EPTJwaow has picked reworked nid vs lulu, started dblade
[Patch 4.10] Nidalee/Skarner Rework General Discussion - P…
| Forum Index > LoL General |
|
Fusilero
United Kingdom50293 Posts
Jwaow has picked reworked nid vs lulu, started dblade | ||
|
ZERG_RUSSIAN
10417 Posts
On June 30 2014 21:32 PrinceXizor wrote: well think of it this way. if the acceptable variance in MMR between players when matchmaking is say 20, then 1380 and 1400 could be matched together, the first method would try to find 10 players between those values and match them, and whichever side incidentally ended up with more MMR goes purple. in the second method, they find two teams who each within themselves fall within that 20 variance. say a team from 1370-1390 that averages 1380, and a team that varies from 1390-1410 that averages 1400. they could be matched up with eachother in the second system, despite the overall variance between players being double that which is the "acceptable" variance because the variance within the teams and between the TEAMS is within the acceptable limits. The difference may not be great. but it is certainly a difference. So how do you account for when a bronze duos with a diamond in the first example? It doesn't seem possible. Also I'm pretty sure you're using the word variance where you should be using difference. Variance is a measure of spread but it's not in the same units as the spread itself. It's the standard deviation squared. So to be clearer, the variance would be expressed in units of (MMR^2). | ||
|
Zess
Adun Toridas!9144 Posts
On July 01 2014 04:39 Fusilero wrote: http://www.hitbox.tv/ChallengerFaceOff Jwaow has picked reworked nid vs lulu, started dblade I tuned in to see N! get delay aced at Dragon and then lose dragon. Looks like Ocelote has found the scene in which to demonstrate his amazing skills. | ||
|
chalice
United States1945 Posts
On July 01 2014 04:36 wei2coolman wrote: Such a small sample size from OP, and it's not like his mmr jumped significantly outside normal variance. Posts like these just make me shake my head, who da fuck cares, just play the game and learn to carry. Even if you're on a losing side of a composition, doesn't mean there isn't anything to gain by playing it out. lol there's plenty of champ selects where it'd be worth it to dodge just from an enjoying the game point of view even without the possible mmr/lp benefits. | ||
|
Fusilero
United Kingdom50293 Posts
Hey guys guess who's disabled again! | ||
|
chalice
United States1945 Posts
On July 01 2014 04:46 ZERG_RUSSIAN wrote: So how do you account for when a bronze duos with a diamond in the first example? It doesn't seem possible. Also I'm pretty sure you're using the word variance where you should be using difference. Variance is a measure of spread but it's not in the same units as the spread itself. It's the standard deviation squared. So to be clearer, the variance would be expressed in units of (MMR^2). are you trying to tell him that MMR doesn't vary? if the MMR varies then CLEARLY variance is the correct word to use here. | ||
|
Zess
Adun Toridas!9144 Posts
On July 01 2014 04:54 Fusilero wrote: https://twitter.com/RiotDeficio/status/483699650114121728 Hey guys guess who's disabled again! Let me guess, it's Sion right? | ||
|
Alaric
France45622 Posts
Ryze/Katarina/Skarner/Morgana/Tristana was a damn scary team comp to deal with in the late game, especially when Kat started getting fed. Thanksfully, with Irelia/Yasuo/Xin/Nami/Vayne we had a bunch of cc to deal with her and some damage potential past midgame too, although ours was more single-target oriented, and 3 melees + one short-ranged AD. Tristana really was a darn thorn to get to (especially as Irelia, I'm not tanky enough to go back in and I'm reliant on burst). We ended up winning because we managed to catch people alone through rotations in their jungle, especially after we got the first inhib and basically danced around Skarner. Last fight had Skarner flash-ult Vayne over a wall and Katarina kill her (getting low and using zhonya thanks to QSS), then for some reason they all stacked on her so Nami set up Yasuo and my ult for sheer destruction. Still, we died so fast when Tristana didn't split us up and they got their AoE burst, despite my first item post-triforce being locket... And that comp scaled hard. | ||
|
Ryuu314
United States12679 Posts
