Greedy Goblin

Monday, June 6, 2011

The fact of the bad group

Tobold dismisses the common "this group sucks" claim as self-excusing: "Mathematical fact is that if you draw random team members, in 95% of the cases you get a team which is average, plus or minus 2 sigma (standard deviations)." He practically claims that those who blame a "bad group" are actually responsible for the failure and just shifting the blame.

Of course he is wrong, there are lot of bad groups, but the reason why "an average group is below average" isn't straightforward. It has two simple and one social reason.

The first reason is that good players tend to form guilds and fixed groups, removing themselves from the pool of random players. With 2 hard mode kills I'm a fairly good WoW player, up in the top 5%. Your chance to bump into me at a random Zandalari heroic is exactly 0 as I never go with a non full guild team (actually I don't go at all). On the other hand the worst players are rejected from every teams, so they are overrepresented in random groups where they cannot be booted instantly after their first stunt. The same applies to "/w for inv" guilds, where bad players are also overrepresented, as decent player wouldn't pick them with a stick. So while an average group is average, an average LFD group is below average.

The second obvious reason is that player skill is not following standard distribution. The question "did you read external sources about this game" is a yes/no question and I'm completely sure that the variation between the "no" and the "yes" group is larger than the intra-group variation of either groups. I'm also sure that the "no" group is more popular than the "yes" group. Something like this:
We have a sampling problem: since every blog and forum commenter read something about the game (the forum or the blog post), they all belong to the "yes group", so it's true that "average commenter is above average", therefore they are right when observing that "the average team is worse than me".

But even aside these, the third, non-obvious reason guarantees that most groups are bad. Actually this reason would guarantee that 80% of the 5-man groups are worse than you, even if everyone would use LFD and their skill distribution of players is normal. Vast majority of players are social, therefore feel uneasy being "jerk" with other players. Therefore the worst player of the group is rampaging free without being corrected or removed. While the wipes are technically on him alone, the observer is right to think that the group is fine with this performance, strike that, protecting the failer, and call the observer "being mean" with a "friend". Knowing this social rule, no one speaks up against the failer, therefore the group performance is limited by this guy.

You have 80% chance to not be the worst of 5, and in this care, you are right to believe that "with 4 clones of me we would perform better than with this group". Of course the other 3 not-worst players have the same right, so there are 4 players in every 5-man social group who are hindered by a "bad group".

Of course they are actually hindered by sociality, the inability to criticize, identify and fight against M&S. Where a social say "I was in a fail group" an asocial would say: "I was in a group with 2 OK and 1 good player, but we had an atrocious moron, so we sucked for 15 mins until one of us finally could initiate a votekick and get rid of that filth".


Grim said...

Was the "no one speaks up against the failer" supposed to have something to do with reality?
Then you really have not visited LFD. When someone fails, he gets called out on it. Often flamed. Sometimes people who don't deserve it get flamed and kicked but I can count my experiences with some failer actually being protected on one hand and still have fingers left. And I do a lot of LFD.

Torpid said...

Giant Wall of Text Incoming

About the 80% chance not to be the worst player part, you also have a 80% chance not to be the best player, therefore by your argument it would appear that one would also be right in believeing that "with four clones of me we would perform WORSE than with this group." Of course this is not the case, with Mr. Completely Average Joe as an example, he would have a 80% chance not to be the best or worst, 60% not to be best, 2nd best, worst, or second worst, etc. The net result of this is that everything cancels out and in the long run he is exactly dead average, i.e. "With four clones of me we would perform exactly the SAME as with this group." This is because unless Joe Average is the first or second best player of the group, replacing the the other four members with clones of him would result in a net loss 40% of the time, when Joe is worst or second worst, and breaks even when Joe is 3rd best and worst. (This is assuming that (worst + best)/2= Joe Average)

On a less technical and more anecdotal note, my personal experience is that the people running the Zalandari Heroics (henceforth ZHs) are generally above average, after the initial two days where everyone and their grandma's goldfish was trying to run the ZHs, it seems that the complete morons simply couldn't finish the heroics and stopped queueing for them, with the result that there's practically no instance in which I get an exceptionally bad DPSer. (With the definition of exceptionally bad being below tank DPS (which is usually around 10k, the tank's DPS, that is), I said less technical, not untechnical, heh.) In fact, the people in the Zalandari LFG adapted FASTER to the ZHs than did people in the Guild, though I had an admittedly small sample size. I.e. they started being able to reliable clear bosses (reliable as in only rarely NOT one shotting a boss) only two days after the ZHs came out, whereas the runs I had with the guild was still wiping at that point, but my sample size is naturally limited since I stopped doing so after the LFG provided groups capable to clearing ZHs in around 45mins so I can pick up Satchels of Exotic Mysteries. It should be obvious that I only ever did the ZHs as a tank, and the other two tierces may have different experiences.

