79 replies to this topic

[Horned Bull]

Serapth, on 03 January 2013 - 02:02 PM, said:

Just to be clear, I am not proposing a solution, I am proposing a fix.

Well as a short time fix this might work quite well.

Serapth, on 03 January 2013 - 02:02 PM, said:

[color=#959595]In the meantime, we've seen additions, like ECM, that actually shift the game even more strongly in favour of premades.[/color]

[color=#959595]Before ECM you had "additions" like LRM/streak boating that stronly favoured premades (LRM premade party stomping pubs).[/color]

[Serapth]

Korm, on 03 January 2013 - 03:00 PM, said:

Well as a short time fix this might work quite well.



[color=#959595]Before ECM you had "additions" like LRM/streak boating that stronly favoured premades (LRM premade party stomping pubs).[/color]

I dont want to turn this into an ECM *****-fest thread, but yeah, streakcats were annoying, but you could counter them at > 270m. Now you get streak-ravens and streak-mandos that can't be countered at all. If ECM was the fix for streaks, they missed by a friggin mile. But yeah, premade packs of streakcats were almost as irritating, ECM certainly hasn't made the problem in any way, but the ability to guarantee a number of ECM mechs on your team is a huge advantage for a premade group.

Edited by Serapth, 03 January 2013 - 03:05 PM.

[kalami]

I rarely lose in my pre-made group so to me that says it's something wrong. That is the main reason I usually go PUG because it's more interesting; yes from time to time jumping all over some PUG groups is fun, but id much prefer a challenge and to play with my friends.

I do have hope that PGI is going to fix it; just have to be patient; hopefully it won't drive away too many of the casual people in that time period.

Edited by kalami, 03 January 2013 - 03:05 PM.

[Horned Bull]

Serapth, on 03 January 2013 - 03:04 PM, said:

I dont want to turn this into an ECM *****-fest thread, but yeah, streakcats were annoying, but you could counter them at > 270m. Now you get streak-ravens and streak-mandos that can't be countered at all. If ECM was the fix for streaks, they missed by a friggin mile. But yeah, premade packs of streakcats were almost as irritating, ECM certainly hasn't made the problem in any way, but the ability to guarantee a number of ECM mechs on your team is a huge advantage for a premade group.

The real problem here is... The infamous lagshield. Without it commandos and ravens (not to mention cicada) would be pretty fragile. Countering a fast streakcat was actually a pain - you weren't able to run away from it in a heavier mech, and in a lighter one, the strakcat's pilot needed only one or two salvos to put your mech down.

Without ECM/Streaks pubstomping premades would probably just stack AC20 catas/srm catas/lrmboats.

In a group you can easily field a cheesy as hell team just be using the current flavour of the month.

[JSparrowist]

Whoever says that matchmaking is working as it should and premades don't usually win against pure PUGs needs their head examined.

I have primarily been a pugger since I started playing MWO until recently when I started dropping with some other randoms I met while playing. We don't use TS or any other VoIP solution to coordinate but we have an unspoken understanding that we stick together and cover one another in battle and my wins vs losses has SKYROCKETED! Taking just a small bit of the "random"ness of who you play with out of the equation can make all the difference. I am fairly confident that if we added VoIP into the mix as well, we'd be damn near unstoppable against pugs.

Those of you who say that everything is fine are the ones who abuse the system. I have no doubt about that. You resist the OP's idea for a short term fix because it would for the most part negate your advantage and hinder your abuse of the system.

I don't believe that ELO will solve the problem at all, but i think it may help the problem a bit.

[Grits N Gravy]

I keep reading that many people think the elo system will be a decent and balanced approach. Fundamentally it will not be for a multitude of reasons.

Currently they are ranking us and using that data to examine the accuracy of the system and then tweak. One of the fundamental problems with this system is that the data they are gathering is biased. Elo systems work best when player's skill is distributed in a bell curve or Gaussian across the population.

