|
Please keep the QQ to a minimum if you do not like this update. We are happy to hear your reasoning for not liking a ranked system, but no "OMG VOLVO WHY" posts. |
On January 07 2014 06:43 oo_Wonderful_oo wrote:
Will ask question here just for lulz.
If anyone knows probably, Maelk finished his placements 1-9 and got 5110.
I know people, who finished as 9-1 and got 4200.
Where is logic, where do we have to find it?
? You start at your old hidden mmr so maelk probably had ~5300 and 9-1 "people" probably had 4000 mmr before ranked came out.
|
The land of freedom23126 Posts
And how do you count this hidden mmr?
Winrate? They have 57%, i don't think it's very little in comparison to Maelk's 63.
Other stuff?
Is there any formula to count mmr or it's still "roll the dice" stuff like in all MOBAs?
|
On January 07 2014 06:45 Kuroeeah wrote:Show nested quote +On January 07 2014 06:43 oo_Wonderful_oo wrote:
Will ask question here just for lulz.
If anyone knows probably, Maelk finished his placements 1-9 and got 5110.
I know people, who finished as 9-1 and got 4200.
Where is logic, where do we have to find it? placements dont really matter, it's based on your unranked mmr before the calibration matches
They do actually I lost and got 90 points in calibration matches , if u win most of them u can increase ur rating by ALOT.
|
I've got 4100 solo and 4700 party, I only play withy brothers (3700, 4600, 3200 - I have 3
Bros :p) we never actually play all together it's usually just a 2 stack.
I started on 4300 party and 4200 solo. The fact that that we win way more party games over solo I feel is simply because I play mid and very rarely lose (either equal or ahead or I get ganked 4 times in 4 mins) leaving the other lane to one of my bros and a random. We talk a lot of shit on the mic and tell people what to do and they listen, and we win games because of it.
*3200 rating bro is an exception he's very n00b :p
However it seems when I play with my 4600 bro the players are not much better than the ones I verse at 4100 solo. It could be the AU servers not having a large enough player base to match close mmrs at all time tho.
Dis just my experience with it
Edit: I feel as though it would be near impossible to get to 5k rating on au
servers from 4k without picking earth spirit every game
|
On January 07 2014 06:49 oo_Wonderful_oo wrote:
And how do you count this hidden mmr?
Winrate? They have 57%, i don't think it's very little in comparison to Maelk's 63.
Other stuff?
Is there any formula to count mmr or it's still "roll the dice" stuff like in all MOBAs?
It is not counted. You always had an mmr in Dota2. It was just not shown.
|
On January 07 2014 06:49 oo_Wonderful_oo wrote:
And how do you count this hidden mmr?
Winrate? They have 57%, i don't think it's very little in comparison to Maelk's 63.
Other stuff?
Is there any formula to count mmr or it's still "roll the dice" stuff like in all MOBAs?
Its the same as the current mmr system only they didn't show you.
|
The land of freedom23126 Posts
LOL
So i'm now curious what formula does current mmr system have :D
Now it seems like it's even worse than League of Legends's one for example.
|
no they are identical lol. There is no "calculation" involved.
Win -> go up in mmr
Lose -> go down.
The amount you go up and down is based on the calculated probability of your team winning based on the MMRs of each side. The matchmaker attempts to make the probability of victory 50-50 but especially at higher rankings i'm sure 60-40 or worse isn't uncommon.
The system isn't perfect but it is in fact conceptually best-possible. The implementation may have its pitfalls but valve has a LOT of data and they seem to be pretty pleased with the outcomes.
|
The land of freedom23126 Posts
Arguable against best possible but ok.
I still can't stand two things then.
First, when did mmr start to count, to apply that hidden elo.
Second, on what points is it based because logic points is winrate (but it's not working seems like), performance (it can not be counted), some stack/solo play (not working as well, same to League of Legends).
|
MMR has always existed since the "find match" button was a thing.
You start at average, and if your team loses when the system says you are average, it bumps you down to "below average". Depending on how certain the system is regarding how good you are, it moves you more or less every time you play a game.
Winrate is irrelevant, as it's possible to have a sub-50% winrate and play with pros consistently. Not terribly likely, admittedly, but possible. The thing about a winrate is that unless the system CANNOT find you equal-skill opponents, you will eventually average very close to a 50% winrate simply because thats what happens when you are exactly as good as your opponents. If you are better than your opponents, you will win and the game bumps up your rating and makes you play harder opponents until you are playing against people of your skill. Whereupon you will win half your games.
