I didn't read this post but knowing mmr before a match starts will lead to early flaming and could result in losing a game you could have won if people didn't attack each other. I honestly think the biggest difference in my W-L is the mentality of my team. It seems everytime my team has an asshole we lose. When the team supports each other we win. It may not be a fact. But I believe heavily in chemistry. The moment your team starts shitting on each other you have already lost.
I win alot and have alot of bad teammates too. Learn to adapt to bad players. If you can't adapt to bad players, then you don't deserve to win.
I find the current MMR to be nothing more than a rough guide to consistency. I've seen 3800 players outperform 4500 ones more than a few times already, so I'm inclined to believe that showing the MMR before a match isn't really going to achieve anything but fuel flame wars. Heck, I've not bothered looking at the MMRs after a match after the day the patch came out, so I'm not seeing any incentive for keeping it around anyways.
@listen to me
The carry (ember) and mid (necro) were 500-600 MMR lower than everyone else. Everyone else was about the same.
@admiral love handle
You have a lot to learn then, because I bring up a very good point.
@barack brobama
Winning with "bad players" on your team amounts to bullying your way onto important roles just like they do and asking nicely for wards at certain spots. That's not the problem. Read what I wrote - the problem is when they unknowingly bully themselves onto important roles, because they don't realize they are 100s or 1000s of MMR beneath other players on the team. Once they have chosen to go mid, there is nothing much that can be done. If you read the post, you will find a story in there that I encountered (I can link the match) where literally a 4.1k MMR player called mid and went toe-to-toe with a 6.2k MMR player. That is just absurd. He obviously would have let the 5.7K MMR player on our team handle that spot had he known the true discrepancies in skill.
http://dotabuff.com/matches/567451228
@cartographer
Nice, dude, but if you look a little further down, I'm not talking about when players are near in MMR. I have seen some huge discrepencies, and any player in their right mind would not WANT to go mid in those situations. Do you think the 4100 MMR player that got 0-10ed or something by the 6.2k MMR Whiz in my game would want to repeat the experience? He chose mid under the assumption that he was the best in the game/around the same as everyone else. Had he known the truth, he would have conceded mid to the 5.7K MMR player.
http://dotabuff.com/matches/567451228
"Winning with "bad players" on your team amounts to bullying your way onto important roles just like they do and asking nicely for wards at certain spots. "
Wrong. I play support 24/7 and I will support even the worst players and through support play I can will a team to victory. Support is the the most important role on a team. Most people cant play support. Buying wards and courier isn't "supporting"...That is why a good support can make any1 better. If you think going mid or playing a carry will help "bad" players....you need to learn dota bro.
Haha, I thought I can do so also, supporting when I had a high mmr. I was wrong. I lost 9/10 calibration matches were about 7 of them were won and our pro carries threw them
@admiral
I think the problem here is your definition of "noob". You probably mean a teammate that has 300 MMR less than you or something. Take a look at that game where Storm went 17-2. He had 6200 MMR versus a guy with 4100 MMR. That is absurd. You can't will shit when your mid feeds a professional player. We desperately needed our 5700 MMR player to go toe-to-toe with him in mid. There is no argument here, it's just fact.
its actually really frustrating recently i played a Captains draft and i drafted an alch for my team, and STATED were gonna run it as a ganking\mid alch so you know thats grabbing medallion maxing stun and spray... but this kid slam picked alch as soon as the draft was over and went mid which was right and went gloves of haste tango level 1 goblin green level 2 spray level 3 greed 4 stun 5 greed.
thanks alch, preciate the free loss asshole didnt even get a fast midas took you 7 minutes. LOSER
\rage lol.
there needs to be some kind of way to tell whos gonna be the loose cannon i agree
Also here is some nice sample data. The Drow+ Venge laned against the Mirana, raping both of them hard.
Also love handle, supports not the most important role on the team, every role is pretty much equally important.
but if you wanna say there is a most important role its probably the mid hero, but i dont think it matters anyway
so how long did you have to stack with your wisp buddy picking korean in order to break 5k in solo queue?
alwayswannasuck, thinking isn't your lot in life. You should focus on manual labor or going pro in dota.
Please sign in to post comments.
http://dotabuff.com/matches/570475854
This sort of thing happens all the time. a match up will have 2 players at some level of play (call it X mmr) and 3 other players that are 400 or 500 MMR below X. These players pick carry and mid. The other carry and mid are manned by the other 2 players with X MMR meant to balance off your two. We lose mid. We don't have late game in our carry since he isn't skilled enough even to farm correctly. There are other similar configurations (1 player on each team is X, 1 player on one team is really good with 4 mediums and the other team has 2 goods with 3 mediums, etc.) but consider just the 2 X and 3 X-400 setup like happened in that game.
