97 replies to this topic

[Abivard]

How utterly boring and predictable will the 3/3/3/3 games be?
Will they be changing lances to 3 mechs so we have 4 lances?
Won't 4 man groups lock up the MM?
Will 4 man groups also have to abide by a 1/1/1/1 make up?
Just how many people want to have the same mix of mechs every drop?

[Abivard]

Who will decide what Mech I can pilot? It sounds as if I am not to really be allowed that choice, instead I must conform to others wishes by compromise.

[Craig Steele]

Abivard, on 07 March 2014 - 11:29 PM, said:

Who will decide what Mech I can pilot? It sounds as if I am not to really be allowed that choice, instead I must conform to others wishes by compromise.

No, you can pick anything but MM will arrnage the players around that strucutre.

Some people are hedging that this might see failed to finds if you want to play a mech class that is really popular as the population cannot support the structure (eg, 50% of people want to play Assaults, so 33% of them get failed to find as only 17% can be matched with the other 50% of the player base 17% each Heavy, Medium and Lights)

But alot will depend on population and demand. It might be a problem when the release a new mech for sale I guess, but who knows?

[Toong]

HammerSwarm, on 06 March 2014 - 01:36 PM, said:

This is a contentious topic, but it's an upcoming feature and so I hope this is the place to discuss it.

The argument against strict tonnage as the matchmaking criteria of choice was that people couldn't play the mechs which they paid real money for. That if you had just bought a banshee and you called up your friend and he had bought a banshee that if they limited teams of 2 to 160 tons you couldn't play them together.

Additionally I think it was said that: if you queue up to play in 12 man match you might be the guy who is told you only get 90 tons for your assault, not 95. or 45 not 50 etc. and this would lead to people unhappy that they couldn't play their money mechs.

Etc.

-----

Has the idea that 3/3/3/3 creates a situation where if 1 class isn't queued as much as others that no matches may start and no one might get to play been discussed?


Has any of this been addressed? Am I jumping to conclusions? or is this like a thing that could happen?

I don't think you're jumping to conclusions, but neither do I think this is an issue, and here's why:

Your fear of people being left out of the queue would be correct, if everyone was matched simultaneously to one, and only one game. If we took everyone online at a given moment, and then paired them up under the 3/3/3/3 model, and the distribution wasn't perfectly equal, then yes, there would be a sizable amount of people left without games.

HOWEVER:

there are thousands of games running at any given moment, and there are anywhere between dozens and hundreds (without hard population numbers, it's hard to accurately guess) of new games being matched and launched each minute. In addition, queues are treated on a first come, first serve basis. The longer people wait in a queue, the more relaxed the matchmaking rules become, which has the end result of favoring people who have been waiting longer.

Because of this, there is a constant cycle of people being switched in and out of games. One set of people leave a game, and another set of people enter. Because of the preference for people who have been waiting longer, the entire playerbase can be pictured as a giant "Conga Line" of people in various stages of play. Most people will more or less stay in the same relative position to each other. So if Frank starts a game three minutes before Cassandra, and they both take the same amount of time to set up their 'mechs and play a full match, Frank and Cassandra will probably stay more or less three minutes apart in the "cycle."

The reason this is important is because adding the 3/3/3/3 system, in effect, splits the one giant line of people into four smaller lines. As you accurately point out, obviously not every line will be exactly the same size. But since it's a cycle of many games starting and stopping, the four lines don't have to be. Since the game's matching people up 6 at a time from a cycle of thousands of players, any one match has no issue being made.

SO.

While it won't make games fail, it will have a different effect: since the lines are all different sizes, but all lines are in equal demand, some lines will have to move faster to keep up with the matchmaking. The end result? People playing a minority weight class will have shorter queue times than people playing in a majority weight class.

Because the population of the game is so large, the actual difference in weight between classes will probably be only a few seconds, small enough to be invisible when compared to the average variation in queue times anyway. While it's possible for the 3/3/3/3 system to cause matches to fail, this would require a hugely unbalanced class distribution. I know popular opinion is that 75% of the population play assaults and the remaining 25% play heavies, but I suspect that the actual numbers are much more equal. In addition, there are two reasons why I doubt this will ever be a problem: First, a lot of people play heavy 'mechs because they feel they have to, and once class limits are imposed, that actually frees up a lot of people to play the lighter classes; and second, the longer the disparity is between queue times, the more likely people are to switch to a more "in-demand" class.

IN SUMMARY:

The 3/3/3/3 system won't cause matches to fail, though it might tack on a few seconds to your queue times if your weight class is extremely popular at the moment.

