Jump to content

Game Balance And Matchmaking


81 replies to this topic

#1 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 23 April 2019 - 01:09 PM

A nice video I stumbled upon https://youtu.be/NXD...j_Qk?t=2483
Starting around 41:23 they start talking about ELO and W/L ratings.

So it seams that the idea of the PSR rating is indeed better for 12v12 matchups with random people then an ELO system.

Also it makes me wondern, how is the new faction play system working? Is it an PSR or ELO system or a mix?

PS If you see the entire video, there is a lot in it that makes me think about MWO and why its not working so well compared to other games.

Do we have a noob tube? Whats with the learning curve and stuff...

Edited by Nesutizale, 23 April 2019 - 01:15 PM.


#2 justcallme A S H

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2020 Referee
  • CS 2020 Referee
  • 8,987 posts
  • LocationMelbourne, AU

Posted 23 April 2019 - 07:35 PM

FP will have nothing to do with QP PSR, thankfully.

Paul confirmed that earlier in the week.

PSR is not necceasrily bad. However because of WHAT goes into it within MWO - PSR is bad. PSR would be fine if wins didn't count for so much along with other things to affect rises and doesn't give you enough to go down/drop.

#3 Catnium

    Member

  • PipPipPip
  • Bridesmaid
  • 72 posts

Posted 23 April 2019 - 11:29 PM

TBH I hardly lose rating when I lose a QM in the current MM system.
but that's probably because I also actually do something during a match that gives me the score or whatever that is needed to not lose rating every time my team is made out literal window lickers.
You know the kind .. assaults or heavy's with lurms and a gabilion lazors they can never fire anyways since heat is a thing with tiny tiny engines so their mech can barely move because of all the useless **** they put on there, Those who spend the match in 1 location always zoomed in staring at the hill their humping while crying about nascar's and stuff.


those ppl do F all all match and they deserve to drop rating like a rock.
If your one of those ppl you get what's coming to you.

I hope the next PL rating im grinding towards has less of these ppl because grinding through 2 lvl's now we're barely half the people seem to know how to play in each match is getting really really OLD.

Also wtf do new pll start around the middle of the rating levels and not just at T5 where they belong ?
if anything it created a small elo hell around the entry level where half your team is stock noobs and the rest are either window lickers or actual ppl who know a thing or 3 about MWO.

Edited by Catnium, 23 April 2019 - 11:40 PM.


#4 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 24 April 2019 - 12:53 AM

IIRC the "startin at the middle" was made so those who are worst have a place to drop while the others get at least half good teams?
I don't know, it was something along the line.

Also PSR rating is, or should be, designed in a way that you don't get punished (so much) for beeing in a bad team as an ELO system would do.

ELO is just "do you win or loose"? It dosn't care for you haveing a bad team.
PSR is "okay you lost but as long as you tried to carry the team you will still advance a little"

Question is what data goes into PGIs version of the PSR. IIRC its kills, assists, damage and I belive some more.
Problem is, as far as I know, PGI never shared with us whats going in there and they also never showed us any of their data if the system works as intendet.

We all have just our subjective view on these things or at least I don't have a database here that shows me all the data from all matches and stuff. I think there is a list where you can see your personal performance (Jarls list or so?) but that dosn't tell you much if the PSR system is working.

I wish PGI would be more open with such things, same as balanceing. They tell us what their thoughts are behind the patches but we don't see their data, so we all go in with our limited information, trying to figure things out.

Also for people talking about the PSR only going up. That is not true. I am stuck in T2 for....its a long time. Also I don't want to get better, I just play to blow things up and stomp around, not trying to be the best. The PSR system is reflecting that pretty good and keept me at the level. So it seams to work as intendet?

Edited by Nesutizale, 24 April 2019 - 12:56 AM.


#5 Willard Phule

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 3,920 posts
  • LocationThe Omega Company compound on Outreach

Posted 24 April 2019 - 03:39 AM

It puts the right number of people on each team. That's pretty much what it was designed to do. It doesn't separate based on weight class, experience or loadout. That's complicated and this is, after all, a minimally viable product.

#6 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 24 April 2019 - 04:15 AM

From what I understand the PSR system puts people of different skills together because we don't have the numbers to fill teams propperly as they should be. That is not the fault of the PSR system to me but a problem of lower player numbers.

