23 replies to this topic

[SilentScreamer]

After reading PGI's post regarding the High Alpha Damage Meta
https://mwomercs.com...ha-damage-meta/

This is not a thread discussing the high alpha meta itself, nor any specific build. But rather a challenge to the community and PGI to answer the question, why continue to attempt to balance mech's based solely on tonnage? Why not move to a point-based system? This proposed system like the current system would IGNORE specific builds, instead merely attempt to match Mech's based on average performance rather than tonnage, and hopefully reduce Quirk and Weapon adjustments as a result.

I suggest that instead of pidgeon-holing mechs to perform according to their tonnage, transition MWO to a Point System used from the Match Maker and Faction dropdecks.
- First, each mech assigned a point value equal to it's tonage. This is the starting point.
- Second, if a mech grossly out-performs or under-performs assign a point handicap or adjustment based on that mech's performance for the purpose of Quickplay Matching Making. +5 /-10 etc.
- Third, regularly evaluate performance compared to others of it's tech-base only in regard to balancing Alpha/ Manuverabilty/ Survivability/ Performance. If minor adjustments are needed use Quirks. If major changes are needed adjust the point handicap.

Ex: If the optimal mixed pod Timberwolf, Gargoyle and Ebon Jaguar all perform similarly in-match they would have the same point value despite the tonnage differences. Omnis would have different Value adjustments based on stock pods or mixed. Similarly each varient would be rated on specific match performance. The Locust 3V would have a seperate adjustment from the 1M or 3S models etc. The performance-equivalent-match for a Nova might be a Thunderbolt 5SS or a Quickdraw, not a I.S. 50 tonner. Additionally, the point value could be different for each game type; Quickplay, Invasion, Scouting, and Solaris7 as some mechs are good performers in one or two, but bad in another.

Just including a few mechs as an example:
15 points - LCT-1V, LCT-3V, LCT-1M, LCT-3S, SDR-5V
20 points - LCT-1E, LCT-3M, LCT-PB, CMD-1B, CMD-1D, CMD-DK
25 points - CMD-2D, Mist Lynx, Piranha, OSR, PNT, UM
30 points - Arctic Cheetah, FS9, JR7, JVN
35 points - Adder, Kitfox, WLF
40 points - CDA-2A. CDA-3M, Ice Ferret, Viper
45 points - CDA-3M, CN9, ENF, Arctic Wolf
50 points - GRF, SHD, WVR
55 points - BSW, DRG, HBKIIC, Nova
60 points - QKD, Stormcrow
85 points - BLR, Ebon Jaguar, Gargoyle, STK, Timber Wolf
100 points - Atlas, King Crab, Marauder IIC
110 points - Annihalator, Dire Wolf, Kodiak

Again, the above is just a quick example. Deeper consideration.would need to be given where each mech model and sometimes variant would be, but isn't it time we considered that maybe the Match Maker should not consider all identical tonnage mechs as equal? Hopefully Quirks and Weapon changes could be reduced by such a system.

[El Bandito]

Technically PGI already started giving points to mechs with Solaris' division system. Thing is, it is very hard to accurately gauge all mechs' power level, since there are hundreds of variants with their own strengths and weaknesses. PGI can try, but I doubt they can satisfy the community by a long shot.

Edited by El Bandito, 12 June 2018 - 12:01 PM.

[MechaBattler]

Might as well go with Battle Value. Every piece of gear used to determine overall value.

[Nightbird]

BV useless even in TT. No armor no HS mechs with insane BV possible lol.

[MechaBattler]

Well doesn't have to be exactly to TT's BV system. Obviously things like hard point placement and number of hard points of each type. As well as hitbox sizes relative to others in it's class would have to be taken into account.

[SilentScreamer]

El Bandito, on 12 June 2018 - 12:00 PM, said:

Technically PGI already started giving points to mechs with Solaris' division system. Thing is, it is very hard to accurately gauge all mechs' power level, since there are hundreds of variants with their own strengths and weaknesses. PGI can try, but I doubt they can satisfy the community by a long shot.

You are right, PGI ranked mrchs into brackets for Solaris7. That MIGHT be usable with Quickplay, but the likely result would be same maych quality as now with dramatcally increased match wait times. I take it from your response the divisions aren't working well? I haven't played Solaris7.

MechaBattler, on 12 June 2018 - 12:07 PM, said:

Might as well go with Battle Value. Every piece of gear used to determine overall value.

Going to a BV would be too much for the Match Maker to handle. Ideally we would want to increase match quality withput tanking wait times.

[process]

It seems to be a more direct goal to try and balance by tonnage over any other variable, especially since your point system is essentially just a proxy for tonnage. The tolarance a point system provides could just as well be converted into compensatory quirks or stat adjustments.

Solaris does do similar scale, but I would argue that's largely because Solaris caters to very narrow play styles that don't represent the roles available in a team setting.

