112 réponses à ce sujet

[Terraflare]

Hi everyone. Over the past month I have noticed many threads created by players who are frustrated after having been kicked from a Gold lobby, seemingly for no reason, equally by hostile/non-hostile hosts/groups. While I don't advocate senseless and most importantly, misinformed kicking, I'd like to share some of the lessons behind why this occurs, and hopefully if you feel wronged about a kick, you might feel better.

For any Economics graduate, there are 2 key forces at play here. 1) Asymmetric Information and 2) Adverse selection. Briefly speaking, 1) revolves around the fact that a private agent (referred to as "the seller" possesses unobservable qualities that the "buyer", (in this case, the host) is unaware of. In the real world, these are your individual qualities - your intelligence, willingness to put in effort, trusthworthiness, responsibility, creativity etc. In a coop shooter like ME3, these will comprise of qualities such as your skill, experience with other FPSes, willingness to cooperate, winning mindset, and in general efficiency/capability at this game. 

How is this relevant? Think about a job interview session. There are 2 candidates - one who walks into the room, with a polished resume, with a 1st class degree from a top university, has represented his school at 4 international sporting events, and started that hedge fund society. The other is what one might call an oddball - on paper, not too bad, a decent degree from a middle university, decent grades, participated but doesn't have anything particularly stellar down on paper. Let's say the 1st guy actually has an IQ of 150, and the 2nd guy has an IQ of 3000. 

The interviewer has 10 minutes to decide who is best suited for the job, and who gets the "thank you, please try again next year" formality. Why does the 1st person get picked, even if his inate ability (his IQ), is far surpassed by the 2nd guy? The answer is clear, because IQ/ability is always a hidden piece of information, the employer will never find out the true quality of a person until the person is hired. In the eyes of the interviewer, the choice is clear - the first guy has "jumped through the hoops", however meaningless they might be. He has put in enough effort in life to achieve the achievements/accolades (albeit, paper), in order to 'prove' that he is capable. All this effort is not costless - it takes time, patience and determination. While the interviewer cannot observe the candidates' raw IQ/ability, he can, via his resume, observe exactly the qualities he would like - the determination and responsibility to get something done.

Relate this back to your game. When you enter a lobby with a low N7 rating, a "suboptimal" build (whether or not some of the builds are indeed suboptimal, is another story), and a low character level, you signal undesirable characteristics. Never mind if you are actually SpawN / cooler behind that keyboard. From the rank of your weapon, I can tell how much/how little you have played and bought packs. From your character level, I can tell whether YOU think a level 5 character will do just as well as a 20 because you think "you are good". From your N7, I can tell whether youve been sitting behind that keyboard playing for 200 hours instead of 20. Hell, even from the numbers in your nickname, I can tell your age (MrDog94 vs MrCat85) These are all imperfect approximations of what I wish I could observe (your true ability), but it's all I have.

A high N7 rating doesn't mean a player is better, it simply signals to me that this person has been playing longer, and is thus more likely to have run gold matches multiple times/know what's going on. Underpinning this is the fact that gaining character levels/upgrading weapons/getting a high N7 is a COSTLY signal that cannot be FAKED. This is the core concept underlying why lobbies kick - I would rather pick a guy who has ticked all the checkboxes (mundane and imperfect though they may be), rather than risk a terrible game with a person who might have simply found a way to shine and "make things work" despite not ticking the boxes. The act of ticking the checkboxes signals desirable characteristics in itself.

This brings me to 2) Adverse selection. Robert Akerlof won a Nobel Prize for illustrating this concept, underpinning used car markets and pricing. He observed that a brand new car (peach) straight out of the showroom, will see its price plummet when it is immediately put on the used car (Lemon) market, the day after it was bought. Why? To summarise and put it simply, this is because that the very fact that it is in the used car market signals to potential buyers, that on average, it is a lemon (bad car) as opposed to being a peach (good car), and is thus worth a lemon's price, not a peach. The market collapses, because potential peach sellers (people who want to get rid of good cars for legitimate reasons, such as moving out of the country), will refuse to sell their cars at lemon prices (they know its worth more). This is a self-fulfilling cycle - in the end, the only cars left in the market, traded and sold, MUST be lemons, and such, any car that enters the market must be a lemon.  

Again, how does this relate? Think of there being 2 types of players - peach and lemon-types. Peach types know the game, know how to make things work, know which combos work, which dont, what class synergies are good. Lemons don't, they are bad at this game. The peach player will never choose to use a poor loadout/character/equipment/mods (on average). Why would I use a BW + a claymore on an Asari? If I want to use a claymore, ill pick a good class with it. If I want to snipe, doing it with an Infiltrator is most efficient. What this implies is that anyone who is using a poor/inappropriate loadout is automatically categorising himself into the market for lemons - ie. he is poor player. Whether or not this is fair is another issue, because again, rational agents (the host), are not willing to risk a bad game for the offchance that you can make that Avenger IV loadout work in ways he/she has never thought possible. As such, they get the boot. Notice this, similar to Akerlof's lemons vs peaches, is a self fulfilling cycle. Poorer players cant join Gold - cant farm credits fast enough - cant buy packs - cant upgrade weapons - automatically enter the lemon market by having lower leveled equpiment. The only way to break out of the lemon market is to again, jump through the hoops - farm enough credits on lower levels until you have enough credits to eventually distinguish yourself as NOT being a lemon. 

