my friend tells 4k is shit. some of the worst players stuck in 4k... i mean flat 4k! he goes to 4k and again falls to 3.9k. yeah ofc 3.3k is seriously frustrating. had a hard time increasing from 3.3k to 3.8k(atm)
ggwp welcum yet another retard who doesnt know what comprehension mean
4k = similar to 4.7k
became
4k = harder than 5k
smurf logic
o cool so we can have this douchebag
but other "offensive/derogatory" comments aren't
anyhow, yeah i agree that the difference between 4-4.7ish is generally minimal
3.9ish and below is generally a larger skill difference
and it makes sense to me. it's not really elo heaven/hell, but the basis of it seems logical
i guess, for those of us who play enough and make it (and deserve it) to the 4k MMR bracket
end up "meeting our match" so to speak
if you took away all the games i was "carried", as in, i barely did shit all
and if you took away all the games where there was an intentional griefer, i.e. destroying items, taking items, feeding, boosters, etc
im pretty certain my MMR wouldn't change much
idk, at this point, it's difficult for me to analyze how i win or lose in full detail
am i on my "A game?" when i go on a winning streak? or am i getting carried? maybe it's just a bit of luck, or it could be a combination of things...
what about a losing streak? am i "creating bad vibes because i had a bad game or two?", am i getting a bit unlucky, or could it be a bit of both and more?
u expect 4000 mmr and 4100 mmr people to be much different? at that point it's mostly about luck or if you're really good but then you'll easily get to 5k
4000 to 4700 is a huge difference in skill... This is how I've generalized similar player skill:
4000-4300
4400-4600
4700-4800
4900-5100 (maybe just 4900-5000)
You can even tell it's somewhat true based on the end game MMRs. Below 4300 nearly everyone has the same MMR. As you get higher, the MMRs tend to be more spread out because there are less players because the skill gets higher. I've noticed MMRs tend to vary significantly once you get around 4500 to 4600.
I'm pretty much vsing 4.5k players all the time as a 4k flat player.
No I do not find them amazing.
I just failed to carry 3 horrible 4.5k players.
So yes I do not feel there is a much of a difference. 4.65k Invoker, dagon eb midas. 0-3-7. His 4.5k supports? Level 1 jungle farm as SK died to neutrals. Yes, I sincerely do not believe there is a huge difference. Feels like 4k flat to me.
seems like to me you're just trying to subtly tell people you're better than most of the 4k players you solo queue against hmmmm
Seems to me some people can't accept the fact most 4k players are at the same level? If you want to take it as me showing off, so be it. I don't know how saying being breakeven in games in both 4k and 4.7k sounds like showing off. It sounds more like me trying to show how similar the two brackets are.
Some people seem to be butthurt that their 500-700 mmr difference isn't as big as it seems.
500 MMR have a huge diff. Mostly 4k players have very poor game sense, they do not know how to rotate efficiently and how to gank core heroes blah blah blah, and they farm most of the time/run around inefficiently. Same for 5k MMR comparison, more efficiency , less deaths/overextension etc.
Wizard, entire 4k isn't the same. That's like saying 3000-3999 is the same. Or 2000-2999 for that matter.
The skill differs greatly between 4100-4300 and 4600+. Even though people has bad games at all mmrs, which may make them look like total douchebag, or account buyers rather. It's a team game, and you can't expect your team-mates to have their game every time.
Everythink from cockiness like playing when tired as fuck or drunk.
I don't understand, if you've been pwning all these 4700 players why haven't you ranked up to their rating yet? Or you aren't really winning more than 50% of your games?
There's a flaw in your theory somewhere. You can't say oh in this match I was mid vs a 4700 mid because your team mmrs are all about even so despite you supposedly winning mid, if you're at 50% you didn't do something well enough to win more than 50% of your games and rank up.
it might seem youűre better cause your team just does a lot more... now try doing the same shit with fucktard teammates...youre clearly not good enough to win with them cause you need to be with skilled players toeven do smth...conclusion?? you suck and youre not even close to 4,5k player...
and YES, big difference between 4k flat and 4,5k and above
See, Mark is an example of a person who lacks comprehension skills.
When did I ever said I stomped them? I win and lose against 4.4k/4.5k just about as much as against 4k flatters, hence I'm saying there isn't much skill difference.
inb4 another smart alect will come soon.
@dd-sama, I guess so. It might be that I haven't gone against 4.6k/4.7k often enough to see the gap in skill.
4k scrubs throwing my games all the god damn time. if you're below 5k, then ure shit.
case closed.
No wizard you are the one who can't comprehend. If you are less than the ratings of the opposing team it means people or multiple people on your team are higher so the team mmrs even out. Those people could very well be carrying you. If you were actually as good as a 4500 player you should win more than 50% but you aren't good enough to win more than that and rank up yourself.
