87 replies to this topic

[EndlessBoogie]

Joseph Mallan, on 20 December 2013 - 05:19 AM, said:

I am not impressed. As a PUG my win loss record is almost the same(54%) as my Team percentage(58%).

That proofs nothing...

A 12:0 is a win and a 0:12 is a loss , are those games fun? I don´t think so...
A 12:10 is a win and a 10:12 is a loss, are those games fun? Yes...

So this "OH look my w/l as pug is just the same, so everything is fine " is just dumb.

[NextGame]

Given that there has been at least 3 threads on this topic today in this forum section:

I think the people complaining about the matchmaker are conveniently forgetting the frequency with which it works for them rather than against them.

Yes, there needs to be some kind of equipment battle value applied to tighten things up but realistically everyone knows that if a side goes 2 or 3 down that side will tend to take a heavy defeat regardless of any weightings that might be applied.

I think people complaining about defeat should think about whether they are genuinely complaining about the mechanics or actually would find something to blame their defeats on regardless and need to instead man up for realsies.

protip (lol) on how to emulate teamspeak play on public drops for those lone wolves out there that feel hard done by but are too timid to party up with people:

1) find the biggest mech on your team and follow it around.
2) shoot the same stuff that the big mech shoots.
3) profit.


It would of course help immensely if long awaited features such as voip were introduced, but lets face it PGI haven't been particularly interested in delivering on many of the original goals for a while.

incidentally, regarding weight balance, the counter to top heavy teams is to have a lance of spiders. If you want to moan about broken stuff, might as well point out all facets of it, otherwise you aren't giving the full story

Edited by NextGame, 21 December 2013 - 10:42 AM.

[Sandpit]

I'm glad someone finally got the nerve to post on this

[NextGame]

Sandpit, on 21 December 2013 - 11:30 AM, said:

I'm glad someone finally got the nerve to post on this

I'm glad that there's someone like you parroting the same line in multiple threads, it makes me feel much better about people who have opinions don't necessarily align with my own, in that at least they have an opinion to discuss. As such, thank you for your valuable service to the community.

[Novakaine]

SOLDIER UP AND ENLIST
It's more fun with friends.
Unless you're a antisocial mom's basement dweller.
Then maybe not so fun

[Sandpit]

NextGame, on 21 December 2013 - 11:34 AM, said:

I'm glad that there's someone like you parroting the same line in multiple threads,

That's my point exactly. This guy gets it

Oh, and MY opinion?

MY constructive posts and ideas on this? It was posted in the first 5 or so duplicate threads on this exact same subject.

[MischiefSC]

@ wwiiorgre;

Here's the problem. I'm pretty critical of PGI, the matchmaker (I posted like a hundred times across 3 threads on the topic not long ago) and game balance. I'm not having the same experience as you though. I'm not seeing or experiencing what you're seeing and experiencing and I track my metrics and in game metrics pretty closely. I'm something of a math nerd. I'm currently averaging about 7.5 out of 10 matches within 150 tons balanced between teams though my time to find a match is close to 90 seconds. As I've been trying to level both Locusts and Tbolts I'm losing about 60% of my matches at present, normally my overall win/loss is about 58% so it's been a painful swing but I have only myself to blame. Battlemasters were a sad, grim, painful experience and I'm still very pissed at PGI for absolutely shafting such a trademark mech as the Battlemaster with flat-out ridiculous design decisions. Awesome style torso width, but with less torso twist! It's like leaving nothing to chance!

Elo is not capitalized - it's a mans last name. Not trying to be a jerk but when you're trying to discuss something it's important to be able to come across as knowing the subject. Elo also has nothing to do with how many premades are on each side. Elo simply measures how often you win or lose matches in a given weight class against opponents of a similar skill level (on average). Elo is then used by the matchmaker.

I'd absolutely agree that the matchmaker needs work and that matching players within a given 'range' is going to give a drastically better experience than matching high/low to a target. To make that feasible you need to 'fatten up' the bands of Elo scores so there's enough people within a given Elo range to fill matches with weight matching - that's where switching to a Gaussian distribution on score is important and reducing the 'k-factor'. The k-factor is how many Elo points you gain or lose for how the match turned out. If it's a high number you'll move ranks faster but it'll be less stable and less accurate. Lowering the k-factor will give a more stable player population curve and make matchmaking more accurate.

Finally, premade vs pug Elo. When you drop in a premade it averages the Elo of your whole team, gives you a 'bump' and then puts your average into the team Elo score when trying to balance for a target mean. This means that just the presence of a premade skews the teams total Elo score just a bit. This goes wild however when those premade players pug.