[Andi Nagasia]

the Tonnage Paul was refering to was Clan Guass Builds, and the weight of C-Gauss weapon,
as of all the Clan Weapons C-Gauss is the least balanced to its IS Counterpart,

Mainly C-Gauss is 3Tons Lighter 1Crit Smaller, and has 10% more chance to Explode,
the 10% greater chance to Explode doesnt make up for the 3Tons 1Crit saved,
which is why Paul said something needs to be done about it,

Edit-
also having a BattleValue system like you propose may not help in mwo,
as you can have 2 players build a LCT-1E very different ways, and both may not be equal,

to make it work well you would have to have it calculate all the equipment in the Mech,
much like a True BattleValue system, which is allot more for the MM to try to Balance,

Edited by Andi Nagasia, 12 June 2018 - 02:56 PM.

[SilentScreamer]

process, on 12 June 2018 - 02:37 PM, said:

It seems to be a more direct goal to try and balance by tonnage over any other variable, especially since your point system is essentially just a proxy for tonnage. The tolarance a point system provides could just as well be converted into compensatory quirks or stat adjustments.

Clinging to a tonnage system requires Quirking of the Enforcer 5D till it equals a Nova since they are both 50 tons. If we gave up Tonnage = Equals Tonnage and used a point system we could find a much better equivant opponent in match for a 30 ton Spider 5V till than a 30 ton Arctic Cheetah. Hopefully then Range/Heat/Cooldown Quicks could be kept down to 5% or 10% and Armor/Structure Quirks could lowered as well.

Andi Nagasia, on 12 June 2018 - 02:53 PM, said:

the Tonnage Paul was refering to was Clan Guass Builds, and the weight of C-Gauss weapon,
as of all the Clan Weapons C-Gauss is the least balanced to its IS Counterpart,

Mainly C-Gauss is 3Tons Lighter 1Crit Smaller, and has 10% more chance to Explode,
the 10% greater chance to Explode doesnt make up for the 3Tons 1Crit saved,
which is why Paul said something needs to be done about it,

I believe you misread my original post. While you made a valid observation, specific builds and Alpha Meta are exactly NOT where I wanted this discussion to go. Last time I looked there were 18 pages over there.

Edited by SilentScreamer, 12 June 2018 - 03:50 PM.

[process]

SilentScreamer, on 12 June 2018 - 03:15 PM, said:

Clinging to a tonnage system requires Quirking of the Enforcer 5D till it equals a Nova since they are both 50 tons.

I don't see a problem here, other than PGI's ability to do so.

[SilentScreamer]

process, on 12 June 2018 - 03:52 PM, said:

I don't see a problem here, other than PGI's ability to do so.

Nova was 2014 right? So after 4 years maybe PGI should try a different approach?

[El Bandito]

SilentScreamer, on 12 June 2018 - 02:23 PM, said:

You are right, PGI ranked mrchs into brackets for Solaris7. That MIGHT be usable with Quickplay, but the likely result would be same maych quality as now with dramatcally increased match wait times. I take it from your response the divisions aren't working well? I haven't played Solaris7.

Divisions are not working well, cause you can't jam over 200 variants into 7 ranks and not expect a lot of them to be useless. To be fair to PGI, their changes to Solaris is only starting.

Edited by El Bandito, 12 June 2018 - 05:53 PM.

[Tarogato]

Any type of mech rating system ceases to matter once you are actually in the game fighting other mechs.

Who cares that your SDR-5V is only worth 15 arbitrary points when it has to play in the same match as every other mech?

Let's say instead of tonnage, Quickplay matchmaker does use an arbitrary BattleValue-like system. Who cares that your SDR-5V is worth 15 points when the enemy team is compromised of also one SDR-5V... as well as nine assaults? How are the SDR-5V's supposed to accomplish anything if they are not competitive with the other mechs on the battlefield?

Sorting mechs into all kinds of different ratings doesn't matter if you aren't actually separating them into separate buckets.

As for Solaris, we already have it. It called Divisions. There are seven of them and they do not go by tonnage. They are also extremely poorly thought out, but that's beside the point.

[FupDup]

Tarogato, on 14 June 2018 - 11:26 AM, said:

Any type of mech rating system ceases to matter once you are actually in the game fighting other mechs.

Who cares that your SDR-5V is only worth 15 arbitrary points when it has to play in the same match as every other mech?

Let's say instead of tonnage, Quickplay matchmaker does use an arbitrary BattleValue-like system. Who cares that your SDR-5V is worth 15 points when the enemy team is compromised of also one SDR-5V... as well as nine assaults? How are the SDR-5V's supposed to accomplish anything if they are not competitive with the other mechs on the battlefield?

Sorting mechs into all kinds of different ratings doesn't matter if you aren't actually separating them into separate buckets.

