What's my real MMR?

Hey guys, I would like you to try my new custom game.

Mod consists of 5 different challenges and according to your performance it will calculate your MMR.

Workshop page: http://steamcommunity.com/sharedfiles/filedetails/?id=1134432887

Obviously, the hardest part of mod making is to tweak that MMR evaluation, which I had calibrated only with my very own attempts, so if anyone from any skill bracket would like to participate and help, just send me your endgame screenshots. Thanks!

I am opened for any suggestions and criticism.


  • edited September 2017 Posts: 273

    Your game is just measuring some mechanical skills, not MMR.

    Being good at lasthitting doesn't make you good at the game (there are supports and roamers that don't lasthit at all. I suppose those things are non-existent in your MMR bracket). Lasthitting under tower is not a measure of anything. Even the best pros can't lasthit everything under tower.

    Creep blocking cannot be done with every hero (example CM, techies because they are too slow) so that's not a measure of MMR. Creep pulling, stacking and contesting enemy pulling and stacking IS the measure of MMR.

    Invoking spells is not a measure of MMR, its a measure of reflexes, speed and muscle memory. You need 1000 games of Invoker to be perfect at invoking spells. I have seen Invokers in every skill level (from 2k to 8k MMR) varying from 10 APM to 500 APM. Invoker's player APM and skill (MMR?) doesn't mean anything if you have BKB and a coordinated team. If your Invoker is Miracle, you can split push him to victory and completely ignore him or keep the pressure on his lane from the start of the game. SO, invoking spells is not a measure of MMR as well.

    Skillshots are a measure of MMR? Maybe, maybe not.

    Microskills are more than just controlling multiple units (RTS aspect). Microskills include, as far as I know, managing mana, cooldowns and usage of all spells and items. Did you include testing for all that? Let me ask you this too: Are Starcraft proffesional players good at Dota? The answer is: Depends on the hero and player's knowledge of the game. You can confirm that fact by seeing players like Day9, SingSing, Korean players etc.

    TLDR: Your MMR evaluation is shit. And you can't measure MMR without a team, because, you know, Dota is a team game.

    Suggestion: Include a quiz with questions about the game (from easy to very hard). And things like aggro manipulation to measure MMR.

  • Thanks for your reply. I am not so dumb, it is clear that you can not measure real mmr like this, without teamfights or objectives. My mod was mainly made for motivation/challenge/meme. But all of used techniques are definitely important in Dota2. Also it was made for low-mid ranked players, def. not for 7k+ (yet).

    100% there will be difference in performance between 2k and 5k player in lasthiting, even for support abusers. Also game itself includes a lot of randomness i.e. lasthiting under the tower is weightened much less than at lane. There is also a big portion of randomness(units damage), not everything goes perfect in every game.

    Creep block is questionable.

    Voker is one of the most popular heroes, so I have used him, everyone must have played some games with him as you said 2k to 8k. Reflexes, speed and muscle memory? That's exactly what I was aiming for with that challenge.

    About micro I am the least happy, but have not found any better solution yet.

    Quiz is one of the other things I was considering to add but with no success.

    I am looking forward to implement code, which would send me results to analyse and make evaluation better and then maybe some kind of leaderboards.