Anonymous said...

I submit bad is subjective and you are using it as if it were objective. You have no impact on the definition of a good or bad ZG group because you are not in the population.

If the worse 94% of the players left then you would be about 16th percentile and would be considered a very bad player even if your skills did not change; the population changed.

Note also that you kicking M&S will help you but will have no impact on the population or what defines good or bad. Blizzard has hundreds of millions of dollars riding on making sure those players can play the game. So you can kick them out of your group; but other groups will need to accommodate them or Blizzard will need to change the game.

Squishalot said...

Gevlon: "Vast majority of players are social, therefore feel uneasy being "jerk" with other players. Therefore the worst player of the group is rampaging free without being corrected or removed."

That may have been the case before LFD was created. The use of vote kicking would suggest, however, that your theory is incorrect - if anything, 'scrubs' are removed with more frequency than pre-LFD days because there is no social consequence to kicking someone from another server.

The primary reason why people may not be being kicked is because of the timer penalties associated with kicking people from LFD groups implemented by Blizzard, not because people are social.

Gevlon said...

@Torpid: the team is more or less limited by its worse player. So it doesn't matter if you are world top player or just OK, you will equally wipe when ArthasDKlol pulls 2 groups.

5 OK players > 4 Method members + arthasdklol

dobablo said...

Getting poor players in LFD is great. You need bad luck to get a group that is incapable of clearing a heroic in sensible time, and if that happens you kick/requeue as required. The reason I enjoy getting poor groups is that I deal 50% of the group damage. If I was with equally skilled players we would clear the instance 10 minutes faster, but I wouldn't get my ego stroked.

Gevlon said...

@Grim: votekick won't solve the problem. Social people still doesn't criticize, the failer doesn't know he sucks until the "you are removed from the group" shows up. And HOW he failed is still unknown.

Squishalot said...

@ Gevlon: Grim is perfectly correct in what he is saying. Flaming for bad performance generally occurs prior to votekick, and poor performance is highlighted very clearly, generally speaking. Being told to attack skull / not pull / do more DPS / interrupt / not stand in fire, and being kicked soon after makes it clear what the person is being kicked for.

You're claiming that people in groups don't speak up. Grim and I are curious where you are getting your data from. Your PuG heroics?

Grim said...

Votekicking without criticizing is exactly what an asocial would do.

A social feels the need to help his fellow player, or at least the need to tell him he sucks, so that everyone sees how much better the social is.

Seriously, where the hell did you even get the idea that socials don't criticize? Most of the party chat in LFD is someone criticizing or flaming.

Espoire said...

I think the biggest, most powerful reasons for real bad groups are being limited by the weakest member or mechanics which magnify tiny differences in strength.

For example, the Lich King's defile heavily punished a single poor player, and similar mechanics would lead to a lot of bad random groups.

Likewise, in battlegrounds, when two random teams meet up, relatively minor differences in power and cohesion tend to lead to crushing defeat or victory.

Complaints about a "bad" team tend to be legitimate in the presence of mechanics similar to the above two, and excuses in other cases.

Anonymous said...

A lot of things depend of the encounters mechanics.

Depend what Blizzard design.

If you have mechanics where a bad player can wipe the group , your group is not stronger then your weakest link.

( encounter mechanic example : "You are the BOMB !!! In 5 sec you will detonate doing 1 milion damage to peoples in 50 yards area.Run away or you will wipe the group " )

4 Paragon players + Arthasdklol will wipe very hard at this mechanic.

Péter Zoltán said...

"He practically claims that those who blame a "bad group" are actually responsible for the failure and just shifting the blame."