[HammerSwarm]

Craig Steele, on 06 March 2014 - 05:59 PM, said:

I get the numbers you posted sure but those calculations all hinge on 1m players all pressing the button at exactly the same moment and the MM doing that calculation instantly and every match running for the exact same length. The reality is the population will trickle in and trickle out of MM and with search time and variable game lengths so I can theory craft that there will be zero impact.

If it's not out why are so many people so critical of it. What's the source of information that is causing people such grief?

You can't get over the math of it. If there is a pool of any size (any number number large enough) the weight class with the lowest representation will dictate how matches get launched. it's true that I assumed all launches will take place simultaneously for the purposes of dramatic presentation and simplicity.

The fact remains that the weight class with the lowest representation will still act as a gate keeper for all matches to be launched. So those left on the outside might be a rolling number but in a population of 1 million and just one class of lower representation(5%) you could still have 1/5 of your population on a rolling 'fails to find match'.

If the player pool were static and 3/3/3/3 were representative of the weights looking for drop it'd work great. I'm hoping to identify potential problems before launch so we can avoid a catastrophic failure.

Story time, Look up the ford pinto, might have been before your time, anyways they identified a design flaw late in the process and decided that it'd be too costly to fix, and the cost of dealing with it afterwards would be less than the cost of fixing it. That's because the further you get in a design process the more expensive things are to fix. If it's on the drawing board you just add something in. If you've launched the product you could be looking at a costly recall and fix.

I'm trying to help make this game better. What are you trying to do?

Edited by HammerSwarm, 08 March 2014 - 09:03 AM.

[HammerSwarm]

Toong, on 07 March 2014 - 11:48 PM, said:

Because the population of the game is so large, the actual difference in weight between classes will probably be only a few seconds, small enough to be invisible when compared to the average variation in queue times anyway. While it's possible for the 3/3/3/3 system to cause matches to fail, this would require a hugely unbalanced class distribution. I know popular opinion is that 75% of the population play assaults and the remaining 25% play heavies, but I suspect that the actual numbers are much more equal. In addition, there are two reasons why I doubt this will ever be a problem: First, a lot of people play heavy 'mechs because they feel they have to, and once class limits are imposed, that actually frees up a lot of people to play the lighter classes; and second, the longer the disparity is between queue times, the more likely people are to switch to a more "in-demand" class.

What is your assumption of player pool weight composition when you make this assumption? I started another thread in feature suggestions to get a handle on what mech class people play most, their go to, so far it's like 1/4/6/7 Which I would guess is representative of most of the ques based on the matches I see. Maybe it's a little skewed by a low sample size; we'll see.

If you have 60/120/150/180 how many matches can spawn under the current system? 21 matches with 6 people failing to find match.

How about with 3/3/3/3? 10 matches with 270 people failing to find match. Now what amount of rolling matchmaking fixes that?

So if you have 10% of people from each class change on the next go around. and they know exactly which class is needed though some feedback mechanism also not outlined in the initial reveal of this feature.

0/60/90/120 becomes 27/54/81/108 creating 4 more matches, idling 174 people.

If it's just a queue and you stay in it until matches form I can tell you that the wait time for the assault class will be over twice what the average match length is for my totally made up sample population.

And that's ignoring an ELO, because we'll have 3-4 buckets of players in 3 game modes The Idea that indefinite ques and rolling matchmaking somehow fixes anything is laughable.

You only have so many matches possible no matter at which point in a window of time they launch if the composition underlying group of available players doesn't match the ridged structure of 3/3/3/3 you're going to have friction. The more skewed that group is. The worse it gets.

I'd like to know they have thought of this. So I made a thread.

Edited by HammerSwarm, 08 March 2014 - 09:01 AM.

[IceSerpent]

Void Angel, on 07 March 2014 - 10:53 PM, said:

First, your #1 point proves my own. There is no reason that a person who has 4 or more friends can't play with friends unless they are the odd man out. Yet, this gets said over and over again. Since only one out of every five people can ever be frozen out in this way, and not all players are desperately pining to play with their buddies (sniffle, sniffle, cue the tears,) how can you claim that PGI is "not letting people play with their friends?" The more you analyze this argument, the less reasonable it is shown to be.

You're missing an important point - people want to play with all of their friends currently online (up to the maximum team size obviously, but that limit is unavoidable due to technical issues). A situation where you have 6 unit members online and you have to split into 3/3 or 4/2 just because PGI messed up the queues is equally unacceptable. "Letting people play with their friends" = "any number of players up to the maximum team size has a place (queue) to go to".