You can kinda see that when you play during peak times and off times. During offtimes your teams are way more mixed and single persons can carry a team. During peaks you get more drop callers and people do more equal damage, have equal matchscores. So system is working...it seams at least to me. Again as long as we don't have the data its a personal impression.

As for seperation by weight and loadout. I think the higher you are ranked the less these two factors are a problem because of the third part, player experiance/skill.
I would say the best players that we currently have can, with their favorit loadout and mech, pretty much beat everyone else that is less skilled and would have a 50/50 win chance against someone of equal skill.
Thats at least what I get from the video and since these guys are more experianced game designers then I am, I tend to take their word for it for now.

Edited by Nesutizale, 24 April 2019 - 04:16 AM.


#7 BlueStrat

    Member

  • PipPipPipPipPipPip
  • 239 posts

Posted 24 April 2019 - 05:59 AM

The problem I see that affects everything in MWO is a general lack of interest/passion for the game by PGI's management and developers, as they do not regularly (or at all in many cases) play the game they are tasked with working on. They have no "feel" for the game nor the player's experience.

It's like an emotionless AI writing music from a spreadsheet. It may be mathematically correct, but nobody really enjoys the result.

Edited by BlueStrat, 24 April 2019 - 06:01 AM.


#8 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 24 April 2019 - 07:20 AM

To me Russ allways sounds like he is falling asleep when I hear him in AMA or podcasts but I think he realy is into Battletech. So there is some passion, can't say for the rest of the team.

As for playing the game. Your game experiance depends heavly on your skill level and I think most devs are terrible in their own games because they don't have the time / energie, after working 8 hours a day on it, to spend another 2-3 hours with gameing to be on the level some competetive players are.

I don't want to see the outcry when they start to do thing by their guts as potato level players and balance everything so potatos have fun.
(Devs if you read this, I have no clue how good you are so correct me if I am wrong).

So yes they will most likely go by numbers to figure things out. Also that can be something good, if you are not to invested and have a little distance.
The more you are invested the harder it will be for you to drop a bad idea or change values. Espacialy if YOU like them people start to not care if they are good for the game.

I feel that currently myself, trying to come up with an idea for a boardgame. I had to drop mechanics I realy liked but where not working and I spent to much time on trying to fit it in somehow instead of moveing on.

Edited by Nesutizale, 24 April 2019 - 07:22 AM.


#9 MechaBattler

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 5,119 posts

Posted 24 April 2019 - 10:57 AM

We need to accept that matches will always be a mix of players of different skills. However if the PSR for solo were actually accurate and place people in their appropriate tiers. The MM could theoretically place an equal amount of the appropriate tiers in each match. Instead everyone floats to the top and each match is a mix bag. With the assumption being that enough games played is equivalent to skill and learned experience. But a lot of people get carried into higher tier just based on wins.

#10 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 24 April 2019 - 11:38 AM

Posted Image

This is the best info I have on how it seams to work with PSR. (Thx @[color=green]Catalina Steiner)[/color]

When your team losses you need a very high personal performance, everything else either keeps you at your level or you fall.
If your team wins a lower personal performance allready raises you.

Question is, what is a low, high or very high personal performance? If the barrier is low enough everyone will rise.
Maybe a solution could be to take the highest, average and lowest performance of a season and adjust these values so that when people over time get better, the overall average score rises so you get a finer seperation in the higher tiers as well as people who don't keep up with the competetive players will be better seperated into levels of their own?

#11 Tarl Cabot

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Tai-sho
  • Tai-sho
  • 7,642 posts
  • LocationImperial City, Luthien - Draconis Combine

Posted 24 April 2019 - 04:00 PM

This has the actual Match score numbers themselves.

If the MS for Win/loss gap was not so far apart, ie closer to a zero sum by increasing the win thresholds for movement up, including a small movement down for a low MS while slightly lowering the Loss thresholds... OR the values have a more dynamic movement/thresholds.

That is just PSR for tier. Again, just an example...for the MM have it use more than just Tier(s) but also average MS for last 30 days/etc. And MM should have selection from 2 tiers, based on the initial seeded player then the next player from a lower/higher tier...

