able to play the lane against people significantly better at dota than they are.
medusa and morph are part of a stack. i don't know if there are any other stacks in that game, but there is one point that hasn't been brought up yet, but relentless pointed this out in another thread: party and solos play together. so if i'm rated 5.5k party (which i'm not, but lets say I am) and 4.7k solo, there's a good chance I will drag down the team if I don't play support and stupidly decide to take mid instead when the other players are all more highly rated, because my party rating is inflated via stacking with better players. the party rating and the solo ratings, for most people, will not be significantly different (+/- 500), but for some they can diverge significantly.
in that case, i suppose you can blame matchmaking and the way it "records" people's performances.
You know full well that someone playing on an unfamiliar hero will be slower to react, take longer to make basic decisions, worse at last hitting etc etc. Compound this with massively disadvantageous picks and it is very easy to see why you think that their skill level is far inferior. But this isn't necessarily the case.
Yes, but that is not what is happening here. We have 6 of the players with stats on so we have at least some information.
These are the records on the heroes used before this game.
Radiant has CM, Medusa, Morphling
CM has 18 win, 4 loss
Medusa 1 win, 1 loss
Morphling 1 win, 1 loss
Dire has Lina, Puck, Sven
Lina 15 win, 9 loss
Puck 1 win, 2 loss
Sven never played
There is not an advantage in heros being unfamiliar. Watch the 64% win Lina lane against the 74% win CM. Watch Lina place far better wards. Watch Lina cast spells better, harass better, lasthit better, move around the map better... outclass the CM in every aspect of the game.
Maybe the annoymous Pudge had not played much pudge... but he can't have played much less than the Puck who was against him who came into the game at 1-2 win/loss. Nevertheless the Puck was far better than the Pudge. I would expect a Puck to beat a Pudge mid, but it was a walkover. The best thing the Pudge managed to do was kill himself.
The Morphling may have been unexperienced with morphling... but he dies to the sven, who has never played sven before this game... morphling dies because his positioning is bad and he is slow. The sven may not have played sven in the last 500 games but he knows dota. The morphling is totally unprepared to face an enemy who might do something like actually kill him early in the laning phase. Morphling plays like laning is only for lasthitting and no one would dare actually try to kill you. He doesn't lose because he is morphling... he loses because he plays like a nub vs a player who can recognize when you are out of position and kill you.
So the question is Why is the Lina way better than the Crystal Maiden? Why does the Lina know how to ward so the CM can't deward it? Why can the Lina control a hero well enough to harrass, deny, and lasthit with attacks... but the CM can't? Why is the Sven way better than the morphling? Why is the Puck way better than the Pudge? Why is the bristleback competant, but the Luna spamming 10 level 2 lucent beams in a futile effort to drive a bristleback out of the lane? Why is the Luna unprepared to face a Lina that can actually hit him with a stun? If the Luna was really a similar level player should he not expect the Lina to be dangerous and stay back?
I think that the best explanation of these things is that the players on Dire are underated and the players on Radient overated so that while their MMRs are similar, they are NOT similar in actual skill at dota.
You guys deserve a nobel or some shit for talking like normal people in a dota forum, seriously.
Your analysis is plausible. I haven't looked at the replay personally and it will be difficult for me to objectively discern the skill differences since I typically analyze by comparing with what I would do, so as a player in the game I'll take your word for it.
However, the point still stands. Your explanation is a good one, and probably the most likely one. But that then is a problem with the rating system itself. Everyone plays in a range of values, but matchmaking is only capable of using one value, and a possibly inflated/deflated one - yes, when everything lines up and you have a team of underrated players against a team of overrated players, one team is far more likely to win before the picks even start. This is something that will happen even in the Elo system for chess, where one player might be really a 1400 player but rated at 1300 after losing a series of games while inebriated; if he plays against a 1300 player who just scraped his way there after a few lucky wins, of course the first player will be much more likely to win. But this is an inherent fault of the rating system itself, which can only hope to capture the approximate skill level of the player.
If your case is that matchmaking is flawed because of these games, I concede that such games will happen now and then within a large sample size, but that it isn't attributable to matchmaking. Furthermore many games that were possibly balanced before picks but ended up in a one-sided stomp will be mistaken as such games.
Guys MMR is just a number, you cant measure skill in a game with so many variables like dota with just a simple number, the system can get closer maybe, but not 100 % exactly ( hell i would say its not even 80 % ), i mean in most of the situations a 5k player will be better than a 3k player, but lets say, for example that someone only plays ranked once and when that person gets his mmr ( lets say 4k ) and he never plays ranked again, in some months the guy ( if he actually tried to improve ) will probably play like a 5 k player, and when he is back to ranked he is going to stomp all of his games, and this system will not work because it will not give every player a balanced match ( the enemies of this player will play at the 4k lvl when he is playing at a 5 k lvl ).