Quote

2. The matchmaker is not set up to match a team of exactly X players with exactly Y players. Leaving aside the effects of such a restrictive setup, this is not the claim that is being made. What's being said, I kid you not, is that a group of 5 players on one team and 11 players on another team would be balanced at any given Elo, because "they had to play as a team to get to that Elo, while the solo players got there on their own," as one poster put it. Do I really need to explain the obvious fallacy? Do you even have the data needed to predict the rate of match failures of such a hypothetical matchmaking setup? The answer rhymes with "go."

We all know that matchmaker is currently FUBAR, that's not what we're talking about. We're talking about how it should work. You also misunderstand what is being proposed - we're suggesting much lower range of group size variance, maybe up to 4 (i.e. 8-man on one side vs. 12-man on the other at most). Nobody is suggesting matching up 5v11, although theoretically it can be done if the team with 5-man gets some other advantage - higher Elo or higher BV for example.

Quote

3. Speaking of fallacies, your stated argument is invalid on its face. It both begs the question and is false to fact. The Devs have given reasons why they're going to do what they're planning - it's not "just because" they're "unwilling or unable to split the queues properly." So, that part is simply untrue, even before we get to your fallacy of begging the question of what's "proper." I have yet to see a convincing argument as to why going to the expense and trouble of fragmenting the player base (which will skew the data they use to try and balance things,) is worth having 5-11 man teams. So that's a classic argument based on facts not in evidence - the textbook form is, "When did you stop beating your wife?" On that note, I guess we can add a third fallacy - your attempt to shift the burden of proof to me to refute you when you haven't actually offered any proof.

The reasons given by devs are simply not true, as in "objectively incorrect". It doesn't "fragment" the playerbase any more that adding a new game mode does and for some reason I don't see people complaining about addition of skirmish "fragmenting the playerbase". Right now we have 2 completely separate queues (PUG and 12-man). What is being proposed is to also have 2 queues - solo only and group, with solo players having an option to queue for solo only, group only, or both. Groups of any size can only go into the group queue. How does it fragment the player base? If anything, it decreases the fragmentation because solo players get the ability to enter two queues instead of one. Lack of objective reasons for the current mess leaves us with two options - devs either don't know how to do it right, or they don't want to do it right.

[Thunder Lips Express]

as a great white whale, i will now not spend real money on another hero/package until i find out how this will work. i don't mind playing any weight class but i don't want to not be able to play one that i spent real money on.
i just hope that in practice it works better then it sounds

[Novakaine]

Tonnage limits are only valid way for a truly fair drop model.
If the limit for said match limit is 650 tons.
Then both side will get 650 to do with as each side choose.
And of course I'm talking about private matches.
That insipid 3/3/3/3 should never get out of the barn.
Oh wait minute......no teams left.
PGI hates teams
Never mind forgot where I was.

[HammerSwarm]

Ecliptor, on 08 March 2014 - 09:46 AM, said:

i just hope that in practice it works better then it sounds

Me too.

[Toong]

HammerSwarm, on 08 March 2014 - 08:59 AM, said:

What is your assumption of player pool weight composition when you make this assumption? I started another thread in feature suggestions to get a handle on what mech class people play most, their go to, so far it's like 1/4/6/7 Which I would guess is representative of most of the ques based on the matches I see. Maybe it's a little skewed by a low sample size; we'll see.

If you have 60/120/150/180 how many matches can spawn under the current system? 21 matches with 6 people failing to find match.

How about with 3/3/3/3? 10 matches with 270 people failing to find match. Now what amount of rolling matchmaking fixes that?

So if you have 10% of people from each class change on the next go around. and they know exactly which class is needed though some feedback mechanism also not outlined in the initial reveal of this feature.

0/60/90/120 becomes 27/54/81/108 creating 4 more matches, idling 174 people.

If it's just a queue and you stay in it until matches form I can tell you that the wait time for the assault class will be over twice what the average match length is for my totally made up sample population.

And that's ignoring an ELO, because we'll have 3-4 buckets of players in 3 game modes The Idea that indefinite ques and rolling matchmaking somehow fixes anything is laughable.

You only have so many matches possible no matter at which point in a window of time they launch if the composition underlying group of available players doesn't match the ridged structure of 3/3/3/3 you're going to have friction. The more skewed that group is. The worse it gets.

I'd like to know they have thought of this. So I made a thread.

My assumption for pool weight composition is something like 15/20/30/35 percent. People love to remember "that one time" they got that game with four highlanders and four stalkers, because it reinforces popular opinion that the game's turned into AssaultWarrior Online. They also love, when pressed, to adamantly defend this position by taking the memory of those occasional (and memorable!) mismatches and multiplying them by like fifty. I'm convinced this isn't actually the case.

