Something similar to this is highly valuable for matchmaking work but is in demand at Riot in general:
http://www.riotgames.com/careers/statistician
http://www.riotgames.com/careers/business-analyst

I do not have any Screen shots of these games but two games stuck out to me this week that were made unpleasant do to poor match making.

I played a Ranked game were my team had all 50+ wins in ranked. The opposing team had atleast 3 people with less than 20 wins. To say the least we destroyed them this occurred right around the 1200 elo range.

On that same note I played a Normal game this week with some friends, we are all around the same number of normal game wins 500-800. The opposing team had 3 people greater than 1200 wins and their other teammates were about our skill level. This time my team got stomped hard.

Both games were summoners rift.

Thanks, it'll be awesome if these are still open when I'm ready to graduate next year, hope we get to work together some time.

the most unpleasant experiences are usually when I have a duo on my team. we almost always end up losing and I dunno why even.

Forgive me if I'm forgetting a variable, but I don't see how this is a difficult issue to fix. If there's 5 guys who average around 2000 ELO, fighting 5 guys who average around 1800, why not just split the difference and have some of the members swap teams.

2200
2000
2000
1900
1900
(Average 2000)
vs
2000
1800
1800
1700
1700
(Average 1800)

If swapped appropriately, the ELO difference could be

2200
2000
1900
1700
1700
(Average 1900)
vs
2000
2000
1800
1800
1900
(Average 1900)

Resulting in a 1900 vs 1900 ELO Match. The formula to figure out the best ELO vs ELO match up of 10 players would not be difficult for any computer to solve, if coded correctly.
Of course, not all ELO differences would result in a perfect ELO match up, but I would say that you could have every match near a <100 ELO Difference between teams, which I feel is an acceptable amount of difference.

1) What is your Summoner name? Crimsonkid5

2) Ranked

3) Summoner's rift

4) Yes
a.1(duo)
b.I was 1300ish and my friend 1400



6) Any specific reason(s) you felt the game was unbalanced?
We had a troll twitch who followed the enmy jungler giving him 3 kills. a 07 Xin and a 0/8 udry
one of my first ever duo q's and ever since I always get trolls in ranked. There are 7/20 games(duo/single q) where The situation was not similar to this one.

7) A screenshot of the end-game score. (none because it was season1 and have lost the picture since then

I just thought of something that might help matchmaking a little without going into a bunch of complicated mathematics.

Make it so when you queue for ranked you pick the roles your willing to play (top, mid, bot, jungle). Then have the system try and match players of equal elo to the same lane as well as how it already determines matching. That way even if the system does put you with high elo players when yours is low, it will more likely have the high level on the enemy team laning against the high level on your team.

It wouldn't solve all problems your trying to resolve. But I think it would definitely be a step in the right direction.

EDIT: this would make champion selection an overall better experience as well.

just another quick note. Didn't list bot as AD carry and support because that's only the meta. You don't have to play bot like that, kill lanes can work perfectly fine (in lower elos mostly, But once you start getting to the higher elos you should probably know how to fill either role anyway).

Also it's common knowledge that top is a solo lane and you have a jungler, That's not really a meta anymore. That's pretty much set in stone. This would just solidify the concrete that's already in place.

Everyone here agrees that the most annoying thing you can have in a game is AFK's, ragequiters and trolls, just that the MM can not predict, has no way of knowing if someone is going to be a troll or not, this part we have to leave to the tribunal. What everyone also complains is playing with players who have no idea what they are doing, the only way I found to know if someone is experienced or not is the number of games played, so I thought a formula that calculates a "true elo" that would be used by MM:

true elo = [(total games) x (won +1)] / [(leaves +1) x (loss + 1)] x Elo

The more matches you play you'll be more experienced and his "true elo" will be higher, also note that if you leave a game you will be heavily penalized and your "true elo" is halved in the first leave, this leave leavers together.

I put a friend of mine as an example, he has 142 games with 75 wins and 67 losses, and his Elo is 1257: His "true elo" would be [(142 x 76) / 68] x 1257 = 199.493 and when the MM was look for a teammate and opponents they would have their "true elo" near that number.
I put my friend as an example because I have another friend with 2 wins, that's right only two victories, and he has 1315 Elo. Putting in the formula: 2 x 3 x 1325 = 7.890 realizes how vast is the difference? He even jokes that does not know how he won both matches and said he will never play ranked again to not lose the Elo lol.
I have 1727 games in normal mode and started playing ranked this season, with this amount of games I can say I know how to last hit, I know about putting wards in important spots, these things of the game that seems simple but it seems that few people knows. A little story just to illustrate what I mean: Before yesterday I was in a game, win a fight in bot lane I was with half HP I went to dragon called my teammates (I wrote in the chat that I needed help several times and ping) so they were all running to gank top, in short, I could not kill the dragon alone had to run and the top player escaped.
Another thing I noticed is that the ELO system of the LOL is not the same one used for chess, in chess that the higher your ELO more ELO you lose if you lose to a less experienced player. Let me give an example: Last week I was playing with my 1257 friend in duo queue, I had 750 ELO in the day, first game we won, he and I won 6 ELO, until there normal I was playing with someone with more ELO than me. Second game we lost, I was thinking "Omg he will lose like 50 ELO" we both lost eight ELO, yes only 8, this is less than a queue dodge. Want people to stop complaining about the duo implement the REAL ELO CHESS. It would be like: OK 1700 sir you want to help your 900 friend get ELO fine, if you win (you have to win) you get 5 ELO, if you lose you lose 100 or more. The people will think twice before the duo, then they're going to call his friend and say: Ok dude let's play a normal and see how you play, do you play suport or tank or jungler? let me explain some things about the game ...
I really hope that MM is improved because I got tired of playing with people who do not have any notion of game strategy and think only of killing, leaves the base fall to chase players, not to mention leavers, trolls ...

OK, Heres the real problem with the matchmaking and elo.

1: with more than 80 champs available, no mather how good you are, you cant play them all at
the same level and team matchup often require you to use a champion your not familiar with for
the good of the team. Going to calculate elo after champion selection for matchmaking would be a pain so I'm really not sure if theres even a way to fix this. I'm at a point where I dont even try new champs in coop. I dont carry as hard with a champs I use for the first time, but at least my opponents provide a challenge. (blink at expert bot)

2: When I started ranked game, there was only solo queue, ended up in a lots of games where whole team was negative cept me, leavers, ragers, teleport revive eve etc, ended up in the 800 elo ( 965 atm). I dont know what is my normal game elo but for now, even if I climb back a lil (74-85 ranked), I still feel like a teachers and If I dont feel like taking the lead on the chat, to ensure my party group up, wards, tank and support protecting the carry, dps focusing the right target, warding, moar ward baron, dragon, warding my jungle if they invade and repeat after evry fight that we have to focus the carry its almost insta lost. So yeah, I end up having more fun in normal game since It seems that those players are more of the same elo as me, cept when they try champs. (615 win - 595 lost normal game)

3: When I Queue with low level friends, Its always a rofltruck, either us or them will simply crushed the other team, and once agains, I have no clue about how to balance this, I ended up playing snowballing champs in these kind of games to keep a good win - loss ratio. I'm still not sure how my lvl20 friends (he was lvl 7 when we started playing pvp together) elo is affected by this since I know that all I have to do is find the low lvl, get fed, and wait till they surrender...or they find him and yeah..GG

4: Have you seen my bear Tibber?

solo queue
80% of games are 20-25 minute surrenders
W/L is balanced but only because which side is getting destroyed flips not because the matches are even close to even.