As for Solaris, we already have it. It called Divisions. There are seven of them and they do not go by tonnage. They are also extremely poorly thought out, but that's beside the point.

I don't support such a system, but I think that you'd have to be matching it so every individual player had a similar BV rather than having a "team pool" with everybody's combining together.

[SilentScreamer]

Tarogato, on 14 June 2018 - 11:26 AM, said:

Any type of mech rating system ceases to matter once you are actually in the game fighting other mechs.

Who cares that your SDR-5V is only worth 15 arbitrary points when it has to play in the same match as every other mech?

I keep telling people in this thread, specific build balance is not what this proposal is suppose to fix. Yes, builds matter, but the PILOT matters even more than the build. So do not worry about specific builds or pilots, team A and team B have equal chances with the match maker. The point of this thread is:

REDUCE QUIRKS

How? Mechs classified/rated by the matchmaker according to performance not tonnage.

Why? PGI constantly trys to adjust mechs so every 50 ton Clan and I.S. mech has identical performance in-match to every other 50 ton mech. When a new mech of a certain tonnage is introduced and overperforms other existing mechs of that weight get new Quirks to compensate...which in turn means others weights need boosts. This is called Power Creep, it is bad for game longevity.

Because: PGI is constantly messing with Quirks for outliers that Overperform: Timberwolf, Kodiak, Annihalator, Arctic Cheetah, Stormcrow; or Underperform: Spider 5V, Locust 1V, Enforcer 5D, Orion, Awesome PGI could just make a quick adjustment to the mech's point rating. Simpiler, easier, no Power Creep.

Edited by SilentScreamer, 16 June 2018 - 08:28 AM.

[Tarogato]

SilentScreamer, on 16 June 2018 - 08:02 AM, said:

REDUCE QUIRKS

How? Mechs classified/rated by the matchmaker according to performance not tonnage.

[...]
PGI is constantly messing with Quirks for outliers that Overperform: Timberwolf, Kodiak, Annihalator, Arctic Cheetah, Stormcrow; or Underperform: Spider 5V, Locust 1V, Enforcer 5D, Orion, Awesome PGI could just make a quick adjustment to the mech's point rating. Simpiler, easier, no Power Creep.

Who wants to play a mech that is crap? I don't care that a mech gets their a "point rating" lowered, and I don't care that the matchmaker is giving me a perfect 50/50 odds at winning. Neither of those are gonna make me play a crap mech. But quirks will.

Sure, by all means uses some type of BattleValue or performance-metric-informed ratings in order to improve the quality of matchmaking, I support that. But that will do nothing to improve mech balance, and to think that such a system would enable you to reduce quirks demonstrates a fundamental misunderstanding of the difference between balancing playable characters (mechs) and balancing players against each other (matchmaking).

[SilentScreamer]

Tarogato, on 18 June 2018 - 07:17 PM, said:

Who wants to play a mech that is crap? I don't care that a mech gets their a "point rating" lowered, and I don't care that the matchmaker is giving me a perfect 50/50 odds at winning. Neither of those are gonna make me play a crap mech. But quirks will.

We'll have to disagree then.

I re-state - not every 30 ton mech can be an Arctic Cheetah; especially a Spider 5V. The Quirks you would have to give it would be beyond even what we saw during CW Phase 2. So don't bother trying to Quirk it that much, just have the MM throw it in match to something more comparable.

Tarogato, on 18 June 2018 - 07:17 PM, said:

I support that. But that will do nothing to improve mech balance,

Again, I disagree. I think you might want to broaden your mind. Quirks are a crutch for mechs with bad hitboxes/hardpoints. Some of them so bad, no Quirks assigned thus far have made them viable. So start rating mechs seperate from tonnage and we won't have to deal with absurd levels Quirks.

Also, if all mechs were truly balanced, the Leaderboard would show Light and Mediums as on level with Heavy and Assault Leaderboards. That will not happen unless Quirks become even more bloated than they are now.

Tarogato, on 18 June 2018 - 07:17 PM, said:

and to think that such a system would enable you to reduce quirks demonstrates a fundamental misunderstanding of the difference between balancing playable characters (mechs) and balancing players against each other (matchmaking).

Matchmaking looks at BOTH Skill and the Mech. I know you read one of Paul's more recent posts on the matchmaker because you replied to it. Did you not understand it or just write your reply poorly? Also, mechs could only be considered a MMO "Class" capable of balance if you locked their equipment or play them Stock. There are too many possible builds to apply MMO logic here.

Paul Inouye, on 30 May 2018 - 02:10 PM, said:

So as I mentioned on Twitter, I stole the keys to the Match Maker configuration tool. My main purpose was to look at Solo Queue and it's current performance and the fact that Tier 1 players and Tier 4 players were being allowed in the same match right from the get go.