Currently there are 3 distinct curves for player skill. People who exclusively group, people who only solo drop and people who mix, all having their own skill distribution curves. When you combine them, you end up with is a bimodal mixed distribution curve that is right skewed. This means that the majority of players will have scores below the mean of the aggregate score. Which generates an artificially high standard deviation.

With good matchmaking you really only should be fighting people within 1 standard deviation of your score. The way scores are being computed, combine with the effect that grouping has on win rates; Will result in a standard deviation so large that it's will allow people of wildly different skills to face each other as often as they do now.

The system is only a marginal fix, the best will never face the worst and everybody else is the same. Really what is need are better metrics than win loss ratio for determining skill and matchmaking. A weighted matrix of, an individual's average percent of the team's total damage, the amount of cap points an individual scores, number of kills, assists, win ratio ect ,ect. Would provide a better basis for the rank of an individual's skill and thus a better matchmaking. This approach allows for players who are not in groups, but lose and contribute; To either not lose as much ranking as other poor performers on their team or to actually gain ranking in loss, depending on how you set up your system. Couple something like that with a "ratings reliability" function that's found Trueskill and glicko systems and you would actually end up with a decent multiplayer matchmaking system.

Elo systems have been abandoned by many games because of their inherit inaccuracy and vulnerability to exploitation. WoW for example, had to add personal rating to stop win trading, and eventually abandoned the system all together. http://wow.joystiq.c...stem/#continued

There are effective metrics for measuring whether the match maker functions as well. An increase in the number of battles which result in a draw and having players with a mean score achieve a 50/50 win ratio, all indicate a matchmaker that functions well. It remains to be seen if that will happen here.

Edited by Grits N Gravy, 03 January 2013 - 07:15 PM.

[JP Josh]

how bout the friken pugs learn when to run away from the main brawl and when not to many of time when its a 8-0 pug slaughter its not due to pre maid

its because we either had 2 disconnect or 3 dis/afk or my entire team retreated to base leaving three of us guarding the main traveled routs.

[Martini Henrie]

ELO will be there to stabalise the matchmaking process, not to make it perfect. That was never the point. The point that is being missed about 'pre mades' is that they are as diverse as puggers. They are not all waac players in the same way as puggers aren't.

Most of the time we do not, in our corps, specify what mech or type has to be taken. We just play in what we feel comfortable. The other reason we use or have TS is because we bring friends who play other games with us and one central hub is simpler to administer than having a load of separate TS servers to deal with. I think my TS server list has about 5 on at the moment.

To the previous poster, a good friends list is golden, even if you don't use TS. There are a lot of games that don't have integrated chat for one reason or another, and personally sitting back and moaning about it when there are free ones available moots the point.

What is the ratio of pre made to pug? Nobody has ever posted it, however to some it has wrongly become a demon that sits on their shoulder. I really do not see this as an issue myself.

[TB Freelancer]

Wizard Steve, on 03 January 2013 - 01:19 PM, said:

The player base isn't big enough to split.

You're right to an extent. But pitting pugs against groups is a sure fire way to keep the player base small. If it doesn't get big enough to split pugs and groups, then the game should fail.

Elo won't solve anything. If it works as described, it will be of mild help, but lopsided matches will continue and pugs will find themselves on one side of it or the other far too frequently. It would work great matching individual players, but groups with coms and mechs chosen before the drop will make all the math basically worthless, still leaving individual players with a sour experience more often than not.

The game is hemorrhaging pugs at an alarming rate and the effects are pretty clear. I've gone from seeing games full of pugs every other match, to a night full of seeing one or two groups I recognize per match, and encountering a handful of the same groups multiple times over the course of an evening. Also I've been getting a lot more (I only pug) failed to find a match, or long waits before finding a match than I did just a month ago.

That isn't good at all, and Elo as they plan to implement it won't do a damned thing to stop the bleeding.