Don't get me wrong, tonnage is a major issue. Probably the single biggest problem with matchmaking right now. But I'm a medium pilot. I've long since gotten tired of people telling me mediums are dead, to the extent that I now make a point to keep track of how many mediums there are in my games. Over the past four or five months I've been watching, I've seen 4-5 mediums in every match, along with 3-4 lights, with the rest being a mostly unpredictable amalgam of heavies and assaults. There's still the occasional game where the numbers are way off, like a team of half lights and mediums versus a full team of heavies and assaults, but they're rare.

HammerSwarm, on 08 March 2014 - 08:35 AM, said:

You can't get over the math of it. If there is a pool of any size (any number number large enough) the weight class with the lowest representation will dictate how matches get launched. it's true that I assumed all launches will take place simultaneously for the purposes of dramatic presentation and simplicity.

The fact remains that the weight class with the lowest representation will still act as a gate keeper for all matches to be launched. So those left on the outside might be a rolling number but in a population of 1 million and just one class of lower representation(5%) you could still have 1/5 of your population on a rolling 'fails to find match'.

If the player pool were static and 3/3/3/3 were representative of the weights looking for drop it'd work great. I'm hoping to identify potential problems before launch so we can avoid a catastrophic failure.

Story time, Look up the ford pinto, might have been before your time, anyways they identified a design flaw late in the process and decided that it'd be too costly to fix, and the cost of dealing with it afterwards would be less than the cost of fixing it. That's because the further you get in a design process the more expensive things are to fix. If it's on the drawing board you just add something in. If you've launched the product you could be looking at a costly recall and fix.

I'm trying to help make this game better. What are you trying to do?

He's also trying to make the game better, by trying to make the forums a place without toxic hyperbole. While you've handled yourself nicely for the most part in this thread, this feels like a senseless attack on someone just because they disagree with you. Also, your Pinto story has nothing to do with the topic at hand. If you're trying to make a point, please tie it in to your argument.

Anyway.

The lowest-played class will, as you say, play gatekeeper, but the effects of that chokepoint will be felt based on the relative size of your class and the smallest class. In easier terms, the effects will probably only be felt by the most played class.

We only have 100% to work with, so the numbers can only really skew so far. Even if only 15% of the playerbase plays lights, for instance, the remaining 85% is split between three classes. The actual difference is only an average of 18% or so, with a wider difference between lights and assaults. Mediums would feel almost no issue, being around the same percentage as lights. Heavies may experience some slowdown, but assaults would take the brunt of it. Thanks to the 3 minute rolling matchmaking though, any one game would have a minimal chance of actual failure even though there's twice as many assaults running around as lights.

Yes, games would fail -- if pressing 'launch' made a single match attempt. But the matchmaker continuously tries to make a match while that button's pressed. Assaults will fail more often, just as you fear, but all those failures are within the context of the 3 minute matchmaking window, rendering those failures invisible. The end result is it would, on average, take longer for an assault player to get into a match, which would, in turn, cause the four different "lines" of people to move at different speeds such that while any one game won't fail to find people, individual people will have to wait different amounts of time within that 3 minute context. And the more people there are playing, the smoother that time difference is.

Now you said you're looking for solutions. I didn't get that feeling from your original post, but I've got one anyway: Why not take a cue from Planetside and give people C-bill/XP boosts for playing a minority class? You don't have to show actual class percentages; just an icon on the front page or on all your relevant 'mechs that dropping as a medium right now will give you a 3% boost or something. Maybe even have some kind of strange "kindness meter" that goes up the more you play a minority class. There's plenty of possible incentives to get people to drop what needs dropping.

[Void Angel]

OK, I'm just going to list the major fallacies in each point here...

IceSerpent, on 08 March 2014 - 09:35 AM, said:

You're missing an important point - people want to play with all of their friends currently online (up to the maximum team size obviously, but that limit is unavoidable due to technical issues). A situation where you have 6 unit members online and you have to split into 3/3 or 4/2 just because PGI messed up the queues is equally unacceptable. "Letting people play with their friends" = "any number of players up to the maximum team size has a place (queue) to go to".

I can't be missing a point you haven't made yet, nor is playing with definitions an acceptable way to argue. You're also still arguing based on value judgements which depend on your argument being correct. There is a difference between "PGI won't let me play with my friends!" and "PGI won't let me play with as many friends as I want."