My main thought process was to identify the major time hogs when it came to creating the teams of individual players.

Here's what WAS happening:

1) Player skill Tier was set to 3. (Tiers are 0 based; meaning If I wanted Tier 1 and Tier 2 to be the only Tiers mixed, I'd have to set this value to 1). This value was set to 3 because of complaints of long wait times when it was reduced to 2. Theoretically, this number should be set to 0 BUT that would result in Tier 1 and Tier 2 players sitting in queue for a VERY long time.. think like 45 mins +. That being said, players capable of playing at Tier 1 to Tier 3 skill levels will allow closer competitive gameplay while maintaining shorter wait times for matches, so a value of 2 would be ideal here.

2) The Match Maker was relying on building teams using the 3/3/3/3 weight class restriction rule. We all know when looking at the weight class distribution in the current quick play queue information that this is going to be hard pressed when you have 12% of queued players playing lights, 15% in mediums, 60% in heavies and 33% in assaults. When you have queues like that in conjunction with the restraints of the above point dealing with Tiers, you can see where this is compounding to the problem.

2a) To add to the above point, the match maker would refuse to budge on 3/3/3/3 building for a minimum of 60 seconds before releasing the restriction slowly to allow a combination of up to 5 per weight class. THIS was the key to the wait times being seen in Solo QP. Once the valves opened up, teams were being created a lot faster.

Edited by SilentScreamer, 19 June 2018 - 10:34 AM.

[Tarogato]

SilentScreamer, on 18 June 2018 - 07:55 PM, said:

I re-state - not every 30 ton mech can be an Arctic Cheetah; especially a Spider 5V. The Quirks you would have to give it would be beyond even what we saw during CW Phase 2. So don't bother trying to Quirk it that much, just have the MM throw it in match to something more comparable.

If the SDR-5V needs a 60% cooldown quirk to be viable, then give it a 60% cooldown quirk. What's laughable is that it used to have a 30% cooldown quirk, and then it was nerfed. The literal worst mech in the game... got a 5% nerf. And nobody plays it.

And let me repeat what I said before - matchmaker is irrelevant to mech balance. Nobody wants to play crap mechs. They are all played on the same battefield anyway. There are no special servers where the SDR-5V is competitive with the other mechs fielded around it. The SDR-5V will still encounter Hellbringers and Annihilators and Huntsman and Warhammers. Who wants to be in that position? Only the masochistic and oblivious choose to play with such a handicap.

Show me how your matchmaking system would make playing the SDR-5V more appealing.

[FupDup]

Tarogato, on 19 June 2018 - 11:02 AM, said:

If the SDR-5V needs a 60% cooldown quirk to be viable, then give it a 60% cooldown quirk. What's laughable is that it used to have a 30% cooldown quirk, and then it was nerfed. The literal worst mech in the game... got a 5% nerf. And nobody plays it.

And let me repeat what I said before - matchmaker is irrelevant to mech balance. Nobody wants to play crap mechs. They are all played on the same battefield anyway. There are no special servers where the SDR-5V is competitive with the other mechs fielded around it. The SDR-5V will still encounter Hellbringers and Annihilators and Huntsman and Warhammers. Who wants to be in that position? Only the masochistic and oblivious choose to play with such a handicap.

Show me how your matchmaking system would make playing the SDR-5V more appealing.

I think he's implying that mechs like the Spider 5V would only encounter other similarly trash mechs (i.e. ALL players in the match using crap mechs not just one or two guys). From a "relative balance" standpoint the 5V wouldn't feel as weak when it's only fighting other 5V's and isn't fighting Hellbringers or other meta robots.

I don't agree with his stance though, I'm with you in wanting all gundams to be viable on a level playing field.

[SilentScreamer]

Tarogato, on 19 June 2018 - 11:02 AM, said:

Show me how your matchmaking system would make playing the SDR-5V more appealing.

Did you miss example list in the first post? It does not list every mech but the Spider 5V is there with the Locust 1V and Locust 1M in the lowest bracket?

FupDup, on 19 June 2018 - 11:09 AM, said:

I think he's implying that mechs like the Spider 5V would only encounter other similarly trash mechs (i.e. ALL players in the match using crap mechs not just one or two guys). From a "relative balance" standpoint the 5V wouldn't feel as weak when it's only fighting other 5V's and isn't fighting Hellbringers or other meta robots.

I don't agree with his stance though, I'm with you in wanting all gundams to be viable on a level playing field.

Yes FupDup, eventually the "crap" mechs would move down in the point system while better mechs moved up. Giving the MatchMaker enough players using those mechs for a complete 12v12 might never happen.

As for your last point. Ideally yes, all mechs would be equally playable. The Leaderboards and various challenges show they are not all equal. So do we try to deal with the bad mechs as they are (which is to not play them) or try to find a way to bring them out of mothball?