mmr inflation is so low that u can assume it doesnt exist at all without being wrong
u just suck
^ oh come on, disappointment all around. I was always beating around low 4k, then ppl said it was shit, now I am beating around 5k and ppl say it's gone to shit too. So you mean to tell me if i ever get a blue star I won't be able to act all high and mighty ? :(
People in 3k are 50th percentile on opendot still. You are in top 10%of players at 4k, including account buyers.
actually it's top 20% over 4k, and top 10% over 4.3 k
Ok, I stand corrected, you are in top 10% at 4300. Still, for a 5k -6k player I might as well just have changed the spelling.
anyone care to share an idea on why the sudden spike at 4k mmr ?
seems like an outlier.
thx
people stop playing when they hit x000 because they know they are garbage
similar thing happens at 5k and 6k if u look
@teeth (JDF8) so basically tou reach XXXX K mmr and your like " OK got to my goal, never playing again"?
hmm seems strange but does explain it. I guess if you hit 4k mmr after 3 years of playing, and are scared of dropping back to 3k that makes sense.
Thanks mate
People are like "wow 5k is trash haha" when in fact a majority of dota players r 2ishk. If ur 5k ur still a massive huge minority and leagues above ur average dota player
but who cares about being better than the average player
the average player also probably does not play so much dota lol. or if they do they just do not spend much brainpower trying to learn
Because everything is relative, if u have lets say scientists of smartness like einstein or hawking or one of those absolute geniuses, ranging to dr. fucking oz, just cuz einstein is above u, doesnt make u a stupid scientist. Cuz turns out there r hundreds of dr. Oz's, all more retarded than u. Its R E L A T I V E.
top 1% (not 10%) cann't be higher than 4k6 (already very unlikely)
It's much more likely to not exceed 4k4 by now.
For top 1% being 4k6 (top 10% being 3k8 then) you already need :
• 1 million new smurf account that calibrated at 5k (which is mostlikely false)
+
• More than half of the player than joined since 2013 being better than 60% of the player base of 2013 (which is obviously false)
I did the calculation quite a long time ago, but if you were to remove the high mmr smurf account, players that joined since 2013 would need (for half of them) to be better than ~95(99?)% of the players of base of 2013
.
.
PS: if everysingle account were calibrated at 5k, (since valve's number got released) top 10% would be ~4400 mmr, and top 1% would be 5200 mmr -assuming no1 abandonned dota2-
The numbers Valve released were for unranked, I don't know why people assume ranked has the same distribution, especially considering how trash unranked mmr is in general.
Just taking a brief look at leaderboards, top was 6k on release, top is now 9k, bottom was 4.9k on release, bottom is now 6.4k. Clearly some inflation has occurred, more inflation the higher you go, don't know how much it's affecting 4k bracket but 5k/6k is definitely inflated by a bit.
unranked and ranked work exactly the same, except you don't see the number. The calibration doesn't change anything, except when you try to abuse it.
+ are you really trying to describe a bell curve, with the least significant part of it ? (it's like trying to describe earth, by looking only at the leaves of trees, and nothing else)
top 1% was 78000 players, you are going to the top 200(800~). The increase of player, and more time to play, can explain (are explaing?) alone the "dispersion/inflation" of the very top.
Top 1% was not and is still not even close to the leadboard.
Last but not least, you anyway cann't get any inflation without the new people joining being better than the one that are already there (and smurf account/fresh new account to be sold). Even having 1 million of such account cann't even make a difference big enough.
.
.
Literally, the only way for "such" an inflation to happen to somwhere else than the very very top is:
The population of dota2 is no longer following a gaussian curve (which would be a premiere in the world). Literally everything we can observe is following a gaussian curve when the sample is big enough. (and dota population is big enough)
The bell curve is just spreading out as people make more and more fresh accounts, clearly the top 200 are inflated, it's not ridiculous to assume that ranks 201 downwards are as well. More inflation the higher mmr you go, 4k probably isn't that affected but 5k is a bit and 6k certainly is. I got on leaderboards with 5700 about a year ago now I'm off the bottom with 6556 without any significant change in relative skill. Note that bigger servers (EU) are more inflated than smaller servers (unless you seriously think that Midone improved 700 points of mmr just by moving to EU, or Ana got 2000 points from the same move).
Unranked obviously works on an altered system to ranked, if you've ever played it at all you'll notice that you get way more trashcan games in general. If I were to tryhard unranked I could easily get 75% winrate permanently (in fact I have 75% winrate last 6 months mainly randoming in 5 mans with real life 4k mates) which would never happen in ranked as I would raise my mmr to the point where I would get harder games and my winrate would drop. If you ever played Hon they had regular PSR compression in order to keep mmr uninflated, I hope they do something similar in Dota.
If you look at other games theres inflation going on, idk what causes it it's probably just a problem with Elo based systems. For example in chess inflation is an issue, Dota works on a very similar system and you'd have to be deluded if you think the same thing doesn't happen here. On release 6k was fucking incredible now shitstains like me can get 6.5k spamming moron carries with 2 buttons and the micro ability of a one handed child, if you don't think that's evidence of inflation I don't know what is.
http://en.chessbase.com/post/rating-inflation-its-causes-and-poible-cures
@SS.CUTNPASTE wow I actually agree with you sir. Back in my days 5k were pro level but now all my friends were getting 5k mmr's like 3k before. Oh well looks like im still bad at dota. 6k in 2017 = 5k in 2013 = MMR INFLATION IS REAL = I DIDNT IMPROVE
You are all wrong. Inflation from newer players (population of the left tail of distribution) is different from the inflation that occurs when the entire curve becomes populated.
A huge and continuous increase in skill of current player base + substantial influx of new players + permanent reduction in skill floor of the game = mmr inflation across all brackets. I.e 6k = 5k in the past, 3k = 2k in the past.
This is Assuming all existingplayers generally increase skill at same rate and new players increase skill at a slower rate.
In other words, the only way this type of inflation persists is if game knowledge is highly entrenched within time. Meaning an existing player will always maintain an advantage in terms of skill over a new player because he has played longer. Or because of lowering of skill floor so more retards can suddenly play (when before they couldn't). This is very likely an incorrect assumption.
So what actually happens is after new players slightly inflate mmr, the average of the curve returns to where it was. They learn at same rate as existing players. They learn the game, approach their true skill over the long run. Existing players don't improve indefinitely. New players can catch up. These are all more reasonable assumptions.
But the right tail of the distribution is pushed out. This is why more populated servers have higher mmrs.
Inflation of right tail is observed. But that does not mean the average has inflated. 4k could well be 10% still. Although ppl who were 6k in the past could be 8k now. However their percentile rating would likely not have changed.
mmr allways have problems, is like skill rating in overwatch or also IQ in real life
Is hard to determine an skill number for humans ..
Чтобы оставить комментарий, пожалуйста, войдите в систему.
is mmr inflation real