IceSerpent, on 08 March 2014 - 09:35 AM, said:

We all know that matchmaker is currently FUBAR, that's not what we're talking about. We're talking about how it should work. You also misunderstand what is being proposed - we're suggesting much lower range of group size variance, maybe up to 4 (i.e. 8-man on one side vs. 12-man on the other at most). Nobody is suggesting matching up 5v11, although theoretically it can be done if the team with 5-man gets some other advantage - higher Elo or higher BV for example.

Starts with an fallacious appeal to common opinion, and moves on to completely missing the point made. Also contains statements that are false to fact, e.g. "nobody is suggesting matching up 5v11." Again, I did not pull those arguments out of a hat. You're also inserting extraordinary claims with no support whatsoever - such as the assertion that 12 coordinated players (with the option to select 'mechs which work together well) is a balanced match against an 8-man team and four random pugs. Clanner, please.

IceSerpent, on 08 March 2014 - 09:35 AM, said:

The reasons given by devs are simply not true, as in "objectively incorrect". It doesn't "fragment" the playerbase any more that adding a new game mode does and for some reason I don't see people complaining about addition of skirmish "fragmenting the playerbase". Right now we have 2 completely separate queues (PUG and 12-man). What is being proposed is to also have 2 queues - solo only and group, with solo players having an option to queue for solo only, group only, or both. Groups of any size can only go into the group queue. How does it fragment the player base? If anything, it decreases the fragmentation because solo players get the ability to enter two queues instead of one. Lack of objective reasons for the current mess leaves us with two options - devs either don't know how to do it right, or they don't want to do it right.

You seem to have confused "nu-UH!" with a logical refutation. Simply stating that the developers are "objectively wrong" is not sufficient to justify your position as you go along your merry way. Nor is the reasoning you present here valid - people do not complain (much; they do complain) about having extra game modes because the value brought to the game is worth the extra trouble to the matchmaker. Your proposal would absolutely complicate the matchmaker - it amazes me that you can claim otherwise when your own reasoning requires the opposite conclusion. I simply cannot fathom it. You're proposing three queues (solo, either, and group) with grouped players only able to queue for the "group" queue. You also have special rules added in, like "the difference in the number of premades on both teams cannot be greater than 4." Three queues feeding into two sets of matches with special conditions... that's more complicated. It's obvious. You... amaze me.

Indeed, you have failed once again to redress any of the problems with your statement, failed to produce viable objections to my reasoning, and largely refused to support any of your conclusions. This is not a convincing way to engage in argument or polemics, and you have succeeded only in embarrassing yourself. In the end, PGI's reasoning that teams greater than 4 would have an undue impact on the game seems valid - it falls on those who disagree to show if this reasoning is in fact incorrect, or how the supposed benefits of 5-11 man teams would be A) real, and B) worth it. You haven't done so, and as I pointed out to you before, you probably can't do so without demographic data which PGI has and you do not.

Edited by Void Angel, 08 March 2014 - 09:01 PM.

[Gladewolf]

Oh dear, Void Angel again, just ignore him. You will lead a more fulfilled life that way. Did the original premise for this post come from someone that saw this happen over and over again in WoW? That was a very large player base....and when tanks and healers weren't on....PGI may want to try this on the test server first.....The OP(and Ice Serpent) may have a VERY good point.

[Craig Steele]

HammerSwarm, on 08 March 2014 - 08:35 AM, said:

The fact remains that the weight class with the lowest representation will still act as a gate keeper (1) for all matches to be launched. So those left on the outside might be a rolling number but in a population of 1 million and just one class of lower representation(5%) you could still have 1/5 of your population on a rolling 'fails to find match'.


I'm trying to help make this game better. What are you trying to do? (2)

(1) True, but your assumption about time is where your calculation falls down. That percentage may well be picked up by the next set of games ending before the 'failed to find' expiry. You have no basis to conclude one leads to the other without understanding what the percentage is and what the rotation is.

(2) I'm trying to make sure alarmist half thought out deductions don't feed the haters of the forums into a frenzy of bile and venom that has no basis. Is that OK?

[IceSerpent]

Void Angel, on 08 March 2014 - 08:53 PM, said:

I can't be missing a point you haven't made yet, nor is playing with definitions an acceptable way to argue.

I didn't say that you were missing my point, I said that you are missing a point of the discussion in general.

Quote

You're also still arguing based on value judgements which depend on your argument being correct. There is a difference between "PGI won't let me play with my friends!" and "PGI won't let me play with as many friends as I want."

Yes, there is a difference from a standpoint of formal logic, but it's immaterial in this particular case. We always could play with 1-3 friends, so "PGI won't let me play with my friends!" complaint always meant "PGI won't let me maintain a team of my friends all the way up to the maximum team size". You are arguing semantics here.

