Broken Matchmaking

First Riot Post
12
Comment below rating threshold, click here to show it.

Fr0g9

Member

10-28-2013

Whats so hard about determining someone's in-game skill level? Nobody, not one decent player, wants people who are complete infants at this game in the same match as them, ON EITHER TEAM. There is nothing fun about wrecking someone who doesn't even put up a fight at any point, there is nothing fun about having 1 or 2 people on your team who are simply below average.

If you actually manage to win with a couple of baddies on your team, how good you think you are had nothing to do with why you won. The only reason you actually win a game like that is because the other team also had some terrible players on it.

A simple website can tell me who's good, who's average, etc. I don't see what's so hard about implementing in-game matchmaking so we can get some actual competitive games going.


Comment below rating threshold, click here to show it.

LaochII

Senior Member

10-28-2013

No, a simple website can tell you what other players recent run of form has been. It tells you very little about how they will perform in their next game. That's why matchmaking seems broken at times; Riot has not yet learned how to predict the future.

I hear their R&D department is working on it for Season 4 though.


Comment below rating threshold, click here to show it.

Fr0g9

Member

10-28-2013

Quote:
Originally Posted by LaochII View Post
No, a simple website can tell you what other players recent run of form has been. It tells you very little about how they will perform in their next game. That's why matchmaking seems broken at times; Riot has not yet learned how to predict the future.

I hear their R&D department is working on it for Season 4 though.
Overall Kill+Assist to Death ratio tells you a lot. Narrow it down to each champ to see what the averages are per champ, what the players average is in comparison, and you can tell how good everyone is. The numbers don't lie.

And I really hope they get this ironed out for the next season, this matchmaking is by-far the biggest problem in the game right now.


Comment below rating threshold, click here to show it.

NaginataX

Senior Member

10-28-2013

Quote:
Originally Posted by Fr0g9 View Post
Overall Kill+Assist to Death ratio tells you a lot. Narrow it down to each champ to see what the averages are per champ, what the players average is in comparison, and you can tell how good everyone is. Numbers don't lie.
Not really. Go ahead and lolking my last couple matches and take note of how I performed with Amumu recently. THEN check my win % with him. That game you would consider I probably "played like a noob", but in all honesty there was absolutely no difference between my gameplay that game and the ones I've won. The difference was that my team was unable to recognize that ult = engage and by the time they did, they were underfed and couldn't even do damage when I had the enemy team stunned. Does that make me a bad player and/or suck with Amumu? No...it means my score sucked because that team didn't play well. Now if this was a habit then yeah, I'd be more inclined to believe that I actually play poorly and need guides/advice/etc to improve.

You forget this is a team game. Your score is directly effected by how the rest of your team does. If your bot feeds like a buffet, expect everyone to get killed by their ADC a lot more than if your bot lane was dominating. Doesn't mean mid/jungle/top are sucky players, it just means that the gameplay of one individual impacted your score.

Another example. Say you Malphite ult into their team (5v5), but everyone on your team is attacking their tank and you get aced. How could Malphite have done anything different? His gameplay would have been the EXACT same as if your team was smart and focused the ADC/APC. In one scenario Malph dies and gets no assists, the other he stays alive and gets 5 assists. See the problem with saying a player's score is a direct reflection of their skill?


Comment below rating threshold, click here to show it.

Alobo

Member

10-29-2013

Gotta upvote Naginatax, it's true, just last night I had a ranked game where I picked panth last second to go top vs yorick just because someone decided to jungle and not top last second, the trolls, I know, but anyway, according to everyone, I was to get wrecked, right? I get out of lane at 4/0/1, and look at my end score 5/4/3, now you would say I'm a bad player by looking at that and a few other stats, but I wreck as panth, don't mean to be cocky, but while I am not as good with other champs, I consider myself an overall good panth player, a good player in general actually, so ya, Naginatax takes this one again, gg wp bro : P


Comment below rating threshold, click here to show it.

DontDerpOnMePls

Junior Member

10-29-2013

Yeah Kda doesn't always matter, though it can be a decent judge of things....its not concrete as they say... you could have one person who is very good at last-hitting and nothing more or a person who dies a lot but pushes all way to enemy base inhib and tanks for the team and never gets a kill but wins a lot.....Kda can be an elusive stat... & then there are people that view it as if well my kda was even I did well...others more like myself would want twice as many kills over death to say you were decent.....but in gameplay it doesn't always matter...<good points NaginataX>


Comment below rating threshold, click here to show it.