So well I am done, I hope this has been interesting for at least a few readers out there. Kicking is an unpleasant business (sometimes for the kickER too!) - just as job interviews are. You have a very limited time to convey as much information as you can about your worthiness as a player, and sometimes you will get the boot when you dont deserve it.

Do feel free to give your comments!

[Tangster]

TL:DR.
"You have a very limited time to convey as much information as you can" is what I took from it.

Modifié par Tangster, 30 avril 2012 - 04:33 .

[PluralAces]

I'm not going to lie I didnt read this....but

[pprrff]

it's only a game

[osseywildborn]

I think that covers most of the bases.

there is one more indicator you could have added. for xbox theres an achievement called unwavering for completing every map on gold. achievements for players can easily be found by comparing games. i've found that generally these players are much more competent than those without it.

[nwj94]

Interesting I read the whole thing and I understand what your going for. The problem is not when people make rational decisions (N7 23 carrying a mantis and a katana on a adapt in a Gold game) The problem is when people set their standards to high and just kick anyone who doesn't live by their specific criteria. (Your not a SI/GI with BW/M-98, lvl 20 and a N7 200+? Clearly your a noob! GTFO!)

People's beef is really with people who hold an interview where you need to by a Phi-beta-Kappa with a PHD from Harvard so that you can apply for a middle mangers position. People who act like they are the master and everyone else needs to conform to THEIR specific views on what constitutes a good build.

Overall a good read, you clearly put some thought into it.

[Maria Caliban]

A PUG lobby is like a 12 second job interview. If you show up looking like crap, don't complain about being kicked.

[PluralAces]

Maria Caliban wrote...

A PUG lobby is like a 12 second job interview. If you show up looking like crap, don't complain about being kicked.

I like her style  too bad she only plays on PC 

[xtorma]

Tangster wrote...

TL:DR.
"You have a very limited time to convey as much information as you can" is what I took from it.

i got , you are evaluated by what can be seen , and percieved, not by your true qualities. All i can say to that is.....no kidding.

[Terraflare]

@nwj94: Thanks! I was bored doing work so decided to write it hoping a few people would read. Yup, that is a market failure in itself - in the job market there are many many many 'tiers' of jobs that eventually makes the scale quite competitive - ie, top jobs will require that PHD in astrophysics or MBA from HBS, and from there downwards there are varying criteria, so people with different abilities have themselves competitively 'sorted'. In here, since there are only 3 difficulties, no such sorting occurs and the "top" tier market fails spectacularly. It also develops as a result of the poor gun/class balancing - resulting in skewed situations where a Carnifex user will more often than not outperform a equally leveled revenant user.

But as long as there is criteria to be met (who sets them? why are they qualified? why is that MD from Goldman rejecting me?) - there will always be some sort of subjectivity involved ironically, hence me writing this is hoping people dont get too personally offended when they do get rejected.

[VTiiiiiiiiii]

TL;DR

But yes.

[a ViciousFerret]

It was actually an interesting read.

[ol MISAKA lo]

A very good read. I think it covers most of the reasons why it happens. There is also the reason I noticed you didn't mention (idk how you would convey it tho) is the kicking of a character based on what class he uses (racism maybe?).

Too bad 99% of the players on here won't bother to read it, or just deflect it off, saying "it's only a game".

[derheidi]

I think he was just excited about a lecture and now has the urge to reflect it.
Any psych-graduate (...) would say: It has probably a reason, but meh. Flaming is more satisfactory.

(I hope I translated that correct).

[Maria Caliban]

If you have an unusual class/race or weapon load out, the best bet is to play one match with a common class and after successful extraction (or Wave 10), switch to your other one. Play a match with your level 20 Asari Justicar and then switch to your level 5 shotgun build Geth Infultrator.

[RamsenC]

I read it

The problem is when the "buyer" is so uninformed that they can't tell the difference between a lemon and a peach (FBW players). Like kicking out a properly equiped soldier or leaving a Cerb game hosted by an Asari Vanguard. That is their loss though. 

Modifié par RamsenC, 30 avril 2012 - 05:01 .

[Bob Garbage]

People also kick for what seems to be no reason at all. Just the other day I was in a lobby with two Adepts (Drell, original Asari) both with Carnifexs and well levelled (18 and 19 I think) but there was also this other dude who was a level 7 solider with a ridiculous loadout, so he was kicked....than immediately my Geth Engineer (level 20) with GPS VI was kicked out of the lobby, and this was just a geth farming run. The only 'good' reason for this is that they had friends coming or wanted an all Adept squad. Which if that is the case, why don't people just ASK these players to switch to another character? My N7 is almost 600, There's a good god damn chance I have oh you know, all classes at level 20.

[XCPTNL]

Wow, I actually read the whole thing... because it was quite interesting. Nothing really new but the comparisons were entertaining.