This may be true for average or below average players, but not for others, really.

Anonymous said...

Tobold's original premise fails instantly, since the fact that most groups are within 2 s.ds of the mean ability does not mean that they are not bad, which is a subjective term not an objective one.

There is no reason why 95% (or even more) of WOW players might not be bad players. Bad does not mean 'worse than the median'. It's perfectly possible that a top raider would think every single PUG he played with (and perhaps every single 'The PUG' he played with) was 'bad'. The fact some of them were above average players is irrelevant. Similarly 'arthasdklol' might think (so far as he is capable of thought), that the players he was PUGing with were 'good', when in fact they were below average.

To Grim - your first comment is interesting as I am levelling a char at the moment, and my LFD teammates defended a low DPS moonkin to the hilt in Nexus, causing me to leave instead. It wasn't a borderline decision as to whether he was bad, his gear was full agility rogue gear causing his Starfires to hit for about 400 damage!

Gevlon wrote: '5 OK players > 4 Method members + arthasdklol'. I don't believe this is true in any situation, not heroics, normals, 5 vs 5 arena, or anything. Perhaps you underestimate how much better the top players are than OK players?

Sten Düring said...


The reason Blizzard implemented the timer and DR on votekicking was definitely NOT because people are afraid of kicking other players.

Your sample size of troll heroics is exactly zero as per your own statement.

The bads get kicked as soon as the timer allows in a troll hc.

I don't know what it's like in lower tier. Probably not as harsh. But the troll instances cost you an individual 100+ gold if they turn into wipe-fests. You just can't afford keeping a moron.

I'd be surprised if LFD doesn't generate a group that is on average better than the non-hardmode guild-run. This because of the 15% extra topping on your WoW-pizza.

Grim said...

Yes, that happens. It has happened to me as well, but such occurrences are extremely rare. That one time it happened to me was up here as a moron a month ago. Since then I have run at least 30 LFD heroics, never encountering such a group again.

And I don't even remember any other specific time I would had seen that happen.

Jacklepsychomaniac said...

TO Anonymous who said (4th comment):

I submit bad is subjective and you are using it as if it were objective. You have no impact on the definition of a good or bad ZG group because you are not in the population. ""

If the worse 94% left, Gevlon still wouldn't be bad, he would simly not be in the top 5%. You do not seem to grasp that by "bad", Gevlon means some guy who will fail at encounters?

While Gevlon's guild wouldn't be as high as it is now if 94% left, it wouldn't affect their ability to clear raids fast. Plus, they are not a hardmode guild.

Gevlon said...

@Grim, Sten Düring: anonymously pressing a button is different than speaking up. It won't fix the bad player, and you get another player who also have the chance to be bad.

Practically the same bad people are rotated between groups.

Michael said...

Hi Gevlon!

I'm always amused at your ideas of how socials think. Why on earth are you suggesting that social players would put up with incompetence and failing play?

I am totally a social player. I do crafting and enchants for people in my guild for free and they do the same for me. I spend a lot of time helping to gear up and carry along the weaker members. I want the people I play with to primarily be fun and pleasant, and also strong players, but the pleasant part is more important. I grats people for their achievements, I spend way too much gold on pets, mounts and reputation grinds, just for fun. I don't pvp at all, cause I see it as mean trying to kill other players. I like to make funny ascii faces on my chats to show my mood and to indicate humor or sarcasm. I always run around with my happy sunflower out. Or penguins.

All of these things, I believe you would dismiss as social nonsense. Now here's the big difference. Everything above gets cut off immediately at the guild level. Outsiders _know_ that my guild is elitist and snobby and know not to rely on good will or whatever nonsense to make any points with us. Our recruitment criteria are very strict. We actively compete with other guilds on the server. We think very little over completely screwing over outsiders for the benefit of one of our own. We're doing pretty well on the raid scene, at 7/13 heroic modes.

Sooo, from the inside we're a classic social guild, carrying peeps and valuing having a fun time over doing well. From the outside we're an elitist, asocial bunch of assholes. Where would you classify us? As socials or asocials?

Just trying to show how you shouldn't say 'socials believe blah, so they act like morons in this case'. :P There are lots of different types of socials.

Kobeathris said...