Zileas

VP of Game Design

10-29-2013
1 of 1 Riot Posts

Quote:
Originally Posted by Fr0g9 View Post
Whats so hard about determining someone's in-game skill level? Nobody, not one decent player, wants people who are complete infants at this game in the same match as them, ON EITHER TEAM. There is nothing fun about wrecking someone who doesn't even put up a fight at any point, there is nothing fun about having 1 or 2 people on your team who are simply below average.

If you actually manage to win with a couple of baddies on your team, how good you think you are had nothing to do with why you won. The only reason you actually win a game like that is because the other team also had some terrible players on it.

A simple website can tell me who's good, who's average, etc. I don't see what's so hard about implementing in-game matchmaking so we can get some actual competitive games going.
We are making some accuracy enhancements over the next few months, but honestly, it's pretty accurate. It is very easy to confuse a lop-sided game and a lop-sided matchmaking result.

We've all had games where we went 0-8 because we weren't on our top game, or made some early mistakes, or had a bad set of match up, or the other team had great team chemistry, or whatever, and these can drive a snowball loss that makes one doubt the matching. We know from the data that the matching is actually pretty accurate. But we do want to improve it further.


Comment below rating threshold, click here to show it.

Chazer

Member

10-30-2013

I recently had a match in normal queue where our 3 man premade was given a duo on our team with around 150 wins in normals(one of which was lvl 28). The enemy team had 2 duo's with 1150ish wins as their lowest. The game was "one sided" to but it nicely. Is there anything we as players are able to do to help improve matchmaking, i.e. report a case like this where clearly something is amiss?


Comment below rating threshold, click here to show it.

Natasha Wong

Senior Member

10-30-2013

Quote:
Originally Posted by Zileas View Post
We are making some accuracy enhancements over the next few months, but honestly, it's pretty accurate. It is very easy to confuse a lop-sided game and a lop-sided matchmaking result.

We've all had games where we went 0-8 because we weren't on our top game, or made some early mistakes, or had a bad set of match up, or the other team had great team chemistry, or whatever, and these can drive a snowball loss that makes one doubt the matching. We know from the data that the matching is actually pretty accurate. But we do want to improve it further.
Hopefully these come soon.

Hasn't been bad the last two days, sometimes I get great matches, but some weeks I encounter nothing but toxic players during champion select teams with Absolutely Zero Communication, forced into playing (X) Role, or just choose whatever I feel works and ignore my team in game to avoid hearing them rage and say anything in chat at all so I can't get punished by Leagues broken Tribunal system.

The problem with League currently is the random match-making system some days its great other days its full of toxicity AFK's, Feeders, Passive trolls, which passive trolls are the worst because it always gets over-looked by Tribunal punishment because their goal is to make another player on the team say any little thing that goes against summoners code then report them.

If the Match-Making was actually improved to where players could freely leave a game when matched with a team they can't work with and another player in Q joins that game then there would be less issues after all having to Double jungle because an ADC Instant locks something and expects me to support them just because of being last pick or something is pretty stupid. I don't mind playing support or ADC role but I will not play with just any champion lay-out or person I need specifics or I play alone in my own lane.

And since I don't have the ability to dodge a game freely other than the 5 minute penalty I will just lock a jungle and farm the jungle and not care what my team thinks about it because that is exactly what other players do when they instant lock an ADC, or support and want me to bot with this, and better yet I am sick of the attitudes in the pre-game of STFU N00b Last pick Don't play ranked, Go kill yourself comments are pretty toxic but these never make it to tribunal.


Comment below rating threshold, click here to show it.

Xeones

Senior Member

10-30-2013

Quote:
Originally Posted by Zileas View Post
We are making some accuracy enhancements over the next few months, but honestly, it's pretty accurate. It is very easy to confuse a lop-sided game and a lop-sided matchmaking result.

We've all had games where we went 0-8 because we weren't on our top game, or made some early mistakes, or had a bad set of match up, or the other team had great team chemistry, or whatever, and these can drive a snowball loss that makes one doubt the matching. We know from the data that the matching is actually pretty accurate. But we do want to improve it further.
Isn't a lop-sided game a lop-sided match making result?

I'd say 70-80% of my games are either easy wins where I am just completely outplaying the opponent and they are doing dumb things like staying in lane at 2 shot range and my team just dominates. Then it seems like the system, seemingly, decides I've won too much and gets me team that will not win.


12