Genuinely believe this is the last thing Musty needs to see right now
I'll gift him candy crush and hello kitty island.
*** Official Dota 2 Thread ***
- Thread starter Purdy
- Start date
I'll gift him candy crush and hello kitty island.
Now as Im only 3 games away from getting my MMR I think I might just go play candy crush now and be done with it
I'll join you.
Soldato
- Joined
- 29 Aug 2010
- Posts
- 8,267
- Location
- Cornwall
I don't think MMR should be based on if you win or lose, well not entirely. I think your KDA should have an impact also. Maybe it could be used as a multiplier.
So for a win: 25 x KDA = MMR increase.
For a loss: 25 / KDA = MMR decrease.
Of course I'm very new, so what do I know.
So for a win: 25 x KDA = MMR increase.
For a loss: 25 / KDA = MMR decrease.
Of course I'm very new, so what do I know.
I don't really know how accurate these kinds of things are tbh. I've seen multiple post and pro interviews where people state that regional differences in MMR can be quite large.
I would agree that a 6-7k player is pro in any region but 1-4k can probably vary a lot. Quite a few of the European Pros claim that when they play MM games on US servers they get matched with 4k players that play like EU 3k players.
Don't know how true any of this is but its been stated by multiple people in multiple places.
So supports should get less MMR than carries? How is that even remotely fair?
His is not the perfect example but I have often thought with a bit of thought and work mmr could be personal as well as team based. It's a fallacy to suggest that your personal mmr is determined purely on a team performance which takes it out of your hands to an extent.
Every role could have certain conditions I.e. A carry how many kills you achieve and GPM/XPM (as an example) and say in a match you can win lose 30mmr. 15 based on a win/loss and 15 on your own performance. Obviously still effects by team performance but there should be a neater solution than team loses you suffer especially if you play well.
Every role could have certain conditions I.e. A carry how many kills you achieve and GPM/XPM (as an example) and say in a match you can win lose 30mmr. 15 based on a win/loss and 15 on your own performance. Obviously still effects by team performance but there should be a neater solution than team loses you suffer especially if you play well.
Caporegime
- Joined
- 9 Mar 2006
- Posts
- 58,026
- Location
- Surrey
The simple fact is there is no way to assign a number that represents your skill level in a complicated team based game like dota. It was a huge flipping mistake making this visible in the first place.
The only way a ranking system can work is a team ranking for a fixed team of 5 players.
The only way a ranking system can work is a team ranking for a fixed team of 5 players.
You would need completely different MMR ratings per role played. Otherwise it would not be fair to judge a support player on hero dmg etc when that player is the ward bitch and not expected to deal massive dmg throughout the game.
To be fair at least he doesn't hide his mmr and pretend the game 'automatically' hides it
Soldato
- Joined
- 29 Aug 2010
- Posts
- 8,267
- Location
- Cornwall
But if we all played together all the time we'd all have the same party mmr but I don't think I'm anywhere near as good as you guys. But my mmr would say I was even if I was feeding most games (unintentionally).
As for supports getting lower scores, that's where my lack of experience comes in as I had assumed that supports would gets decent assists.
Hello can you stop replying when we are ingame please.
To be fair at least he doesn't hide his mmr and pretend the game 'automatically' hides it
Dude cmon gotta wait for the grand unveiling of my return to 4.2k
I'm pretty sure during calibration it takes your "unranked MMR" then your calibration matches have 3x as much weighting as normal matches. i.e. +/-75ish MMR. Pretty sure I found an official thing that stated this, can't be bothered looking for it now though.