No I really don't know.
I assume it is on a different scale than normal queuing. I'm talking about team mmr, not party.
I have no idea if 4.3 is good or bad.
Compare to:
http://www.dotabuff.com/teams/1148284
5418 MMR => TI qualifying pro team MVP Phoenix
The highest Team MMRs I recall were about 5.8k for the murs/beesa/supercell/mason team but I have not been paying close attention to it for a year and I know these guys have gone well past 6k and some past 7k by now.
Most people at top MMRs do not use the Team matching because it is not used much (circular problem) so you have to wait a long time and then you get matched to a team very far from you in MMR.
Anyway your Team MMR starts with your party MMR, but after 3 games of calibration it locks in and has to go up and down as a separate MMR score only used when you play Team Matching with that exact team. You have a different Team MMR for each team.
http://www.dotabuff.com/teams
Because the function is unpopular, there are only 19 teams at 5k+ with active scores on dotabuff.
There are 200 teams above 4.3k MMR. There are 6093 active Teams (a year and half ago there were over 160,000 teams). Unfortunately Valve did not make this feature easy enough to use so people have largely dropped it. So 4.3k MMR is currently a top 3% score.
Overall Team MMR is skewed toward the better players who play a lot of dota, although the very top players don't use it much - and really now all players don't use it much although when Team Matching was introduced about 40% of the playerbase at the time tried it.
Current Team MMR distribution:
99% below 4669 MMR
95% below 4141 MMR
90% below 3818 MMR
80% below 3451 MMR
70% below 3185 MMR
60% below 2977 MMR
50% below 2784 MMR
40% below 2591 MMR
30% below 2354 MMR
20% below 2105 MMR
10% below 1524 MMR
Overall I see that the distribution of scores has drifted upward about 500 MMR points since a year ago as the vast bulk of lower MMR players stopped using Team Matching.
Yes, it starts with the average of party MMR then adds some factor that Valve has not explained well, then you calibrate for 3 games where I guess you can move a lot like during the anti-smurf for new accounts. After that the team acquires a separate MMR for those exact players on that Team. You could even make a new Team with the same players and you have to re-calibrate. Or if you change one player you have to re-calibrate.
"Most people at top MMRs do not use the Team matching because it is not used much (circular problem) so you have to wait a long time and then you get matched to a team very far from you in MMR."
That's not really the case anymore, as there was a patch (no idea which one) that merged the queues of teams and parties. The number of active teams should have increased since then.
Valve has attempted maybe 4 or 5 fixes to team matching. So far all have failed. It's really an advertising problem at this point. People do not play it because people do not play it. If there are not a lot of teams involved it just doesn't work well.
Active teams is currently at 6174. That is 0.3% of players. It used to be 40%. I wish people were grouping into teams, but clearly it is not happening.
Anyway, what kind of que times do you see - basically the same as parties? Valve needs to push advertising for this. People don't know about it, or don't see advantages to having a team.
You meet TMM teams now and then (always +5/-9000 ofc) but there is no point in playing it really. If you always play as the same five players then it's better and more fun to just go scrim/play in open tours and if you aren't the same five you have to recalibrate so it's pretty pointless.
They should just have some incentive like maybe TMM ladder reset every month/every second month with top10 or so getting some prizes. The prizes could even be unmarketable if they are afraid of screwing up the economy.
Yeah, I think they would really need to run it like a rolling open tournament. But lots of those already exist. So there really isn't a need for Valve to have an official one. The amateur leagues are beginning to take up the niche.
Por favor, se conecte para postar comentários.
Is 4.3k bad calibration?