Suppose you drop regularly with some amazingly good people and you win often playing with them. Maybe you play a spotter light or coordinate well with them. Alone however you're not nearly as good - the matchmaker doesn't know this so it takes your stellar Elo score into consideration when you drop in a pug game. If your Elo is high enough you're probably dropping with a team that's much lower Elo than you - it will then grab some literally newbie to 'offset' your high Elo score and get your team towards the average score it wants to find for this match. Result? Players who are only a bit better than average when pugging being matches with newbies or hardcore bads and throwing the matches actual ranking way out of whack.

Statistically it's not going to happen as often as you're implying but it happens. It's fixable but I'm not sure PGI has the will to do it. I suspect that MM did *not* turn out as they planed and they either don't have a statistician or if they do they don't utilize him or his calculations are not as attentive to detail as we'd like or.... well, whatever. You can't tune something like a Matchmaker off anecdotal opinions though. You don't 'tune it to feel'. You run it absolutely off of math and you identify the math that will get you the results you want.

Biggest issue is and will always be population. It's got to fit you in a match based on everyone else who hit 'Launch' within 120 seconds of when you did and doesn't fit better in a different match. The issue however is not Elo - it's how the matchmaker uses that Elo score. We need a fatter Elo population distribution (gaussian), we need to focus on matching you to similar skilled folks and not high/low scores for an average, we need to split pug/premade Elo. Then you'll get your more balanced experience. We also need about 1 million players. Please get to work on that.

[wwiiogre]

Mischief,

I know ELO is the name of the man who designed the math, but like programming, garbage in equals garbage out. Is win/loss how skill is figured? Because currently if you have a winning percentage, then matchmaker will put you on a team with other people who have bad scores.

So right off the start PGI's version of their matchmaker is not trying to balance the game in any way at all other than to manipulate win/loss records. Which of course leads to so many games where 4 mans are top heavy on one side, where mech tonnage is top heavy on one side or where elite level 3 equipped mechs are all on one side.

My point is current matchmaker takes none of the things into consideration that would really balance a match. The w/l or k/d of one team or another is not an indication of that teams skill.

A measure of a persons skill can be tested when they face a mech of the same tonnage with near the same equipment costs piloted by someone with about the same amount of drops. That is how you balance a game. Then after those two pilots play around 100 games facing similar pilots with similar mechs with similar equipment. You then have a chance to actually judge the better pilot based on win/loss and k/d. Because now you have reliable data, based on similar statistical groups.

Otherwise you have what we currently have, a bright idea to use a statistical averaging system but having no sound data to base it on that could possibly give you a real assessment of skill of the person piloting the mech. Instead we get new players in trial mechs grouped with other new players none of them on voip, facing veteran players in elite mechs with elite equipment in heavier mechs with more firepower and using voip.

A real matchmaker trying to make a game fun and balanced would never allow the above to happen. NEVER!

@Nextgame, nice 12 man private match. I do not know what point you were trying to make with that screen, but it is completely invalid since we are discussing and debating pug dropping not 12 mans. 12 man drops I have no issue with and I am completely happy with how that works. I am just surprised you got a game in.

The Battlemaster, I am very happy with my Battlemasters and have great luck running them. I either use a Bug Swatter with 4ssrm2, 2xERLL or a LRM build with 2xLRM20 artemis 3ML and tag with bap, both usually do well either in solo pug or with group drops.

Like I said I am familiar with PGI's attempt at ELO and after many months have to give it a completely failing grade of an F. It should be aborted, or drastically adjusted. Because it really does not even attempt to rate the skill of players dropping, it does take into context the equipment or tonnage used. Because of all that you get far more blow outs than close games. At least I have seen it that way and its been overwhelming.

And I am not complaining because I am always on the losing side, the game is not fun when you win 12-0 or 11-1 and its not a challenge. I like games that come down to a nail biter with 1 or 2 badly damaged mechs on both sides trying to stay alive and get that last lucky shot in to win the game for their team.

Its no fun to see on the starting screen all lights and mediums on your team, 4 trial mechs, no 4 mans and start on River City. In the first 2 minutes you realize the entire other team is in Heavy and Assault mechs and just blob to the Center of the map and kill any thing that reveals itself within 1-2 seconds. Even when playing that game in conquest the blob of assaults and heavies could easily control three points and maintain control.

Now the same situation on Alpine or Crimson then the lights and mediums have a chance, but not if the other team is coordinated enough to not extend itself and stay within line of sight of each other.

So sorry PGI your matchmaker is horrible and does not work. It is designed to only manage win/loss records not make a balanced challenging game.

Chris

[Steven Dixon]

To OP, I agree that matchmaker needs some tweaks and weight balancing is one of them, but I'm afraid that I must disagree with your solution to remove elo in favor of balancing based on mech value rather than pilot value. I would much rather face an atlas piloted by an average or unskilled player than a jenner piloted by a top tier player. And Elo is one of the few ways I can think of balancing that.