Gevlon, I think you are correct for WOW random dungeon groups, but I believe Tobold was talking mainly about World of Tanks teams, which is a completely different matter. You can't bring a full group to a random battle, so there really isn't a selection bias where better players can remove themselves from the population. Sure, they can go do clan wars, but I can't imagine someone ONLY doing clan wars, they don't happen often enough.

Anonymous said...

"it seems that the complete morons simply couldn't finish the heroics and stopped queueing for them"

Or, they have learned the tactics, and have geared up. They could've learned

* From patient people (experience via trial and error)
* By looking up the tactics (video, written text, or party chat)
* Or they've geared up from more random HCs, raids, VP, ZH.

Or a combination of these. IOW: the content is auto-nerfed over time. 1 and 2 are from gameplay. 3 is from gear. How much % is from improved gameplay, improved gear, or certain (bad) players simply not doing ZH anymore I don't know.

I also wonder how much % of people have queued as tank or healer for the satchel while generally not being ready for tanking or healing [that] content for whatever reason (e.g. no experience as tank/healer, gear issues, not knowing tactics to take responsibility of that role). Have they quit, or have they learned?

FWIW, I quit doing ZH. I don't need the gear, and I don't need the VP. I've farmed some satchel/BoE on some tank and healer characters, and have sold the contents on AH, giving me some nice profit due to 'pet collectors'. The prices of these BoE objects have now heavily deflated. YMMV, especially depending on realm's economy.

Anonymous said...

"@Torpid: the team is more or less limited by its worse player. So it doesn't matter if you are world top player or just OK, you will equally wipe when ArthasDKlol pulls 2 groups.

5 OK players > 4 Method members + arthasdklol"

As a rule of thumb I can agree with this. Otherwise, really depends on role, fight, and the quality of the other players. There is Cataclysm content where you can chainpull or not CC while previously you could not. Not every group can do this (e.g. requires some interrupting, good gear, etc). Remember: someone in full ilvl 346 is overgeared for random heroic dungeons, someone in full ilvl 353 is overgeared for random ZH, and someone in full ilvl 359 is overgeared for normal raids. The 4 Method players can be on badly geared alts, but theres a higher chance a few of them are carrying the others since else there is no incentive to do random ZH, sans guild rep. By gear alone will these people perform better than arthasdklol no matter how shit player he may be. The reason why people are in random dungeon also affects their social behavior.

..but I disgress. Example of Zanzil fight: if some DPS dies on Zanzil, you can still down him, although this gets harder if you have low DPS. I find in such situation you can seperate a good (or perhaps experienced) player from an average one. A good tank can survive the fight for quite some time using CDs, some DPS can OT the fight, some healers can do some DPS. It just depends on the group's DPS, individual's skill, the communication, and even the group setup. For example, you can 'zerg' the fight instead of killing the zombies or the big add. This isn't always viable, and requires some coordination. Perhaps even kiting. Good players adapt, bad players don't. Just because it is PvE does not mean that there will never be anything to adapt, and this is THE aspect of PvP playing I respect good PvP players for: they adapt to situation, and have learned to do so. If he wants to PvE with me I'd give this guy a chance over someone who has proven to be an average PvE player.

Some game mechanics also have overpowered aspects such as spellstealable spells. If arthasdklol is skilled enough to cast Dark Simulacrum while Zanzil casts Zanzil Fire, he will be able to do good DPS on boss (my tests indicated 400k with 1 GCD on boss), and perhaps even on adds if he is skilled enough. Not rocket science.

A mage can spellsteal Remedy. If he is assigned for this task, maybe he cannot do for whatever reason (e.g. frozen, or less valid reason such as tunnel vision). Whatever the reason may be, the enhancement shaman can purge it. A good enhancement shaman does this for the best interest of the group, while an average one has tunnel vision and does not bother or notice Remedy. Ofcourse, the mage (or other players not freeing the frost tomb) should not have put the raidgroup in this position, but that does not mean that the group should wipe because of that. If I'd be raid leader, I'd assign people on backup for certain tasks whenever possible to give them responsibility. Unfortunately, some seem to think this means they can slack "cuz theres backup lolol", but thats the type of attitude I don't want in a serious raid. We work as a team, there are priorities, and we help each other out when necessary. End of story.

