Lyte you touched on it in one of your replies but it still left me confused. If there were more rewards in ranked for winning earlier and less for winning later (underdog holding back the opponent, like 300), what problem would that create? It would force everyone to play their best from the get go, because drawing out a game if you are winning doesn't earn you anything special. This should only be applied in ranks but it would make it less useful to troll because the more you troll the faster you lose. We know people do not want to lose and this system combined with other systems mentioned could weed out a lot of the trolls and lead to a more general idea of the player skill levels and more competitive games, no real padding of scores either.

What a coincidence. I just won Dominion game that was much too easy last night, and to confirm the mismatch I'd decided to take down our team Elos after the game. I've even considered doing it every game to look for patterns (I need time to compose myself between games anyways). Now I realize that our team Elos were further apart after the game than before, but after is all I've got.

It was a five minute queue time that was supposed to be two minutes - solo queue Draft Pick. I'm leaving the names out, but I have them if you feel you can use them.

February 19, 2012

Ally 1: 1311
Ally 2: 1464
Ally 3: 1447
Ally 4: 1344
Ally 5: 1288
Mean: 1370.8

Opponent 1: 1385
Opponent 2: 1326
Opponent 3: 1289
Opponent 4: 1374
Opponent 5: 1270
Mean: 1328.8

PS: The only suggestion I would have is to make the Approx. queue time more accurate. I often get queue times that are double the Approx. time; while high Elo players get queue times dozens of times longer than their Approx. times, and then stomp horribly. Is there a way to make separate estimates based on whether a player is queued for Blind or Draft Pick? How about making estimates that account for either the time of day or the number of people in the queue? Can you make estimates that account for the number of the user's Elo matches that are in queue?

I can't comment on mismatches, since I was never interested in training to a measured standard until the last two months; Dominion is just so engaging! I hope the problem with mismatches isn't very common...

Thanks for helping!

Hope you really have a good explanation for this one game, and trust me when I say this isn´t an isolated case, this happens A LOT

As you can see, it was a 4v5 that I managed to carry, but my point is not braggin instead look at how many wins those people have, GP left early game, this game for me was A SURE LOSS. This isn´t a comeback or " gj, you carried a game ", it was a clear disparity on how only one medium skilled player could abuse the enemy team, thank GOD my teammates listened to me, their vayne did everything right, but one can do so much against a burst heal soraka and a kennen/riven combo.

Seriously, how can you expect a good system while pairing people that have 108 wins with 0,1,3 ones?

I can guarantee that all the below 10 win people had poor skills/little knowledge, but notice that on the enemy team leona has 31, while all my teammates are below 10 that is clearly wrong.

While your matchmaking is working accordingly, the philosophy and design behind it is completely wrong for this game, you cant narrow down people skill/emotions/persona/ego for just that, there are a ton of things making an influence on each game, matchmaking does not count that, it is just a logical thing not taking into account human behaviour.

I will suggest one thing, you have TONS of data into player stats, damage dealt/ kda/ assists / kill spree the works, why dont you use that?

As a matter of fact, I will post an idea on a thread where I will explain how to use player data in a nearly non-abusive way, I will edit with the link after.

You can´t base people simply around elo, it is beyond wrong, it is doomed to fail based on how human behaviour will affect it in general.

Again, I can assure you that there was a clear skill disparity inbetween the teams, some people didnt even knew they were supposed to mid, or to jungle properly, some of them even said " Wow this is my 3rd game in ranked " while others went " Wow I never knew riven could top " or " How do I jungle ", one of them went into an argument with me about the meta, with such a shallow knowledge that I was baffled, he JUST FOLLOWED BLINDLY THE META, dammit, this is a strategy game, YOU HAVE TO THINK AND DEVELOP WAYS OF PLAYING THE GAME, god knows how many times I picked MF and the enemy followed picking soraka, it is THAT BAD.

EDITED BY LYTE ::: Saved the screenshot, thanks!

I just have one request: Fix smurfing

Will someone PLEASE look at my normal elo.

Ranked games are fine.I get good games in ranked,Im almost back to 1200 elo and ive been winning 70% of the games I play so getting to a good elo isnt going to be hard now that I learned how to play the game.

My normal games elo tho has to be like 200 or something I swear.Never any junglers i mean people are so so so so so bad in the normal games I solo que.When I que with a partner it isnt that bad.I swear something messed up my normal game elo.PLEASE check it for me.I emailed you guys and they said to come to this post so im begging you please look into my normal game elo for me.

ANOTHER example of poor matchmaking. My team plays at ~the 1400-1500 level. We got matched v Riot (phreak, janook, etc.)

Here's an idea:

Take a page from World of Warcraft's book and allow people to queue for the roles they want to play. I can't stand buying a champ and not being able to play it because I am always forced into a role that balances the team. No more fighting over who is jungling and who is going mid or top, and duo queue bot could be really fun. It isn't a perfect solution and doesn't really flex the meta to much, but you might see more unorthodox champs in different roles.