I also agree with the fact that 4mans should be teamed against 4mans but this isn't always possible, MM does at least try to team them up. But again Elo actually helps with this. A team that goes for a 4-man is probably going to pugstomp lots of players but they are also raising up their elo which means that they are going to get teamed up against hirer tiered players. While a bunch of noobs teamed up may not win and wont raise their elo but then because they aren't winning then they aren't much of a threat.

As an aside, I don't know about anyone else but in my matches that are 1 sided in terms of weight it seems that the lighter team often wins (even when straight up fighting, not capping). This could certainly be observer bias, I expect the heavy team to win so I'm shocked when they lose so I remember it more vividly, but I don't think so.

I do agree that rewards should be tweaked to encourage role warfare and lighter mechs as well, though I fear the tone of some of the post might just encourage flame wars and white-knighting rather than an intelligent debate, I guess we'll see.

[Sandpit]

NextGame, on 21 December 2013 - 11:34 AM, said:

I'm glad that there's someone like you parroting the same line in multiple threads, it makes me feel much better about people who have opinions don't necessarily align with my own, in that at least they have an opinion to discuss. As such, thank you for your valuable service to the community.

Just to prove a point

http://mwomercs.com/...it/page__st__20
http://mwomercs.com/...st-system-ever/
http://mwomercs.com/...r-any-one-else/
http://mwomercs.com/...tem-worries-me/

so yea, I'm glad someone finally got the nerve to post on this. Those were found just in the first couple of pages in THIS section without doing a search. SO yea

[L Y N X]

Mudhutwarrior, on 20 December 2013 - 03:43 AM, said:

Primarily the stomps in skirmish seem to be whomever is on teamspeak dominates. Trying to locate and link up on typed chat is difficult at best because the engagements happen so quickly. I see a lot of broken typing with guys just hitting enter so they can regain control.

Premades are basically kicking butt on the deaf, dumb and blind. No skills needed there, just voice comms. Command and control with quick intell wins battles. Lets not be obtuse. We know what the real issue is and it isn't tonnage or weapons loadouts.

If you are saying Teamwork wins? I agree!

I suggest now that you are aware of this that you find a team that fits you. Loners lose, winners work together. Anyone wanting teamwork and willing to work together, come check out Team Special Forces, http://www.team-sf.com. If you do not fit with us, talk to us as to why and we can possibly suggest another outfit that suits your personality better.

[Sym]

You want to save this game...especially after the clan tech is released? I'm just going to say one thing...

Battle Valve.

Edited by Sym, 22 December 2013 - 11:28 AM.

[Sandpit]

Sym, on 22 December 2013 - 11:27 AM, said:

especially after the clan tech is released?

everyone keeps on saying this. It's impossible to say this right now simply because we have absolutely no idea on how clan mechs and tech will work. We have one singular example in a post by a dev on how they MIGHT work. We simply cannot know this right now

[Sym]

Going off the system that we have now, it will only get worst.

But I love how the Devs want to "nerf" the new Clan tech. Ya, that solves everything!

[Sandpit]

Sym, on 22 December 2013 - 12:18 PM, said:

Going off the system that we have now, it will only get worst.

But I love how the Devs want to "nerf" the new Clan tech. Ya, that solves everything!

and again, we have no idea what clan tech stats will be or how it will work. It's complaining about an item before you know anything about the item.

[Accursed Richards]

The thing is, should tonnage imbalance even be a problem? I mean, wasn't role warfare meant to eliminate the race to the assault mechs that consistently plagues BT-based games?

[Roland]

Sandpit, on 22 December 2013 - 12:44 PM, said:

and again, we have no idea what clan tech stats will be or how it will work. It's complaining about an item before you know anything about the item.

Let's be realistic.

Based on track record, do you believe that balance will will improve, or get worse, with the addition of clan tech?

Take your time.

[Sandpit]

Roland, on 22 December 2013 - 04:49 PM, said:

Let's be realistic.

Based on track record, do you believe that balance will will improve, or get worse, with the addition of clan tech?

Take your time.

Well considering I think the game it well balanced now? I think it will be well balanced then. You're welcome to have a different opinion though.

[xMEPHISTOx]

Roland, on 22 December 2013 - 04:49 PM, said:

Let's be realistic.

Based on track record, do you believe that balance will will improve, or get worse, with the addition of clan tech?

Take your time.

As with Sandpit well balanced now or at least as well as it has been since ive been playing from the closed beta days. How will ClanTech change it...no way to know w/out more information on the subject.

[ImperialKnight]

the matchmaker is {Scrap} and we all know it. the ELO balance is questionable at best and there is ZERO tonnage balance