Quote

Starts with an fallacious appeal to common opinion, and moves on to completely missing the point made. Also contains statements that are false to fact, e.g. "nobody is suggesting matching up 5v11." Again, I did not pull those arguments out of a hat. You're also inserting extraordinary claims with no support whatsoever - such as the assertion that 12 coordinated players (with the option to select 'mechs which work together well) is a balanced match against an 8-man team and four random pugs. Clanner, please.

What common opinion? The part about MM not working as it should is a proven fact (based on information posted by Paul in Command Chair posts), unless you define the purpose of MM as something other than creation of even matches.
I'll take you word on someone actually suggesting 5v11 matching. I haven't seen this suggestion myself, but you are right - I should have said "I am not suggestion" instead of "nobody is suggesting" there, just to be on the safe side. It doesn't change anything about the merits of proposed system though.
Regarding the 8v12 assertion: I underlined the "at most" part for a reason, this is the most "uneven" setup where I personally would still call it a fair game. YMMV of course, and if you strongly feel that variance shouldn't be more than 2 (10v12 max) for example, I won't argue.

Quote

You seem to have confused "nu-UH!" with a logical refutation. Simply stating that the developers are "objectively wrong" is not sufficient to justify your position as you go along your merry way.

If someone makes a statement that certain feature will "fragment the playerbase" AND I personally know how to implement that feature in a way that doesn't cause any fragmentation AND there are existing examples of other people having implemented said feature without causing a fragmentation, then I can conclude that that someone is objectively wrong.

Quote

Nor is the reasoning you present here valid - people do not complain (much; they do complain) about having extra game modes because the value brought to the game is worth the extra trouble to the matchmaker.

Sorry, can't accept this argument - value is in the eye of the beholder and one man's junk is another man's treasure. For example, as far as I am concerned, value of skirmish is zero - I have no use for TDM game modes in general.

Quote

Your proposal would absolutely complicate the matchmaker - it amazes me that you can claim otherwise when your own reasoning requires the opposite conclusion. I simply cannot fathom it. You're proposing three queues (solo, either, and group) with grouped players only able to queue for the "group" queue. You also have special rules added in, like "the difference in the number of premades on both teams cannot be greater than 4." Three queues feeding into two sets of matches with special conditions... that's more complicated. It's obvious. You... amaze me.

1. If you seriously believe that "either" option creates a separate queue, you might want to re-evaluate your position on the subject. Having an "I don't care, either is fine" choice doesn't increase the number of items to choose from. If I am asked to choose between chocolate and vanilla icecream and I say "either", there are still only two kinds on icecream on the table.

2. I never claimed that MM would be less complicated, I am saying that current implementation doesn't function properly and proposed solution that would make it do what it's supposed to do is fairly easy compared to other development tasks.

Quote

Indeed, you have failed once again to redress any of the problems with your statement, failed to produce viable objections to my reasoning, and largely refused to support any of your conclusions. This is not a convincing way to engage in argument or polemics, and you have succeeded only in embarrassing yourself. In the end, PGI's reasoning that teams greater than 4 would have an undue impact on the game seems valid - it falls on those who disagree to show if this reasoning is in fact incorrect, or how the supposed benefits of 5-11 man teams would be A) real, and worth it. You haven't done so, and as I pointed out to you before, you probably can't do so without demographic data which PGI has and you do not.

You never provided any reasoning. You are arguing formal logic for no apparent reason. If you want to see objections, you need to make some statements first, fair enough? So, how about you tell me what you mean by undue impact in "teams greater than 4 would have an undue impact" (impact on what?), what you mean by supposed benefits of 5-11 teams (benefits to who? Benefits to members of those teams should be self-explanatory), and what you mean by those benefits being "worth it" (worth to who and compared to what?).

[Void Angel]

IceSerpent, on 09 March 2014 - 08:42 AM, said:

Yes, there is a difference from a standpoint of formal logic, but it's immaterial in this particular case. We always could play with 1-3 friends, so "PGI won't let me play with my friends!" complaint always meant "PGI won't let me maintain a team of my friends all the way up to the maximum team size". You are arguing semantics here.

Not at all. Rather, I am arguing based on semantics - just as you are attempting to do. Semantics is what words mean. You can't argue without that - but you can try to play games by shifting definitions and playing word games. This is what you are mostly doing, while accusing me of doing it myself.

You have yet to provide any of the reasoning I asked for, instead pretending that I "haven't offered any reasoning," and dismissing my counterpoints to your sophistry as "arguing formal logic for no apparent reason." I had hoped that you would have realized by now that this will not work on me, but perhaps you are so far gone that you honestly believe this is valid argument.