"A social feels the need to help his fellow player, or at least the need to tell him he sucks, so that everyone sees how much better the social is."

The new vote kick CDs introduced in (IIRC) 4.0.3 force players to not kick people so quickly because when you agree to a vote kick you get penalized for it no matter why you agree (someone else started vote, DC, AFK, getting boosted, rude, etc), and no matter if the vote kick succeeds. We are _forced_ to be as you put it social, or we're pushed to leave the group.

Coralina said...

For someone that doesn’t run in DF pugs, I think Gevlon is surprisingly accurate in his understanding of how it works.

I am an ex raider who pugs a lot and I think firstly you guys are very lucky in your pug groups. I don’t get groups with one bad player; it generally tends to be three. I find damage output typically varies from 4 to 6k DPS with my tank usually outputting three times the overall damage of the lowest dps. On my healer I can see quite clearly that not only do they fail to avoid mechanics but when I have spoken to them it is obvious that they don’t even use a boss mod and had no idea that a mechanic of some type was happening.

You don’t really see a couple of 5k DPS’ers calling out the 4k DPS’er for being bad so group discussions don’t often take place. Given that I can’t kick them all I just quit. I find more and more that people are simply choosing to quit without saying a word rather than using the restricted kick system let alone attempting to explain where players went wrong – it would be futile as you only receive abuse in return and there is a never ending supply of bad players that you get bored of educating.

That is probably one of the most marked social changes I have seen going from Wrath to Cataclysm. There is no longer any attempt to try and rectify groups; players have learned to cut their losses early and quit. It is also true that five ok players are better than four good players and one M&S; the worse player does seem to determine the overall outcome but when you have three or four M&S you better get out before too much time/gold is lost.

I think other posters have called it correctly in terms of the two new heroics being blighted to a much lesser extent by M&S. I personally find them a lot easier to heal than the original HC’s and they have a far higher success rate than say a GB or Stonecore HC.

Perhaps my most interesting observation (feel free to figure out the explanation) is this: I quit in February and started playing again a couple of weeks ago. In that time the success rates for HC’s appears to have fallen dramatically. I’d say the average damage output and crucially even the knowledge of mechanics is also far lower than it was before my break. Strange how the official forums describe those HC’s as a “twenty minute nerfed to the ground faceroll” yet in reality the failure rates are far greater than they were back in December or January.

Oh and before you put it down to the M&S finally hitting 85; I don’t believe that is the cause. Most of them hit 85 over the Xmas holidays.

Anonymous said...

Tobold missed the point about 'small sample size and tank/dps/healer and ability thresholds.
If you assume that:
(a) a group is 'bad' if either the tank, healer, or 2 out of 3 dps is unable to complete the encounter
(b) there is a continuous ability threshold, with a fixed cutoff for each encounter
(c) that only 20% of players fall below the threshold for a typical encounter
You will find that roughly half of your groups are 'bad'.
Even assuming normal learning curves, assumption (c) will be valid. Eg, I tanked the troll heroics about 10 times with a freshly minted 85 warrior using a mix of BOE and pvp gear. I didn't hold AOE aggro well at all and was kicked in my first two runs. The following 8 were highly successful (Threatplates), but still put me at the 20% failure rate. Afterwards, I stopped because I'd seen everything interesting in the dungeons.
That said, in LFD, the only really terrible groups I've found involved healer/dps teams who were unkickable.

Sthenno said...

For heroic dungeons, if you think that 4 Method players + bad player is worse than 5 mediocre players then you have never seen a good player play the game. While we were greens I exclusively played with my guild, but after the first couple of weeks I would have been very confident four-manning a heroic, and three of us would be able to queue together to carry anything but AFKers.

I can't think of any fights in the original heroics where you can't afford a weak link. Again, in greens you can't because there are a few damage races that you just won't win, but four people in 359s do more damage than five people quest gear.

Raid encounters often have mechanics that wipe the raid if a randomly selected person messes up. Five-player dungeons do not.

Tobold's point, I believe, was this: You can't credit yourself for all of your successes and blame the team for all of your failures. If you are fatalistically convinced that winning and losing is just determined by the team you draw then you have to credit the team with your wins as well as your losses. I personally thought it was always more useful to concentrate on what I could do for the team rather than what the team could do for me.