Posted Image

Edited by Tarl Cabot, 24 April 2019 - 04:01 PM.


#12 justcallme A S H

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2020 Referee
  • CS 2020 Referee
  • 8,987 posts
  • LocationMelbourne, AU

Posted 24 April 2019 - 07:51 PM

Yeah it's again the win condition abd the fact you can basically suicide and still go up that's the problem.

If you're match score is below 210 and you're in a win. It should be no change.

Below 180 - small drop.
Below 150 - medium drop.
Below 125 - large drop.

Or something like that. It would then stop all the players with 190 average match score making it into T1. 190 being below the overall player base average

#13 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 25 April 2019 - 01:04 AM

View PostNesutizale, on 24 April 2019 - 04:15 AM, said:

From what I understand the PSR system puts people of different skills together because we don't have the numbers to fill teams propperly as they should be. That is not the fault of the PSR system to me but a problem of lower player numbers.

Again and again and again ... The amount of players you need in order for a properly working MM to form a balanced 12v12 match using solo players is ... 24. Yes, 24 players. Math OP, plz nerf.

#14 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 25 April 2019 - 01:39 AM

The correct answere is, for MM to work properly, you need more then 24 people. Cause if your 24 players are in a range from T-1 to T-5 the MM is forced to put these people together. The result would be a MM that dosn't match people, just sticking the few together that are there.

So idealy to have a proper MM you need 24 of each tier = 120 people so each Tier is filled with 24 people of the same Tier for each match.

Problem with that is these wouldn't be interesting matches. When you put together people of equal skill the outcome of the match becomes a 50/50 chance...what is basicly a random match and it would be more interesting to roll a dice.

So you would mix up some higher and lower skills anyway to tip things slightly. Is that fair? No but it makes for interesting matches. One side might get an easier win, makeing them feel good. The other side might get a hard fought win, giveing them a good feeling.

For this to work and not getting out of hand, resulting in stomps you need a good veriaty of people with different levels of skill so you can counterbalance a to skilled player at least a little bit. Like a T1 with two T2 or three T3...something like that.
But not to much or you will run in the 50/50 problem again.

So from the ideal state of 120 people you would need to add more because you don't have ideal skilled players online all the time. I would guess that you would need to add about 10% more people at least? So 132 people to beginn with.
I don't know how many players there are online at any given time or on average but just 24 people only works if these are all nearly equal in skill.

------

Anyway PSR seams to need a rework. I think their initial goal might have been to give players a good feeling by showing some kind of progress in the Tier bar. Thats nice an all but dosn't serve the game well.

Changing the numbers for what Matchscore gives you what advance/loose might be a thing.
I wonder what excactly they did with FP....they did something with the MM there right?

Edited by Nesutizale, 25 April 2019 - 01:52 AM.


#15 Wil McCullough

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,482 posts

Posted 25 April 2019 - 02:11 AM

Imbalanced matches don't cause stomps. There's an argument to be made that they're correlated but the relationship isn't causal. Stomps are caused by snowball effect. And when both teams start with such limited and non-renewable resources, the snowball effect is ridiculous. If differences in skill level caused stomps, there wouldn't be many stomps in the world championships. And the stompers and stompees keep alternating too.

Stomps happen when one team, by accident or design, is able to punish early mistakes from the other team. That's why close matches are more rare. Either nobody made a mistake throughout the match or neither team is good enough to roll an advantage into a win. That's one reason why close matches are more common in the lower tiers where it's mostly tater on tater action.

#16 Pilotasso

    Member

  • PipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 365 posts
  • LocationInner Sphere

Posted 25 April 2019 - 02:49 AM

PGI follows the current industry trend to launch a broken game for full price and then promises to fix the game later with a whole bunch of excuses. After getting the money there is no incentive for a quality product anymore. This is why for example, faction warfare like all other modes is half assed reflecting in the stomps we see. Fixing matchmaking is a minimalist approach to address the issue of fixing the game later as promised, and is not surprise it has never gotten the desirable results. If they were to sell a product as traditionally games have been, i.e. release with features and quality and proper product reviews to attract the customers before they pay we wouldn't be where we are today.

This is why PGI offered MW5 bundle and I did not, and will not open my wallet.