[Terraflare]

@ Misaka: Well, it's the same reason. Lets say for example a bad class = Quarian Engi or Drell Vanguard (Not going to argue with anyone over this). Good players in general tend NOT to use these, preferring classes like GI,AA, SI etc. So if you use one of that, again your segmenting yourself into the poor player market. Now, you MIGHT be a GOD with the Drell Vanguard, but again no one knows that.
As this happens over time, less and less people USE drell vanguard, until a point where the market of drell vanguards consist of only new players that dont know much. As such, when you pick it (even with N7 500), I can only conclude that you dont know much.

@derheidi: Not really, this lecture was a long time ago (very long). Just thought it was applicable to the case here. Psyche-grad would prob talk about the innate pleasure people get on imposing standards on others and feeling superior.

[Terraflare]

Maria Caliban wrote...

If you have an unusual class/race or weapon load out, the best bet is to play one match with a common class and after successful extraction (or Wave 10), switch to your other one. Play a match with your level 20 Asari Justicar and then switch to your level 5 shotgun build Geth Infultrator.

Yes this gets around the asymmetric info problem. Its like doing an internship - work your ass off for 2 months to prove you can handle the job, then slight oddities will probably be forgiven in the future. 

[Tangster]

Terraflare wrote...

@ Misaka: Well, it's the same reason. Lets say for example a bad class = Quarian Engi or Drell Vanguard (Not going to argue with anyone over this). Good players in general tend NOT to use these, preferring classes like GI,AA, SI etc. So if you use one of that, again your segmenting yourself into the poor player market. Now, you MIGHT be a GOD with the Drell Vanguard, but again no one knows that.

 
Gee thanks.

Terraflare wrote...
Psyche-grad would prob talk about the innate pleasure people get on imposing standards on others and feeling superior.

I'm gonna feel smug about being good with the QE.

[rtizz6446]

Way to over-analyze such a simple problem. Being an economics major, the topics covered above do not seem to relate to the problems at hand. The reason you get booted from a game is not due to adverse selection, but the fact that most people are so uneducated about the game that they do not know how to judge a players worth. I assume you are going to compare moral hazard to afk players in farming games. Ridiculous that you are trying to apply economics to ME3 multiplayer. While some of your points are valid, they are far too obvious to actually help anyone.

PS Why does this forum not have the word "multiplayer" in its dictionary.....

[eye basher]

Economics i just call it being a douchebag.

[capn233]

I read it.  But it is good the middle school demographic has a voice in the forum with their "tl;dr" posts.

More simply you can just think about it as risk reduction. If I am playing with two buddies, and they are sort of borderline, and a 4th random guy pops in with either a very low level or some oddball setup we are more likely to ditch him.  In this case, we were not reasonably confident that we would beat the missions without the assistance of a 4th player that played at least decently.

Alternatively, if I am playing with a couple guys that I trust to play well we are more likely to go even with a lower level 4th random player, or if the random has some stupid setup like a GPS and Widow on an AA.  This is because we are reasonably confident that the 3 of us can beat the mission regardless of what the fourth does.

The extreme is the completely random game.  Nobody really knows the abilities of anyone else.  From a risk reduction standpoint you want to see setups that you know have a reasonable chance for success on Gold.  You want to see character and weapon choices that make sense.  This even includes mods and equipment.  You do not want to see mods that do not work (ULM on nearly all SMGs, or Shredder on GPS).  When you see that it is a sign of a less experienced player.

And of course I am more likely to play things like Cowboy soldier when playing with solid buddies, and be allowed to do so, because of mutual trust and our assessment of risk.  You don't have that in a random game, so you shouldn't expect that people will tolerate goofy characters and loadouts.

Modifié par capn233, 30 avril 2012 - 05:16 .

[neurovore]

A couple of observations. First, people on this forum have reported surmised that N7 rating can be faked. Even if it couldn't, a high rating only means that they've played the game a lot. In fact, more than half of the players with 800+ rating I've met were rubbish. (Take this krogan with 950 rating who tried to melee a banshee, even if he got instagib-specialized every time.) Some people (such as myself) don't also like promoting. I think my N7 rating is half of what it should be. Overall, I'd consider 150-800 the safe range. Anything lower is less likely to have played more than a couple of characters to level 20. Anything more, and they're probably just not very effective.

Also, since weapon packs can be bought with real cash, people have weapons they haven't "earned". An heir might have a peach car, but not know how to drive it. (And what's wrong with lemons?!)

Not really trying to destroy your logic (though a lot of it is just common sense, to those with common sense), just offering more reasons why looking at crap like that might not give you accurate information. In fact, and I know this is revolutionary, but rather than just treating pixels as a job interview, you might actually ASK that person why they chose a nonstandard loadout or class. Their answer will most likely be much more illuminating than anything seen on the screen. Aside from having voice chat off, which might count against them.

Oh well. It's not like I play Gold anyways. I don't find it interesting, as it favors a strategy where you spend 80%+ of your time stationary. On Silver, you can at least move around and even solo a little (eh, I know people can solo Gold, but that doesn't really seem very interesting either. Just a lot of covering and kiting.)

Modifié par neurovore, 30 avril 2012 - 05:19 .