Sten Düring said...


I grind my weekly 980. Not because I need them, but to have something to do.

My average troll LFD consists of one decent healer, two decent dps and either one good dps or one horrible dps.

I'm not doing the trolls to chat around. I have /trade for that. A horribad gets kicked, and more often than not I don't even have time to start the kicking.

While it happens that we get a new moron it's not very common.

At the ends of the bell clock are the horribad runs. I just leave those and accept 15 minutes deserter debuff. There are also the very good runs. No talking, just fast killing.

Both occur very seldom.

Now, this is anectodal, but my sample size is a lot better than zero. With zero sample size you're not estimating anything -- you're making wild guesses.

Masterlooter said...

I'm not argueing with the content of your post, but the context is very misleading.

Tobold is referencing World of Tanks, where there is no such thing as a "guild run". (There are clans, but it's an entirely different type of game.) For all intents and purposes, in WoT there is only grouping with random strangers. At most, you can have 2 other players group with you, still leaving 12/15 of your teammates be random (80%).

There is no kicking from matches, and aside from players being obviously AFK, or players getting 5 kills in a match, there's no easy way to distignguish poor players from average, or even good ones.

So in reference to Tobolds post - yes, pulling 14 random players out of ~20,000 you will get a TEAM of combined average skill more often than not.

Fernando said...

I cant be playing the same game as most of you guys, I have tanked, healed and dpsed in Wrath, i have done the same on the 85HCs and troll ones and I must say that I have gotten a very minor % of idiots on LFD, and when I do get them, they usualy last only for 15 mins, the only exception that i can remember was BRD maybe after 3weeks of cata coming out, of a hunter doing around 5Kdps, and by somw strange reason no one could kick his ass, even so it was completed without too much stress ( had to skip beuty though).

Fex said...

from what i see in my pugs, the critique gets no further then DO MORE DMG NOOB, or further L2P comments. Proper critique would consist of a complete check on rotation, pointing out spec errors, glyph errors. Helping with proper enchants. Rotation help. Addon help. The only sensible critique that is of any value to a bad player is limited to don't stand in crap and attack skull. And possibly with exceptional good groups. Interrupt more, and don't break CC. Any other information i've never seen come across chat. And even though i'm capable of providing that i don't in LFD either. 1 takes to long, 2 is not appreciated by bad player, 3 is not appreciated by group. End result bad player gets told to do more dps, sometimes to research, and kicked. This isn't helpfull and does not consist of proper critique. Flames with L2P noob don't count in this aspect. So in essence, a pug group does not provide good critique ever. At least i have never seen it in the close to 300 pug runs i did.

Anonymous said...

Having an 80% chance of not being the worst does not mean that you are hindered by your group and that four clones of you would out perform your current set up. You actually have a 40% chance of being below average in a group, a 40% chance of being above average and a 20% chance of being average (in a 5-man). So there is a 50% chance of being hindered by your group and a 50% chance of being pulled by it.
The obvious Gevlon response is, that having one player worse then you will guarantee failure, because they will pull the whole instance, dooming the group to endless wipes. This is a completely clouded view of reality from someone with little recent expierence in LFD's. In 90% of dungeons the worst player is just a new person who hasn't geared up yet and doesn't do much DPS. In this case, the best player is normally just has high on the positive side and so the two equal out.

The reason certain players, mainly elitists and a-socials, feel that their is such a high percentage of moronic players is because of there aggression towards their peers. If you are a completely capable player and someone in your group calls you out because you are trying a new rotation, forgot to cast a spell the second it popped, or used a the shortform of 'you' then your tendency is to be pissed. (Understand that im talking about the majority of players, obviously a-socials are more understanding about being called out on minor details). If the person doesn't mind losing 15 minutes of their time then the easiest way to combat the elitist aggression, is too screw over the person's dungeon by simply acting like a moron. Since an elitist is pissed about a non-filled gem slot, they will be ready to slit there wrists when you start dancing around the instance naked. Obviously this seems extreme, but aggression has a very large effect on people, and cause them to do rash things. Which has been exhibited countless times by some of Gevlon's posts about what he calls M&S.

End of Rant.

Grim said...