There are these possibilities (listing which mid/carry role was manned by player X):
mid,carry vs. none
mid,carry vs mid
mid,carry vs mid,carry
mid,carry vs carry
mid vs. none
mid vs mid
mid vs carry
carry vs. none
carry vs carry
none vs. none
A statistical way to state there is a problem is showing that some of these categories (even though they are rather imprecise not accounting for jungling, offlane, etc. [though you could broaden the study as much as you want]) have very different win rates from 50%. We need the knowledge to assign our best players to certain roles, e.g. always making mid,carry vs mid,carry on average (and hence almost surely being 50/50 chances in the average). If we were to rank these categories in terms of |win chance of one team - 50|, we could perhaps predict something like this:
Has a value of 0 (fair in the average):
none vs. none
mid vs mid
carry vs carry
mid,carry vs mid,carry
First being the highest value (furthest from 0, most imbalanced):
mid,carry vs. none
mid,carry vs carry
mid vs. none
mid vs carry
mid,carry vs mid
carry vs. none
These X-400 players are not malicious, and most would gladly concede important roles if only they knew. But of course, sometimes they're with 4 other people that are X-400, sometimes THEY ARE the X player, etc. They simply don't know. We need to inform players in the beginning of the match so that on average, the buckets closer to 50/50 are chosen by the players themselves using their own reasoning. I personally concede any role to a player with 6k MMR. I respect him, and I want the best chance to win.
A BIT OF FURTHER READING, ADDRESSING COUNTER ARGUMENTS
From what I can see, there are two major counterarguments to this system.
1.) What if players leave when they have players on their team they perceive as bad?
A: We would implement it to where the MMR doesn't show until everyone loads fully into the game. If someone leaves and abandons after this, he gets -MMR to his rating plus an abandon/loss, and the game doesn't count for anyone (except the leaver). After a certain point, as is done now, of course the match will always count for everyone (after FB or however they do it). Any rational player, even if matched with a bunch of horrible teammates, should decide to stick around for what he perceives as a small chance of winning rather than a guaranteed chance of losing (since his leave will penalize him with -mmr and a loss). Furthermore, with his MMR being known from the get-go, the stigma against playing with lower-skilled players should lessen. I predict the lower-skilled players will become humbler and support the better players (if they are not dicks), making a better experience for everyone. Right now some of that bad stigma is an X-400 player accidentally taking mid and getting demolished by an X player on the other team. Why would ANYONE want that (including the X-400 player)?
2.) What if lower MMR players are always "forced" to play this role or that role when they are better at ___?
A: First, don't fret, thinking you will 100% of the time be forced to play any role. Sometimes, no matter how low your MMR is (unless it is near the lowest), you will be player X or you will be with a bunch of people of just about even MMR, and so on. In other words, no one player will always be expected to play support or whatever. Second, we are humans that harness experience, so the problem resolves itself. If you claim he is better at some certain role, by being forced to play other roles, that differential will equalize over time. It's not a big deal. Third, and most importantly, we WANT them to be forced into those roles to make more consistent gaming experiences. This last point is important to deserve its own paragraph.
As it is now, since the categories are symmetric, statistics hide the problem. For example, suppose carry,mid versus none has a win rate for one team of 80% on average. Well, 50% of the time, we can say almost surely, team A will have those odds. And 50% of the time, team B will have them out of all games that have that setup. In aggregate statistics, this is seen as a perfectly fine 50/50 odds for both teams. But the actual experience for players in those games in those specific distributions of players is absolute misery. Playing an 80-20 configuration through sheer bad luck is the last thing anyone wants. The X players get pissed that their important roles are playing the wrong way, and the X-400 players feel bad that they were crushed so hard (especially in mid, carry players tend to be delusional and blame everyone even if they just suck, but they still have a bad time since playing the blame game is very far from fun).
A commonly known thing: That people play better and worse on certain heroes and in certain situations. There's nothing Valve can do about that - they must deal with long-term averages, not finely detailed hero/role performances, and in fact, that's exactly what they do, will continue to do, and will never stop doing. My suggestion is the most reasonable way for them to keep doing what they must pragmatically do (deal with averages) while also giving us heuristics that alleviate much of the distress caused by that necessary, statistical way of judging skill.
A final argument that blows anything you say out of the water is that people can still choose whatever they want to choose. If you truly think at 4100 MMR you deserve to play solo mid more than the 5700 MMR player (I have seen this happen in the MMR end screen before, and the 4100 MMR guy was against Whiz, someone with 6200 MMR, who ended up going 17-2 while the other guy went 0-10) then you are simply delusional. In that example, he actually wasn't delusional since he wasn't armed with that information. But hey, you can do it if you want. And THAT is the main idea here - playing the crucial roles in situations where there is tremendous MMR disparity is actually stressful and not enjoyable for EVERYONE involved, including the low MMR individual who unknowingly just matched himself in a head-to-head against someone 1500 MMR higher than himself. No one in his right mind who put that kind of pressure on himself. It's not fun to be demolished 0-10.
Everyone in this thread might argue against me thinking something up like this:
4100
4213
4117
4050
4000
will have ironclad mid/carry positions demanded due to MMR. That is NOT what I'm talking about. I'm literally talking about MMR lineups like this:
4800
4300
4300
4800
5800
Why on earth would the 4300 guy go mid? He will be, 9/10 times, against the 5800 MMR guy on the other team if he chooses mid (and it wouldn't even be his fault - he has no idea someone on his team has 1500 MMR more than himself). He will feed, be underleveled, feel like he let his team down, etc. And it really wouldn't be anyone's fault other than Valve's for not arming people with the knowledge to make better decisions.
The best compromise is AT LEAST if there are tremendous discrepancies, MMR should be known upfront in the match so players in that particular situation can make intelligent decisions. If you don't agree with that then I just have to think you are trolling or don't know better.