This is not the way to conduct business in the long term, sorry PGI. I miss the time when people in the business knew how to make a loyal customer base.

Edited by Pilotasso, 25 April 2019 - 02:56 AM.


#17 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 25 April 2019 - 03:06 AM

View PostNesutizale, on 25 April 2019 - 01:39 AM, said:

The correct answere is, for MM to work properly, you need more then 24 people. Cause if your 24 players are in a range from T-1 to T-5 the MM is forced to put these people together. The result would be a MM that dosn't match people, just sticking the few together that are there.

Sigh ... MM matches teams, not players. If a combination of T1's and T5's on one team matches an overall skill of a combination of T2's and T3's on another, then its a balanced match.

View PostNesutizale, on 25 April 2019 - 01:39 AM, said:

Problem with that is these wouldn't be interesting matches. When you put together people of equal skill the outcome of the match becomes a 50/50 chance...what is basicly a random match and it would be more interesting to roll a dice.

Seriously? ... 50/50 chance of a win in any given match is exactly what people who ask for a balanced MM want, because not only it means that matches are exactly that ... balanced, but it also means that the outcome is decided by the slightest things every single player does. I.e. your impact actually matters.

And you say its "not interesting" ... Pfffft. Ridiculous.

#18 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 25 April 2019 - 04:16 AM

Its a bit counter intuitive I know.

From what I gethered by what other game designers say its when in a team everyone is equal then each person has less impact on the game overall.
As you have the same skill as everyone around you, you don't matter that much anymore. If you fall, there is someone as good as you are to replace you. Thats what the statistics say.
Your personal "hero" moment or "omg everyone is a potata" comes when there is a differance and not equality.

Quote

Sigh ... MM matches teams, not players. If a combination of T1's and T5's on one team matches an overall skill of a combination of T2's and T3's on another, then its a balanced match.

Yes that is what I said. It takes players skills and groups them together and to do that each players skill has to be looked at. With the limit of 24 players that you had given it would mean that the MM can not make a good match as the numbers of different Tiers would most likely never match to an equal state.
That is why you need a bigger group then 24 people that are searching for a match to have the equal teams you want.

#19 FRAGTAST1C

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Fighter
  • The Fighter
  • 2,869 posts
  • LocationIndia

Posted 25 April 2019 - 04:20 AM

View PostNesutizale, on 25 April 2019 - 01:39 AM, said:

Problem with that is these wouldn't be interesting matches. When you put together people of equal skill the outcome of the match becomes a 50/50 chance...what is basicly a random match and it would be more interesting to roll a dice.

So you would mix up some higher and lower skills anyway to tip things slightly. Is that fair? No but it makes for interesting matches. One side might get an easier win, makeing them feel good. The other side might get a hard fought win, giveing them a good feeling.

For this to work and not getting out of hand, resulting in stomps you need a good veriaty of people with different levels of skill so you can counterbalance a to skilled player at least a little bit. Like a T1 with two T2 or three T3...something like that.
But not to much or you will run in the 50/50 problem again.


Interesting thought but if the MM is supposed to work as intended, then every team should have an equal chance to win, which means 50/50. That's basically what everything fairly balanced will be like - 50/50.

#20 Nesutizale

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 3,239 posts

Posted 25 April 2019 - 05:06 AM

Problem is we don't know what is the intend of PGIs matchmaker. We are currently "theorie crafting" a MM...kinda?

Also is it fair to have different loadouts then and different mechs?
If you have a match where everyone is at the same skill level, wouldn't the mechs they bring decide the match?
Put a meta build against a fun build and have equal skilled players, guess who will most likely win?
I dare say its the meta build. That would mean we have to take mechs and loadouts also into the MM calculation...but with the tons and tons of different configurations, how do you balance that?

Sure you could have an arbitraty rate like the BV in the Tabletop but that also comes with its sets of flaws IIRC.

Frankly MM and balanceing are a terrible difficulte thing. As one said "They still try to balance Street Fighter after 30 years".
MWO isn't nearly that old and for that its running okay.

I hope PGi will at some point open up a bit more and give us a bit of insight to how things work and how their metrics are. They should start tapping into the collective mind that is here to brainstorm ideas and get feedback much faster.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users