"Don't stand in X" and "Interrupt Y" is all that is needed almost all of the time.
It is very rare that the group's dps is so low that it is impossible to finish the instance.

Anonymous said...

Fex, this is because there is no social relationship between the LFD people. They will probably never meet again after the run. They are temporary allies. Why invest time in someone you will never meet again?

In LFD, from day 1, I see people asking for tactics, and they are helped. I see a random person getting free advice, and it is not accepted because "mind your own bizniz, I know my class."

But why on Earth would one want to check what glyphs some random PuG is using, or other advanced yet core class aspects? You'll never meet that person again. If glyphs were the only problem you could probably still carry the person. Some errors cannot be fixed with little effort. The easier solution is either to kick the person, or leave the group.

stubborn said...

I agree with Gev on most of his points here. There should be a more formalized Vote to Kick "reason" window, and the person being kicked should see the comment. Some people still won't fill it out, but I would. I hate it when my buddy VtK's someone and just says "moron;" I want the guy to know that we kicked him because he couldn't heal a trash pull with a 359 gs tank who used all his cooldowns and dps doing 16k each. There's a problem there, and the guy needs to know.

As for not criticizing, that, too is a problem. I won't be a jerk to someone, but if they're making the same mistake twice in a row, they're going to hear about it. I'll be polite, but I'll also be direct. In a recent ZG (I freaking hate the Z instances), we had a fellow who kept standing in green stuff on the snake boss. We had explained the fight to him since he'd never been there before (groan, but okay, let's give him a shot). After his first death, we reiterated the list of things that were bad and told him that he'd stood in one. After his second death, we reiterated it again and told him that if he could not get out of the bad, he'd be replaced. After his third death, he was replaced.

I'm generally a nice guy. I am always willing to help a PuG who wants to be helped. That fellow, though, couldn't perform a basic function and had to be removed so we could continue.

Azuriel said...

The "bad group" generalization is a bit vague as you are either in a group that finishes the instance or you are not. Of the successful groups, you can segment them out like "barely made enrage timer" to "barely had time to loot before breezing through." If you come to expect the latter then, yeah, most groups are bad (to you).

If you use LFD with no expectations at all other than successfully clearing the instance, I would say (a lot) more than 50% of groups are successful. We don't know the exact figures, but Blizzard did admit a less than 50% success rate pre-15% LFD buff/heroic nerf. It is likely much higher now.

Squishalot said...

@ Fex: How is "GET OUT OF FIRE N00B" any different from "Fight at flags, morons!"? Because the latter is what Gevlon would use.

Anonymous said...

Your post reminds me of a phrase, I don't know the exact wording, but it along the lines of "a group is as strong as their weakest link". I think this is very true for wow pugs, because as you point out most of them are at best average. If you took the skill level of the population the curve would probably be left leaning and have a sharper gradient on the right hand side.

Also Gevlon I see a big problem in your graph in that there is a population of players or both have and have not read up. Since the graph represents a yes or no response the two curves should never touch, or you get what I just pointed out.

Anonymous said...

p.s I'm not sure you can even use a bell curve to show the results from a yes/no question. you would be better off showing it as a bar chart, if you wanted to highlight the differences. the current state If all i saw was your graph i would not be able to know what it shows.

Also you claim that player skill is not following standard distribution. I disagree and think that if you came up with a good way to measure player skill the results should reflect a bell curve, although i think i may have misinterpreted what you meant b that statement.

Fex said...

@squishalot did you even read my comment?

I was referring to the crap that the average lfd member yells at another average LFD member. Not what i would say myself.

As for glyphs not being a big problem, Balance druid not having proper glyphs loses them 30% on each tick of Insect swarm, causes moonfire to fall off a lot faster then if it were glyphed, Gives a major Gimp to wrath. All staple spells. Easily make up 20 - 40 % of a boomkins DPS. Don´t stand in the fire really doesn´t `fix` all the wipes out there, And Do more DPS really doesn´t tell a player just `how` to do that. Unless he´s idely playing tetris at the same time as playing wow. Telling someone to do more dps will not make that happen.

Squishalot said...

@ Fex: That is what the average LFD person would say. Other examples:

"you pull, you tank"
"less QQ, more pew pew"

These things provide information as to why the failer has failed. There's plenty of information in the typical LFD flaming.