Boards›[ARCHIVED] Miscellaneous

[Inquiry] Mismatched Teams in TB - This is not a rage thread

Hyrum Graff·5/12/2014, 5:11:14 AM·4 votes·2,425 views

Earlier in the day, I was in a TB match where 4 silvers and a bronze were paired with a bronze, a silver, 2 golds and a plat. Needless to say, we (the silvers) got stomped; I'm kind of befuddled about how two blatantly mismatched teams could have been placed against one another.

###Let me be clear. Though this thread exists because of a bad experience,
##this is NOT a "TB matchmaking sucks, rito plz fix" thread.

Instead, I'm genuinely curious about how the matchmaking system works, such that a situation like this is possible -- and you can probably count on me trying to think up a solution to the problem (if one exists) given the information you give me ;) Could it have been MMR disparities between ranked and unranked? That we were a 5 man premade? Low volume of people using the TB que? Some combination thereof? Something I'm not considering?

People on the TB project: @Lyte, btab, ceciliaicy, davin, DoctahWayne, Drevarius, Fr3nchT0ast, Goggalor, MissMechaZero, RiotBahamut, RiotBlackrook, RiotComplex, RiotExLibris, RiotId, RiotJayway, RiotLobster, RiotLulu, RiotMartlet, RiotNeuroCat, RiotNoyz, RiotSeb, RiotTantram, RiotWorldPeace, Ryedan, scezumin, statuskwoh, stochasticmx, VonBurgermeister, Yegg and Yupster.

Other cool people on the forums who could forward this to the team: @BuboBubo, Riot Girl, Baconhawk, whist, xelnath, morello, ghostcrawler, LtRandolph, RiotSargonas

General: @Riot, reds, rioters, rito

######in other news... THERE'S A RIOT LULU!!!

24 Comments

RiotBuboBubo5/12/2014, 5:32:39 AM1 votes

Hey Hyrum! If you're up for it, I'd like to turn this question back on you and see what comes up. I suspect you'd enjoy it more than a direct answer but would be happy to provide you my guess if you prefer.

So here goes: How do you know that the teams were blatantly mismatched? Whatever the signal, can you think of cases where even well matched game would lead to this outcome?

Kicking this up to the macro level, how would you judge the quality of a thousand or a million match made games?

Edit: rephrased to sound less dismissive, thanks for calling this out Seminole

Garnet Dust5/12/2014, 5:52:37 AM1 votes

I want to know what happened to the patch 4.7 Team Builder updates? Specifically,Solo players can now see which role/position combinations are in high demand when deciding their spec •"Play Again" has been added to the post-game screen, allowing Captains to automatically invite their friends from the same match to a new Team Builder lobby.

Yeroban5/12/2014, 3:36:08 PM1 votes

While I understand there are differences of opinion in the diligence one employs while playing ranked vs teambuilder vs normal SR, I've always thought a player's matchmaking rating, or MMR, should be based on the map they're playing on, not the game type (*excluding novelty game modes like URF, AllForOne, etc). There are no differences in the gameplay between non-ranked and ranked aside from bans. And very rarely do I see players or teams troll or goof off to toxic extremes in non-ranked game modes.

I think matchmaking could be more accurate if the MMRs for modes on the same map were continuously linked in some way; not completely, but partially. I think this would be useful for beginners and players who predominately play one game mode, when trying out a new game mode, as well as those who do not play certain game modes on a frequent basis.

What would be the downsides to more closely linking MMR between varying modes on the same map?

Aezuriel5/12/2014, 6:39:24 PM1 votes

I feel there is a major issue here that is not being explored, and that would explain a major hesitation for transparency in Team Builder. In part, people have mentioned Separate MMR's for separate game types, but haven't really explored the WHY? of it.

From my own experience... My team-builder games have met with MUCH better results (and w/l record) than my previous experience in matchmaking. While I expected to have better games, it strikes me that it is a statistical impossibility for EVERYONE to experience the same raise in win percentage, that I have in TB.

Just talking about it makes me a little nervous... because I don't know what to attribute it to.

  • Have I always been more team focused, and TB allows me to shine more?
  • Have I learned to game matchmaking and my MMR better over time, and a new play mode allows me to use that knowledge from game 0?
  • Am I just lucky? Is my performance in TB a statistical outlier?
  • Is Teambuilder being played more often by persons of a lower skill level than me?

I have a few guesses... but I don't know for sure what is causing the disparity in my performance. So, I can see why Reeto wouldn't want to say anything either. How can you be transparent, when you are still analyzing the data and trying to make it better. Why tell players how to game the system that is in place, when you are not sure if it is monkeyproof (though nothing is ever 100%).

Please don't mistake my tone ... I think having this discussion is a good thing.

But, I don't think we should soon expect the magician behind the curtain to show us the wires and duct tape holding everything together.