Individual performance is not measurable, so they measure team performance (aka victory). If you lose because your teammate sucks, theres a reasonable chance that the matchmaker assumed your teammate would suck and expected you to play better. Next game you might win because your enemy sucked unexpectedly. It all balances out.
|
The land of freedom23126 Posts
It still shows that matchmaking rating is nothing basically.
Best mmr system was in pre-s3 lol i feel.
When you started at 1200 and just play for elo points. You're winning? Get more and more. Losing? Lose more and more. No stupid divisions, nothing unclear.
|
Thats exactly what dota does...
+ Show Spoiler +Actually at the very start in the earlier beta days you could answer a questionnaire and it would give you more or less mmr, but that balances out. Also it was always hidden, but that didn't change what was happening.
|
The land of freedom23126 Posts
No, you hadn't understood my point then :D
We don't know WHERE did we start, WHEN did we start and HOW MANY points were we winning/losing.
So, i hope you will get my point now.
If EVERYONE started at ONE place in ONE day, based on some levels just to not throw 20 lvl with 200 in one game like it's happening nowdays, it could be WAY better that it is right now.
|
On January 07 2014 07:26 oo_Wonderful_oo wrote:
No, you hadn't understood my point then :D
We don't know WHERE did we start, WHEN did we start and HOW MANY points were we winning/losing.
So, i hope you will get my point now.
If EVERYONE started at ONE place in ONE day, based on some levels just to not throw 20 lvl with 200 in one game like it's happening nowdays, it could be WAY better that it is right now.
There's absolutely no reason why that would be the case. Starting at a more inacurate point doesn't not result in a more acurate MMR. I have no idea why so many people think that.
|
how would it be better ? Good players would just have to trash 10games in a row before having the same mmr as before and bad players will get trashed 30games in a row before having the same mmr as before. More whine for nothing
|
@oo guy
There's only so many times I can explain it before questioning your intelligence becomes inevitable.
Since that point was reached a while ago, I'm left with... wishing you all the best. I ain't replying anymore.
|
United States47024 Posts
On January 07 2014 07:26 oo_Wonderful_oo wrote:
If EVERYONE started at ONE place in ONE day, based on some levels just to not throw 20 lvl with 200 in one game like it's happening nowdays, it could be WAY better that it is right now.
LoL never did that either, though. There hasn't ever been a hard reset. Even before the division system, your Elo at the start of a season was based on your past performance, with some compression for the purpose of curbing Elo inflation. People didn't start on the same day either.
|
The land of freedom23126 Posts
Guess, discussions about mmr are always same.
Pathetic and boring like whole matchmaking rating system.
Hope, you're gonna have good day guys (:
|
The land of freedom23126 Posts
On January 07 2014 07:37 TheYango wrote:Show nested quote +On January 07 2014 07:26 oo_Wonderful_oo wrote:
If EVERYONE started at ONE place in ONE day, based on some levels just to not throw 20 lvl with 200 in one game like it's happening nowdays, it could be WAY better that it is right now.
LoL never did that either, though. There hasn't been a hard reset since season 1. Even before the division system, your Elo at the start of a season was based on your past performance, with some compression for the purpose of curbing Elo inflation. People didn't start on the same day either.
It was hard compression.
All bronzies like my friends were getting ~1200, and highs were dropped to 1600~ iirc.
I was 1700 and 2300 after s1 and s2, and was still dropped to 1500 and got into plat 2 after placements.
But there was nothing to start.
Basically, except some hidden thing.
But ok, it's useless to speak about it, all will have their opinions like always and it's great.
|
United States47024 Posts
On January 07 2014 07:43 oo_Wonderful_oo wrote:Show nested quote +On January 07 2014 07:37 TheYango wrote:On January 07 2014 07:26 oo_Wonderful_oo wrote:
If EVERYONE started at ONE place in ONE day, based on some levels just to not throw 20 lvl with 200 in one game like it's happening nowdays, it could be WAY better that it is right now.
LoL never did that either, though. There hasn't been a hard reset since season 1. Even before the division system, your Elo at the start of a season was based on your past performance, with some compression for the purpose of curbing Elo inflation. People didn't start on the same day either. It was hard compression. All bronzies like my friends were getting ~1200, and highs were dropped to 1600~ iirc. I was 1700 and 2300 after s1 and s2, and was still dropped to 1500 and got into plat 2 after placements. But there was nothing to start. Basically, except some hidden thing. But ok, it's useless to speak about it, all will have their opinions like always and it's great.
If you compress people down by the same factor, it doesn't serve as a reset because peoples' relative positions on the ladder are the same. It only serves to curb Elo inflation, not to act as any form of functionally useful "reset".
|
|
|
|
|
|
EPT