Personally I'd limit it to a single pair of two per side in pug matches, and hard groups of 4 or 8 in a group queue in any combination. The only reason people play 4s right now is because they're much more likely to win...and who doesn't like winning all the time?

Just take away the easy button and give groups a simple choice. Play as a pug or with a single buddy against pugs, or group up and play against other groups.

...because the player base is small, PGI would likely have to either leave the unbalanced matches, allow some level of tolerance of team weights, or enforce drop restrictions per group to ensure short wait times. Personally I kind of like that last one since I'm getting awfully sick of seeing drops with 5+ of a single weight class usually, assaults, lights or the odd heavy match. Balanced teams make for much more interesting battles.

EDIT: Because of Elo, I would think hard restrictions or some degree of tolerance in group weight restrictions allowing a weaker team to field more weight to balance things out and make wait times shorter.

Edited by TB Freelancer, 04 January 2013 - 12:34 AM.

[Serapth]

Martini Henrie, on 03 January 2013 - 11:44 PM, said:

ELO will be there to stabalise the matchmaking process, not to make it perfect. That was never the point. The point that is being missed about 'pre mades' is that they are as diverse as puggers. They are not all waac players in the same way as puggers aren't.

Frankly if there is a problem with the community size supporting splitting the queue ( again, not suggesting that ), how the hell is ELO going to work?

I think there is a problem with the matchmaker in this regard already. Consider a premade group composed of 4x Atlases drop into a group together. Now the matchmaker needs to try to find 4 assaults from the PuG queue or other premades to try and balance the match. I have a sinking feeling that this is what leads so often to 8v7 matches being started, as it times out trying to accommodate this configuration then times out when it fails, massively screwing one side.

Now factor in ELO, which is going to add another X factor to the mix, each player has a rating.

So now that matchmaker ( which is already failing hardcore ) is going to have to not only find mechs to match the premades mech loadout, it's going to have to find pilots with comparable rankings as well.

I can see how ELO would work for matching individuals against other individuals, so newbies aren't getting owned by vets and vets have a more interesting experience. When it comes to premades though, it's just going to make it even harder to find compatible matches.

If the community is actually that small, then ELO is damned to failure before it even launched.

Edited by Serapth, 05 January 2013 - 12:08 PM.

[CrashieJ]

how about 4 v. 4?

less strain on the game server-side.
no "pugs" to leave as cannonfodder or such.
less TKers

can be regulated easier than 8 v 8.

problem solved.

[Martini Henrie]

Note the word stabilise, not perfect. Getting the balance right across the drop would be relatively simple within a resonable variation. There is no need to balance the skill level of individual mechs, just the balance over the team. This means your example of 4 Atlas is moot. The teams skills would average out.

I don't see the matchmaker currently being that bad, but then I can work around a lot, and enjoy a good challenge. There is still the point that only a small percentage, of the percentage of 'pre-made' teams actively build 'focused' 4 man drops. The vast majority, like pugs, run what they enjoy.

If the ques were split, from the one reasonable version we have now, where would the recruitment and cross pollination come from? The game would rapidly stagnate into two smaller ques and probably wither...

[Wraith05]

Hotthedd, on 03 January 2013 - 01:32 PM, said:

Even if there are 5 sharks and 3 minnows on each team, the minnows still face an early death. Cadet status does not last forever. Eventually (after learning a bit it is hoped) they will have to swim with the bigger fish.

I don't think you understand how the ELO works. (going with your fish example). It will start you off as a minnow and when you win/get enough skill (however they rate skill) you are bumped up to a tuna level. Then once you win enough/enough skill you are bumped up to a flounder level. Then swordfish, then a shark

With ELO a minnow should never ever be in the same game as a shark unless both the shark and minnow are new players in which after several games the sharks are moved up through the tanks until they are in the one they belong in.

[jakucha]

Yeah, a decent ELO system should make the game markedly better than it is now with just throwing people randomly together.

