EPTGeneral Discussion - Page 3275
| Forum Index > Dota 2 General |
IRC chatter should remain in IRC - http://webchat.quakenet.org/?channels=tl.dota2 Posts that relate to topics with their own thread, such as in News, Tournaments or Strategy should go in those threads. | ||
|
Excludos
Norway8279 Posts
| ||
|
Talron
Germany7651 Posts
| ||
|
Bash
Finland1533 Posts
On July 09 2013 03:40 Erasme wrote: memes ? hu? theres a clear difference between #demonarmy and kappa. In one i clearly show my support to demon and eg, the other is shit you like to bash me no ? any reason why ? if i offended you on mm or here, i'm sorry. Unless i did it on purpose but i wouldnt forget that Esports memes are still memes, I need go no further than to refer to your signature if you're still confused. Both of your examples are equally inane me-tooisms that add no value to anything but give the speaker a sense of belonging through the clutter created. It's like doodling on a bathroom wall, except instead of expressing one's individuality everyone is taking turns to scribble the same joke or statement, and the joke is now that you did it too (again!). A portion of online discourse is always going to be like this of course, and it's easy enough to skim as a minor annoyance, but that's totally beside the point of this whole thing. I'm merely "calling out" your double standard. You seem to think I'm "attacking" you over some personal problem, but you shouldn't have much trouble seeing my actual position. You yourself expressed similar disgust of "Kappa" spammers, was that to you a personal grudge as well? | ||
|
FinestHour
United States18466 Posts
especially jayce | ||
|
SilverStar
Sweden18511 Posts
On July 09 2013 05:41 FinestHour wrote: my signature captures teamliquid, sc2, dota2, and league of legends all in one especially jayce Jayo? | ||
|
Yoshi-
Germany10227 Posts
On July 09 2013 05:41 FinestHour wrote: my signature captures teamliquid, sc2, dota2, and league of legends all in one especially jayce What's the LoL reference? | ||
|
Shaella
United States14828 Posts
Ok. Going back to the rating system thing Can anyone tell me an actual concrete difference between two players that are separated by like, 100 elo or MMR or whatever system you're using. Or 1 league level or something? Whats the real difference. The actual difference. I want something concrete, not something fluffy like 'they're better' | ||
|
canikizu
4860 Posts
"kids" | ||
|
Unleashing
Denmark14978 Posts
A league should be a bit harder to get through just based on a good streak though, but overall in matchmaking it would often just mean better mechanics and game sense and in dota a better ability to play in pubs since pubs are very different from pro games, many people treat pub games as though they were on a pro team when it's a completely different can of worms. A player that can utilize heroes and tactics that work well in pubs will go higher in the pub 'hierachy' than someone who is better at playing as a 5-man team. It will boil down to "they're generally better" but yea. Better at pubs that is. | ||
|
jrkirby
United States1510 Posts
On July 09 2013 05:58 Shaella wrote: I'm sorry, but.. Ok. Going back to the rating system thing Can anyone tell me an actual concrete difference between two players that are separated by like, 100 elo or MMR or whatever system you're using. Or 1 league level or something? Whats the real difference. The actual difference. I want something concrete, not something fluffy like 'they're better' The only thing you can say for certain is that they tend to win more often, as that is the only thing that causes them to have a higher MMR. | ||
|
sirreginold
United States557 Posts
| ||
|
Unleashing
Denmark14978 Posts
On July 09 2013 06:02 jrkirby wrote: The only thing you can say for certain is that they tend to win more often, as that is the only thing that causes them to have a higher MMR. Not necessarily, it's not that hard to have a lower winrate but higher MMR than someone. | ||
|
NB
Netherlands12045 Posts
http://dotabuff.com/players/played Players it got matched with consistently: Dero, ...(kappaface), 37blok, ArtGoD, Tempelchen, Cherepa,.... and thats only the people who i care to notice. | ||
|
Garfailed
Netherlands409 Posts
| ||
|
Blitzkrieg0
United States13132 Posts
On July 09 2013 06:11 Garfailed wrote: I got a smeevil treat, anyone know how to use it? Can't seem to figure it out Go into the compendium to the courier section and you can unlock another color. | ||
TheYango
United States47024 Posts
On July 09 2013 06:04 Unleashing wrote: Not necessarily, it's not that hard to have a lower winrate but higher MMR than someone. Matchmaking is meant to assign you to a rating where you will win 50% of the games the matchmaking algorithm assigns to you at that rating. Winning more than that will cause your rating to increase, winning less will cause your rating to decrease. So if you have two players who are separated by some amount of rating, then you know that: 1) The system expects them to win ~50% of their games at their currently assigned rating (accuracy will vary based on how robust matchmaking is, how many games the player has played, etc. 2) One (or both) of the following: - When the player with the higher MMR was at the MMR of the lower player, he won enough games more than he lost to cause his MMR to increase over time - When the player with the lower MMR was at the MMR of the higher player, he lost enough games more than he won to cause his MMR to decrease over time From these, the implication is that a player who is at higher MMR would, if he were matchmade into games the same way that a lower MMR player was, win more of those games than the lower MMR player (because he did so in order to reach his current MMR in the first place), and is therefore a better player from his ability to win games that the lower MMR player was not capable of winning. The accuracy of this is, of course, up for debate. | ||
|
Cool Cat
United States1644 Posts
although this might've changed very recently who knows, since a few days ago i was on a 12-0 smurf and all of my teammates had 300+ wins as far as i could tell | ||
|
NB
Netherlands12045 Posts
knock yourself out  It is however a tiny bit more complicated than that to solve the existing problems: First of all there exist the party system in of which will affect your mmr one way or the other. Given a party that consist of accounts that belong to different MMR 'ranges', the match MMR accuracy is obviously impossible to achieve and the result of that match will affect players MMR accordingly: The low could get raised too much while the high could get dragged way down. Secondly: What do we want from a Matchmaking system? There has been a lot of opinions around the matter: do you favor speed for game finding or do you favor accuracy? What actually defines a good game per say(close skill? what defines skill?)? And how do you solve the match inaccuracy problem for players at 2 ends of the ladder where inaccuracy is guaranteed? Are you just gona let the minority suffer even though they are the most vocal part of the community? My previous solution was introducing a slide where players could choose how the Matchmaking priority between search time and search accuracy. On the left would be instant search while on the right will be finding players of the exact MMR you have or no game. Too bad, a faulty matchmaking is still way better than a matchmaking that does not function and giving you a game. Introducing such option would simply divide the player pool even more and potentially make the search time absurd. | ||
|
Unleashing
Denmark14978 Posts
There are people that think there were only 600 people online when sing was #1 player MMR-wise? Wut. There was not 300k but definitely not 600 either, i think there was like 40 or 50k? | ||
TheEmulator
28103 Posts
On July 09 2013 06:37 Cool Cat wrote: well a problem with the mm system imo has always been that it places such a large emphasis on number of games played, where even if you're 40-0 on a smurf account you don't get matched with people with more than 100 wins, and likewise even in page 1-5 games, you'll have players who are pretty bad but just have a huge number of games played. although this might've changed very recently who knows, since a few days ago i was on a 12-0 smurf and all of my teammates had 300+ wins as far as i could tell Yes, Eric Tams mentioned on an official dev thread that number of games played was a minor factor in the matchmaking algorithm. Not the MMR algorithm though, just mm. Were you playing solo queue btw? edit: honestly I think that isn't implemented anymore. | ||
| ||