As you have repeatedly ignored my requests that you support your objections properly, I'm in no means obliged or interested in giving you more reasoning to misrepresent. Remedial instruction may be available in the philosophy department of your local institute of higher learning. Goodbye. /ignore.

[TB Freelancer]

Gyrok, on 06 March 2014 - 03:26 PM, said:

LOL...nearly every 4 man premade will be 3 assaults + 1 heavy or 1 assault + 3 heavies, or some mix of the intermediary possibilities.

PGI seriously overlooked the way this will pan out...whatever. It will crash and burn and some *worse* idea will come along...wait and see...

They'll simply have to tighten restrictions groups, and I'm sure it will need to happen. As it is you occasionally get a near 3/3/3/3 balance in teams, and have one lance running one end of the classes (lights/med or assaults/heavy) and the two extremes the situation causes is easy enough to observe.

1) A bunch of really good players monopolize a class and completely dominate the match.
2) A bunch of total newbies monopolize a class and cost the team the match.

Letting a group monopolize a class will occasionally create a gross imbalance. Whether the players piloting them are newbies or veterans decides which way it skews things for the team.

All that said, there are other changes to how matches will be made and how Elo ratings are matched, so I don't think we'll see the kinds of extremes we're accustomed to seeing right now. Not as often anyway....

...for now I'll just wait and see how it plays out in reality before passing judgment.

[Craig Steele]

TB Freelancer, on 09 March 2014 - 08:13 PM, said:

They'll simply have to tighten restrictions groups, and I'm sure it will need to happen. As it is you occasionally get a near 3/3/3/3 balance in teams, and have one lance running one end of the classes (lights/med or assaults/heavy) and the two extremes the situation causes is easy enough to observe.

1) A bunch of really good players monopolize a class and completely dominate the match.
2) A bunch of total newbies monopolize a class and cost the team the match.

Letting a group monopolize a class will occasionally create a gross imbalance. Whether the players piloting them are newbies or veterans decides which way it skews things for the team.

All that said, there are other changes to how matches will be made and how Elo ratings are matched, so I don't think we'll see the kinds of extremes we're accustomed to seeing right now. Not as often anyway....

...for now I'll just wait and see how it plays out in reality before passing judgment.

Yes, but in our little world of Friday night social grinding, we often each take the same mech for comparisons. 3 Jenners, sure, 3 Shadowhawks, why not, 3 Thunderbolts, give it a go. And we laugh about how rubbish each other are with the same mech.

I doubt we are alone and we are certainly not either of those extremes.

[Sephlock]

Actually, come to think of it, what ever happened to the whole lobby thing?

This might work better if we could pick 2 or more mechs (say, one medium and one assault) so matchmaking doesn't waste too much time trying to fit a square peg into a sea of round openings.

[HammerSwarm]

Craig Steele, on 08 March 2014 - 09:46 PM, said:

(1) True, but your assumption about time is where your calculation falls down. That percentage may well be picked up by the next set of games ending before the 'failed to find' expiry. You have no basis to conclude one leads to the other without understanding what the percentage is and what the rotation is.

(2) I'm trying to make sure alarmist half thought out deductions don't feed the haters of the forums into a frenzy of bile and venom that has no basis. Is that OK?

I don't have PGIs hard numbers, the only numbers PGI used to support this radical departure from anything previously discussed was solo/group composition. I'd love to have hard numbers to work with and help them. This thread was about raising the alarm about the potential for a problem and potentially inspire more transparency.

It's okay to try to quell alarmist tendencies of an interested party who is information deprived. Where I took offense was that a page back I felt you were dismissing valid concerns. There is a place for all of us I guess, I was worried you were trolling me. My apologies

Toong, on 08 March 2014 - 02:24 PM, said:

My assumption for pool weight composition is something like 15/20/30/35 percent. People love to remember "that one time" they got that game with four highlanders and four stalkers, because it reinforces popular opinion that the game's turned into AssaultWarrior Online. They also love, when pressed, to adamantly defend this position by taking the memory of those occasional (and memorable!) mismatches and multiplying them by like fifty. I'm convinced this isn't actually the case.

Don't get me wrong, tonnage is a major issue. Probably the single biggest problem with matchmaking right now. But I'm a medium pilot. I've long since gotten tired of people telling me mediums are dead, to the extent that I now make a point to keep track of how many mediums there are in my games. Over the past four or five months I've been watching, I've seen 4-5 mediums in every match, along with 3-4 lights, with the rest being a mostly unpredictable amalgam of heavies and assaults. There's still the occasional game where the numbers are way off, like a team of half lights and mediums versus a full team of heavies and assaults, but they're rare.