[Trauglodyte]

ELO might help with PUG stomping but it won't help PUGs from themselves. The lack of integrated voice chat in this game makes PUGging pretty hard. Then you add in the lack of people wanting to work together - IF ITS RED, ITS DEAD - and complete and total failure for understanding strategy. I swear, if I see another jackass run straight on into the fray, I'm going to face punch myself. Hey newbs, stop rushing the center. It doesn't work when you're herding cats with the other 7 people on your team.

[Redoxin]

So where do the 2 and 3 man groups go with this idea?

Trauglodyte, on 05 January 2013 - 12:58 PM, said:

ELO might help with PUG stomping but it won't help PUGs from themselves. The lack of integrated voice chat in this game makes PUGging pretty hard. Then you add in the lack of people wanting to work together - IF ITS RED, ITS DEAD - and complete and total failure for understanding strategy. I swear, if I see another jackass run straight on into the fray, I'm going to face punch myself. Hey newbs, stop rushing the center. It doesn't work when you're herding cats with the other 7 people on your team.

If I was pugging I would sure as hell leave any voice chat with random people that just happen to be in a game with me. This public, in game voice thing has never worked well in any game, wont work in MWO either.
And the reason that premades are dominating is not voice communication, its them being the better players on average.

[Windsaw]

Grits N Gravy, on 03 January 2013 - 04:53 PM, said:

There are effective metrics for measuring whether the match maker functions as well. An increase in the number of battles which result in a draw and having players with a mean score achieve a 50/50 win ratio, all indicate a matchmaker that functions well. It remains to be seen if that will happen here.

Do I understand you correctly:
If I train to get better, if I struggle to be as good a player as possible and this will not result in any noticable increase in gaming performance, then this game is supposed to have a good matchmaker?

Please tell me I am wrong or this was a joke.
Because with such a matchmaker, why should I play at all? Or at least: Why should I try to not suck?

[Trauglodyte]

Redoxin, on 05 January 2013 - 01:05 PM, said:

So where do the 2 and 3 man groups go with this idea?


If I was pugging I would sure as hell leave any voice chat with random people that just happen to be in a game with me. This public, in game voice thing has never worked well in any game, wont work in MWO either.
And the reason that premades are dominating is not voice communication, its them being the better players on average.

I agree with that entirely. Blizzard failed with integrated Voice Chat in WOW cause it barely ever worked and people wouldn't use it even if they were in it.

As for premades being better, well that's a given. Teams with fully optimized mechs and mostly optimized groups are always going to be better than PUGs that are mixed with trials, non-fully leveled chassis, etc. I'm a pretty good pilot in my Cicada and I average 2-3 kills a game, most of the time, but I'm not enough to turn the tide on a really really bad PUG team. I can't get people to stop running around solo, I can't get people to focus fire, can't get people to stop poking their noses in the Caldera in Caustic Valley, etc.

[Hotthedd]

Wraith05, on 05 January 2013 - 12:35 PM, said:

I don't think you understand how the ELO works. (going with your fish example). It will start you off as a minnow and when you win/get enough skill (however they rate skill) you are bumped up to a tuna level. Then once you win enough/enough skill you are bumped up to a flounder level. Then swordfish, then a shark

With ELO a minnow should never ever be in the same game as a shark unless both the shark and minnow are new players in which after several games the sharks are moved up through the tanks until they are in the one they belong in.

So the minnows will not be playing against the sharks. Seems very much like having a cadet-only queue.

[jakucha]

Redoxin, on 05 January 2013 - 01:05 PM, said:

So where do the 2 and 3 man groups go with this idea?


If I was pugging I would sure as hell leave any voice chat with random people that just happen to be in a game with me. This public, in game voice thing has never worked well in any game, wont work in MWO either.
And the reason that premades are dominating is not voice communication, its them being the better players on average.

Actually it's because they focus fire. That's 90 percent of winning in this game.