You are assuming you are as good because you go solo mid vs their solo mid you happens to be a bit higher. But skill can't be judged in a vacuum that way. If you are really a 4500 player in a 4k body and you have others on your team who are 4700+ to even out the team mmrs you should be winning more than half your games and your rating should be going up. But it's not. So there's a flaw in your theory. You are thus not as good as a 4500 player on the whole, you're not doing enough to get your team wins.
Dota isn't about just one player is it. No matter how good you are you need the other four people to coordinate with you and try to win. At any skill level there will be people who are having an off day or get counter picked way too hard to create an actual impact. No matter what role you play there is only so much you can do. Even if you dominate mid while your off lane has given away 4 kills to a PL in 8 mins.
and MMR to say that it measures skill is bs. It just measures the number of games you win irrespective of whether you were the best player in the game or the worst. So yeah the possibility exists that you might be better than several higher ranked players. Create a smurf account and play with a 5 stack, you will start at an extremely high MMR, but is that skill? No, its just a higher winning ratio.
^
I highly doubt that's what he's saying, and for the record, the post above mine is a perfect example of why most of us don't get along.
4000 - 4700 is less of a difference in skill (on average) in comparison to 3300 - 3900
However, I would say 4600ish+ is when the skill difference becomes a little more noticeable.
This is just from personal experience...
I am not such a stupid or new player to the extent that I don't realize whether or not I am or am not being carried as well as my own contributions to the game. What the fuck do you take me for? Especially since you haven't watched replays of the games? Thats mighty a lot of assuming you're doing there. Please shove your mmr e-peen down the drain and read this post while considering me as an equal.
Firstly, I'm mostly going against players roughly 400 mmr above me to which I feel that despite this gap, there is close to 0 skill difference from what I see in them as teammates and opponents when compared to my flat mmr. Secondly, I said when I played against 4.7k players mid (which I only did like thrice on my friends acc), they were nothing 'overwhelming'. It was nothing like when I went against a leaderboard player like Zenoth, they were definitely better than 4k don't get me wrong.
Cuz like I said in one of the above comments, the skill gap from low 4k to high 4k is a lot lower than in low 3k to high 3k. A 4.7k mid can fuck up and lose to a 4k flat mid but the same thing happening for a 3.7k mid to lose to a 3k mid is far more unlikely. This is mainly due to the damn fact that everyone has the mechanical skills when they reach 4k but some people lack the competency to utilize it consistently, as quoted by Havoc.
My entire fucking post is trying to elaborate that if you're within the 4k bracket, regardless of high or low 4k, you're similar in skill to the other 4k players. Much closer than some people would like to think. Similar does NOT mean the same. You can kid yourself that because you're 500/600 mmr higher than another person, you're gonna win 100% of the time. In reality if both players play at their best, its more 60-40 because the main bloody difference in high 4k and low 4k is decision making, item builds and lineups.
I only feel that in 5k is where your mechanical skill becomes really apparent and you realize you're getting outlaned entirely.
Please stop assuming things. Please stop twisting my words. Please stop thinking you know better. Jeez.
4k bracket is the most retarded bracket
I'm 4.7 atm and I've been on a 15 loss streak or something
sometimes you get really decent players and sometimes you can't win even if you are Dendi
4500-5000 bracket is worse than flat 4k trust me
I think a lot of these threads/theories on how mmr correlates with skill is just a result of small sample sizes (ie win/loss streaks, random factors etc).
That being said, not only do you need a large amount of games to see your true mmr, but you also have to play consistently (good,bad, or the same) to get an accurate rating. People who care enough to graph their wins/losses know what I'm talking about. Look up "Law of large numbers". It's simple, in the long-run you will eventually stable out to where you are suppose to be since you will have played enough games to rule out the randomness of the game (win/loss streaks, lag, tilt, drunk teamates, uncomfortable heroes, etc.).
This is no theory, its fact. However, the general misconception here is that players who are truly 5k+ but linger around the 4ks don't want it as bad as others. In other words, they are not willing to play 50-500 games consistently to win more than half their games.
There will always be a game where you feel completely outclassed and remind you how bad you are. These are the games which are most important if you want to actually get better imho. And if you don't have many of these games then you are definitely climbing up the ladder!
TL;DR numbers don't lie, you will get to where you belong in the long-run
"the skill gap from low 4k to high 4k is a lot lower than in low 3k to high 3k"
The average is around 3500 lol, that means there is less skill gap from low 3k to high 3k, and higher skill gaps at the ends of the mmr spectrum (aka the skill gap from low 4k to high 4k is actually larger)
@Vidra Not true, 4k players are really bad at understanding the game, they will try to argue with you their flawed logic and not listen to your commands. Whereas its easier to win the game in 4.5k>>> with much lesser effort of carrying the entire team as your team would mostly know how to roam and stuffs unlike 4k mmr.
the gist of it is
everyone has different experiences through the matchmaking system
yet no one is willing to just admit there's truth to both statements
ive been amazed by 4k players, ive been shocked more than several times, when the MMR scoreboard shows at the end of the game
i've seen too many 4.6k+ players under-perform, and there's a lot of factors to consider... looking at MMR only, is an assumption, unless you're willing to sit through a replay from each person's perspective (10 times) most of the time, you really don't know how your ally went 0-2, or what difficulty they're going through, and when you happen to be in the same position, it's somehow "different". \
which leads me to:
there's too much of a blame game most of the time, and it pisses me off (the difference is, i don't try to provoke the person i want help from) when others get pissed off whenever some ally "supposedly" made a mistake. ironically, the person who's starting a flame fest, is just as susceptible to mistakes and has likely made mistakes already.