On July 01 2014 04:54 Fusilero wrote: https://twitter.com/RiotDeficio/status/483699650114121728 Hey guys guess who's disabled again! i find it hilarious that Shen is perma-disabled now, considering his ult bug is at least 3 seasons old, but he was never disabled during his 90%+ pick/ban reign. | ||
|
Alaric
France45622 Posts
| ||
|
arb
Noobville17921 Posts
On July 01 2014 05:03 Alaric wrote: Maybe if Crumbzz keeps picking Vi for no good reason at all at least she'll glitch 3 or 4 times in a single game and Riot will be forced to acknowledge her issues. >_< I've played a shitload of games as Vi and never had any of those Issues lol. Ive seen my friend pick her and go thru people with Q and do no damage/not knock them up though. | ||
|
ZERG_RUSSIAN
10417 Posts
On July 01 2014 04:57 chalice wrote: are you trying to tell him that MMR doesn't vary? if the MMR varies then CLEARLY variance is the correct word to use here. Not in the statistical sense. I could be wrong, but what he's describing is a difference in MMR between two teams. Variance is a MEASURE of that but not in the units that he's describing. But whatever, I understand what he's saying even if he's using the wrong terms for it. In any case, he's wrong and the matchmaking algorithm is much more complicated than that. | ||
|
wei2coolman
United States60033 Posts
On July 01 2014 04:51 chalice wrote: lol there's plenty of champ selects where it'd be worth it to dodge just from an enjoying the game point of view even without the possible mmr/lp benefits. Those are few and far between. | ||
|
PrinceXizor
United States17713 Posts
On July 01 2014 05:23 ZERG_RUSSIAN wrote: Not in the statistical sense. I could be wrong, but what he's describing is a difference in MMR between two teams. Variance is a MEASURE of that but not in the units that he's describing. But whatever, I understand what he's saying even if he's using the wrong terms for it. In any case, he's wrong and the matchmaking algorithm is much more complicated than that. I used english not math or statistics to describe what i was saying. sorry, i'm haven't had anything to do with any actual math since sophmore year in high school. got all that crap taken care of then. please explain the matchmaking system since you seem to know how it works thoroughly. | ||
|
chalice
United States1945 Posts
On July 01 2014 05:23 ZERG_RUSSIAN wrote: Not in the statistical sense. I could be wrong, but what he's describing is a difference in MMR between two teams. Variance is a MEASURE of that but not in the units that he's describing. But whatever, I understand what he's saying even if he's using the wrong terms for it. In any case, he's wrong and the matchmaking algorithm is much more complicated than that. i was mocking the way people will argue english semantics instead of just admitting they made a mistake, tried to make it obvious by capitalizing "CLEARLY" but communicating on the internets is hard. i don't think you're wrong at all. | ||
|
Goumindong
United States3529 Posts
On July 01 2014 05:23 ZERG_RUSSIAN wrote: Not in the statistical sense. I could be wrong, but what he's describing is a difference in MMR between two teams. Variance is a MEASURE of that but not in the units that he's describing. But whatever, I understand what he's saying even if he's using the wrong terms for it. In any case, he's wrong and the matchmaking algorithm is much more complicated than that. A team can be thought of as a set of 5 players drawn from a population such that E(MMR) = Team MMR. This will mean that there will be variance between the individual players since each persons MMR is not precisely the team MMR. Each team is another random variable and is the sum of the 5 random variables on the team. This also has its own variance (which is lower than the individual variance). So if we select teams by grabbing 5 players and then finding another team of 5 players at once which are relatively close then putting the higher MMR on purple this is different than finding 10 players and then grouping such that purple has the higher MMR. | ||
|
Alaric
France45622 Posts
| ||
|
Gahlo
United States35160 Posts
On July 01 2014 06:47 Alaric wrote: "MMR" will never accurately represent people's inability to look at the teamcomps and use their brains. x_x And unfortunately how the game evaluates us. | ||
|
Slayer91
Ireland23335 Posts
mmr=skill statistically pretty much the only reliable metric even if it has medium variance | ||
|
MooMooMugi
United States10531 Posts
| ||
| ||
