General Discussion

General DiscussionMMR INFLATION

MMR INFLATION in General Discussion
i commend

    Like our economy, prices tend to rise in different rates. Since the introduction of the ranked matchmaking, the norm has been to be atleast 2500 mmr to be considered knowledgeable at dota. 4000 and above used to comprise 5% of the top players in the world. Is it still the case now?

    I believe that 5K is the new 4K in terms of skill cap. As the year passed, players regressed and progressed from their initial MMR so it has even out the playing field. People went to their true brackets but account boosters and smurfs ruined the system by messing the system.

    As of 2017.
    What are your opinions on this? What is the current middle bracket? What MMR are you considered to be borderline "pro" at dota? What bracket is the trench? What bracket is the ELO hell? Is VHS still really good?

    Garruk

      Dunno if this is legit, but looks good.
      https://dota.rgp.io/mmr/

      论坛展区

        Blood Blood Boooooooooooooooooold

        1-IceTea 🌟

          Base on

          The full dataset doesn't solely track Ranked players, but as many players as possible. Only a small fraction of all Dota 2 players play Ranked, roughly in 1 in 3 or less.

          And lower player more likely to never show their mmr on profile (raj for example) middle should be 2.8K which is much higher compair to past

          Jacked

            No point guessing as there's no data. EstimAtions are inaccurate due to reasons pointed out by licetea.

            It is likely that the average hasn't moved. Although number of higher rated players have increased due to more players

            TripleSteal-

              mmr inflation is <1% and there s no significant reason for it to exist either

              Clippy

                Inflation can't really be a thing since you can't print MMR. What's happened is global specialization in specific roles in Dota. This has increased the amount of inputs of high skill players and therefore allowed the MMR cap to rise to heights never seen before.

                Jacked

                  Technically you can print mmr when monkeys like blunt keep losing games at 1 mmr

                  i have 5 reports to use

                    If the data follows a normal distribution then it should be a good estimator of the population data. And as I can see from the data 4K is no longer 1% as expected. Obviously the mean has shifted higher and there is "mmr inflation"

                    This comment was edited
                    Jacked

                      Only if the mean has actually shifted. And there's no data for this