I am also a medium pilot, IMHO they are the best mechs.

Back on topic the only player experience I am qualified to comment on is my own, and whatever ELO bucket I am in my results are as such, Conquest - skewed light, Assault - skewed heavy, Skirmish - skewed heavier. I think as a matter of psychology we tend to remember our defeats much more than our triumphs. This is further compounded when we remember when we've felt slighted and feel that we lost unfairly. This might explain a bit of why people always remember the worst examples.

I'd love to have hard data. I'd love PGI to be open and honest when they propose giant sweeping changes to the match maker and provide more supporting data than what percentage of the queue is grouped.

I feel as though I acknowledged that all of my numbers were based on personal experiences and not real world data.

Toong, on 08 March 2014 - 02:24 PM, said:

He's also trying to make the game better, by trying to make the forums a place without toxic hyperbole. While you've handled yourself nicely for the most part in this thread, this feels like a senseless attack on someone just because they disagree with you. Also, your Pinto story has nothing to do with the topic at hand. If you're trying to make a point, please tie it in to your argument.

Acknowledged and apologized. The pinto story was clutch though because it's another example of where a deficiency in a product wasn't caught in the design phase and by the time it was in the production phase it was too late to fix it. Ford further compounded it by saying that the cost too fix it would outweigh whatever after the fact costs might arise. The results were terrible for ford.

I hope we're still in the design phase of launch module matchmaking and that they can address this before it goes live. If you haven't noticed PGI has leaned towards a let it go live and then damage control and spin your way out of it. UI 2.0? 3rd person view? Cool shot?

I think we both want the game to succeed. So I'll move on.

Toong, on 08 March 2014 - 02:24 PM, said:

Anyway.

The lowest-played class will, as you say, play gatekeeper, but the effects of that chokepoint will be felt based on the relative size of your class and the smallest class. In easier terms, the effects will probably only be felt by the most played class.

We only have 100% to work with, so the numbers can only really skew so far. Even if only 15% of the playerbase plays lights, for instance, the remaining 85% is split between three classes. The actual difference is only an average of 18% or so, with a wider difference between lights and assaults. Mediums would feel almost no issue, being around the same percentage as lights. Heavies may experience some slowdown, but assaults would take the brunt of it. Thanks to the 3 minute rolling matchmaking though, any one game would have a minimal chance of actual failure even though there's twice as many assaults running around as lights.

Yes, games would fail -- if pressing 'launch' made a single match attempt. But the matchmaker continuously tries to make a match while that button's pressed. Assaults will fail more often, just as you fear, but all those failures are within the context of the 3 minute matchmaking window, rendering those failures invisible. The end result is it would, on average, take longer for an assault player to get into a match, which would, in turn, cause the four different "lines" of people to move at different speeds such that while any one game won't fail to find people, individual people will have to wait different amounts of time within that 3 minute context. And the more people there are playing, the smoother that time difference is.

Now you said you're looking for solutions. I didn't get that feeling from your original post, but I've got one anyway: Why not take a cue from Planetside and give people C-bill/XP boosts for playing a minority class? You don't have to show actual class percentages; just an icon on the front page or on all your relevant 'mechs that dropping as a medium right now will give you a 3% boost or something. Maybe even have some kind of strange "kindness meter" that goes up the more you play a minority class. There's plenty of possible incentives to get people to drop what needs dropping.

This is my other thread about this. I put it in feature suggestions because I believe that it's not a critique of a specific feature and there is no general forum for information gathering.

I believe that the ques would be much longer than you do. Because the average match length is quite long here.

Additionally it's the nature of a launch module that many of the same players will stay in the same lobby if these choose to. If 100% of players in a lobby stayed there then 0% of players without the gate keeper class could form new games.

To my understanding it isn't as though you'll all fall back into the queue every time but that you'll fall from the game back into a lobby. This will also be hampered by 3/3/3/3 if a person or 2 from class XXX drop out and you need them to continue dropping or you'll be waiting on that class to drop again. Sure you could switch but I thought that to switch weight classes you had to drop and queue again?

No mistaking that I want something different than 3/3/3/3 but I didn't think upcoming features was the place to discuss suggestions. I considered feature suggestions the best place to get noticed. That is why I dropped my critique here and my suggestion there. Sorry if you got the impression this was my end all be all thoughts on the subject.
.

Edited by HammerSwarm, 10 March 2014 - 08:07 AM.