What i am trying to say is that people should stop giving matchmaking so much importance, they should focus more on their game and trying to improve, you dont need a number to tell you that you are better now than before.
Like sing is saying here http://www.twitch.tv/sing_sing/c/1348585 you dont win anything saying that your team is awful.
i think the other problem is that dota is a game with many factors. MMR is supposed to account for all of these indirectly, but it's a pretty poor predictor of how well someone plays in any particular game. i have a friend rated 5.3k and he got a bunch of teammates who got outpicked in lane phase, and then got raped by a fed riki. He told me none of his allies (who should be rated 4.5-5k) got a single sentry/dust, and he had to buy all of it as a mid QoP. I mean, shouldn't 4k (and even 3.5k/3k players if they are knowledgeable enough) players know some basics of this game? If there's at least 1 invis hero get detection...lol.
"Skill" in the abstract sense may not even be a linear concept. Someone might be "skilled" in x area but severely deficient in y area. They can still reach a certain MMR just by emphasizing their x/capitalizing on its advantages, but it doesn't mean they're well-rounded.
Lack knowledge of how to use consumables is extremely common in pub games.
Many people do not buy obs wards because the only ward spot they have a little knowledge about is rune warding. And even if they do know ward spots they don't know when and how to place the wards.
Many people do not buy sentries because the have zero idea of how to deward... placing a sentry correctly requires not only knowledge of ward spots generally but watching the enemy supports enough to know which ones they are currently using. Placing a sentry to defend a location means you have to predict where an invisible hero will be having never seen them in the first place. Using dust is easy, really at least 2 and probably 3 heroes should be carrying dust mid through late game against invis dependent teams. But since pubs don't use it much they often fail when they do use it. They dust late and can't stun or slow or don't even see where the invis hero went to escape or they dust late after the invis hero has already attacked. Because they have failed in using dust before, they don't rely on it.
TP Scrolls are not purchased and not used because people try to use them and fail porting out to late (get stunned die anyway), porting in too late (team already died now you add another death), porting uselessly to a tower they can't defend anyway, porting too late base already died... pubs have a lot of failures of tp scrolls in their mind. Though using it may seem simple, its also easy to make a big mistake with slightly wrong timing.
Smokes are often used very badly. People smoke at no targets because ward vision is poor and waste time. People smoke into a surprise fight with the enemy team they can't win. People smoke in enemy ward vision (even creep vision) and then waste time. People smoke 5 so its super obvious that they are doing rosh or ganking as they all disappear from the map. Because of failed smokes, people are afraid to use them and don't buy them.
Salves... who uses salves to heal their team so they can keep farming or pushing? That just never happens unless I do it. It has huge impact, but people think of their items as their personal property. They never think of an outside the box use like, keeping your carry in a lane to farm and push and follow up an advantage after a fight.
Most 4k to 5k MMR players got where they are by massing games of their favorite snowball hero. Do they deserve the score? Yes... for that hero. If they try to play something else they instantly become a liability.
That's a factor that make some matches much less fair than Valve might have set it up. But picks are not separate from MMR. Picking a hero that will win the game is also a skill that has a big impact on MMR. As I always tell people if you pick heroes that you know how to use to win your MMR will go up. If you pick heroes that often lose for you, then you lose MMR.
Valve can try to make the game "fair before picks" but picks are also part of the game. A higher MMR player should pick better. If they don't that mistake might be enough to lose the game. I think people should be expected to make tryhard picks in ranked games... but there is no way to force them to do it.
If you are in a game that was lost by picks you just have to re-focus on your execution. Make playing your hero well for practice the objective. If that does not win the game because of picks... then acknowledging that some games really are not fair will help to ease the frustration of the certain impending loss. You may be helpless to impact the ultimate outcome of a game, but you can still outplay enemy heroes in individual fights.
It's more or less because playing those snowball heroes allows u to control the tempo of the game and allows you to lift really heavy games whereas playing as a support you're pretty much worthless and contribute far lesser
another unwinnable game
http://dotabuff.com/matches/540079321
cliff jungling prophet does nothing all game but feed and make the team rage.
look at his games history/skill builds/etc. beyond a shred of reasonable doubt - he's obviously throwing in as many games as possible but Valve hasn't banned him.
sounds like a forced loss when people get paired with throwers/utter shitcans (i don't mean just people playing badly, i mean literally the worst scum you can find in a dota game).
terra why are you offline on steam
dunno, i just logged on to another account. wanna play?
gonna play with melody and simpsons soon
Relentless while I agree with you I think a lot of people don't take ranked seriously enough to just pick the heroes they are the best with or the most broken right now. Almost everyone I know just randoms. Sure they might repick if they get some weird hero like Meepo or if our lineup is horrible but in the end we play this game for fun and just playing TB/Phoenix/old ES is not fun for everyone. The MMR system should be simple (as it is). I don't think anyone claims it's amazing or anything because it's really nothing special.