"if you have nothing good to say, don't say it" (if you want to win, that is)
nothing really surprises me more than someone who's shocked at the fact that someone could "wrong them", and continue to seek help from the very person they're badgering .___. (i'm not saying i'm against it, but for the love of god, don't expect "better" co-operation).
see you in another life said it quite well, and it's basically a combination of everything + having large sample size of games to obtain a more accurate representation of your skill level/mmr
Talks shit about higher mmr players having low skills
Doesn't win more than 50% of the matches
Obviously the other players don't feel overwhelmingly strong, they don't have 4 hands and 30 fingers.
But they can win consistently to get a better mmr than you do
Burn.
In all honestly the tryhards in 3.3k hell is 2-3x harder than the 3.6k mmr folks. The thing is at 3.6k the draft not team will fail you more often. Where as in 3.3k there is a higher chance 4 scrubs will just lose the game for you. I've enjoyed more 3.6k games than 3.3k games. So I imagine it's the same higher up.
MMR relies on a system designed to pair players 1v1 for every variable you add, i.e. other players, the MMR becomes less useful as an indicator as variables play a larger factor. This is more accurately how the law of large numbers applies to MMR. It is a system which comes from pairing chess players. Part of the problem of MMR is the K value appears to be fixed at the moment after 10 games, in chess you would play maybe 30+ games at a high K value before your K value was reduced to fine tune it. We are stuck on a higher K value which means you see large variance and it compounds the variable effect. For example your MMR is adjusted at the moment by 25, in chess once you had played your 30 games your K value might fall to 10 depending upon whether or not you seem settled in a particular bracket. The only way this could improve is if the MMR system could take on board a larger number of variable per player and slowly reduce your K value depending upon the number of games you have played factoring in things such as KDA / Economy / WL / Relative Success, all of which would require many more than 10 games to establish a true baseline. Things get worse when you consider that players will vary depending upon hero, so really you might need 30 games with every hero before you start to get a true rating and then it could only be an average across all heroes. But compare this with chess where you have fewer variables it becomes easier to see why MMR feels so vague.
Further issues with this arise from people manipulating the system, however the long and short of it is that your MMR feels wildly variable because it is. It is an imperfect system which has too many variables to account for to accurately assess your individual skill so you have games where you are the weak link and games where you feel like you are surrounded by weak links.
Come up with a better system that is tried and tested and perhaps things will improve but without a fair way to compare performance across hero's you will continue to get this variance.
On average, 4500+ players are definitely better than 4-4.4.k players. i can tell from skill difference even before seeing their mmr at end of game
Skill wise, I don't think theres much of a difference. I think in terms of general farming and game sense, there is a noticeable difference.
I haven't been playing ranked so much lately. But at the 2.9-3k bracket, all I can say is that shit varies a lot. You get people who do some amazing things and others who feel like they are playing for the first time after the tutorial missions.
What I'm trying to say is, stop looking at the MMR and try to enjoy the game. I sometimes play stacked vs people who are 700 MMR over me and stomp them, and then I play with people who are like 100 below and get stomped.
MMR is too inprecise to stamp a number on someone's head and say "that's what you're worth" because there are too many factors: team draft, skill on each individual hero, skill for each position (1 to 5), game knowledge, in game execution, decision making, proper skill/item build, etc. You can see some people with 3k MMR play certain heroes much better than other 4k MMR players.
As someone previously said, to get a proper MMR, you'd need to have every player play each hero 30 times to even get close to his actual skill level.
Пријавите се да бисте поставили коментар.
I've been duo stacking more often than solo queueing now with my currently 4.7k friend.
In short, I'm always playing against players 300-400 MMR above me. I almost always take core roles; mid or carry. My friend goes mid when I don't otherwise he goes support (so no, he doesn't 'carry' me in these games) and my verdict is that there is no obvious difference in skill.
I've gone solo queue on his account before for a few games but eh the enemy mids didn't exactly feel overwhelming, in fact some played like idiots. Like the only main difference is a higher constancy of proper lineups. Trilanes/Offlanes become more frequent and you see less of 3 carries.
Or am I just really lucky over my past 30 games that I've gone against scrubs? I know there is a huge, huge difference between 3.3k to 4k but that same MMR gap of 700 between 4k to 4.7k has a much smaller gap in skill.