69 replies to this topic

[Tombstoner]

Mechs by themselves are not what is being balanced. its the players. BV or tonnage is not the issue, its the quality of the players and synergy in play-styles that helps make a good PUG. IE not loosing 2-3 mechs in the first 90 seconds. If you can avoid that chances are you might get a good game.

[TygerLily]

BV would be the best implementation for matching Mechs, in my opinion. Then, "bad variants" would be a steal for your team. Plop in some ace pilots and and put the greenhorns in the "good variants" and you can have a large tonnage for the same price as the other guys lowering their overall tonnage to bring the "good variants"...

And, maybe they discard the heatscale and instead have a BV scale. Want to bring that AC/20 Jager, sure, but it will be f*cking expensive! Bring two or three metas...and you'll only have points left to fill your team with Locusts and Hunchy 4J's...But again, plop in some ace pilots for those "weaker" chassis and you will be just fine.

In a way, this system could make your skill more valuable than your build...I'm all for that. I also think the ELO being assigned a BV would be good...but I haven't thought out how that jives with what I just described.

As far as, "Things are unbalanced. It will never work..." I don't think it's fair to expect things to be perfect before implementing a system...It's hard to see how BV it could make things any worse!!

Edited by TygerLily, 03 March 2014 - 08:00 AM.

[IceSerpent]

Tombstoner, on 03 March 2014 - 07:56 AM, said:

Mechs by themselves are not what is being balanced. its the players. BV or tonnage is not the issue, its the quality of the players and synergy in play-styles that helps make a good PUG. IE not loosing 2-3 mechs in the first 90 seconds. If you can avoid that chances are you might get a good game.

These are two separate issues. Both mechs and players need to be balanced, players by properly implemented Elo and mechs by properly implemented BV. The only relationship between the two is that you can add them together and get some sort of overall "combat value". You don't really need to do that though, as you can have them as two separate metrics in MM. Either way, whatever scheme you use to balance mechs has no bearing on whatever scheme you use to balance players.

[Tombstoner]

IceSerpent, on 03 March 2014 - 08:03 AM, said:

These are two separate issues. Both mechs and players need to be balanced, players by properly implemented Elo and mechs by properly implemented BV. The only relationship between the two is that you can add them together and get some sort of overall "combat value". You don't really need to do that though, as you can have them as two separate metrics in MM. Either way, whatever scheme you use to balance mechs has no bearing on whatever scheme you use to balance players.

But player and mech interact. mechs in MWO are inherently biased by art work and hard point location. any build point system needs to account for those two things. for example the awesome and the stalker. energy point location matters as does CT profile. how are you going to assign a BP value for that... thats why you have to take it out of the equation and base it only on the player and let the mech choice float. you can try to balance by mech but ever one is given the same 100 ton allocation per match.

[IceSerpent]

Tombstoner, on 03 March 2014 - 01:02 PM, said:

But player and mech interact.

They are still separate "values". Both should be accounted for, but one is not a substitute for the other. You can easily have a situation where great player in a lousy mech is equal (as far as battle performance is concerned) to a bad player piloting a meta-build.

Quote

mechs in MWO are inherently biased by art work and hard point location. any build point system needs to account for those two things. for example the awesome and the stalker. energy point location matters as does CT profile. how are you going to assign a BP value for that...

In a perfect world you wouldn't have to, because hit boxes / hardpoint locations would out balanced by other quirks. I.e. STK has high mounted energy hardpoints, but also has huge side torsos and is not XL-friendly. Given that PGI's world is not perfect, you can always have a backup plan of certain "good" mech variants adding a bonus value to BV. BV calculation is nothing but a sum of different values, so adding extra ones is beyond being easy.

Quote

thats why you have to take it out of the equation and base it only on the player and let the mech choice float. you can try to balance by mech but ever one is given the same 100 ton allocation per match.

Can't balance by player skill only when people can field vastly different equipment - it simply won't equalize anything. If we were to go shoot some clay pigeons and you get a flint lock pistol while I get a modern shotgun, it won't be a fair game even if our skills are exactly the same.

[Zerberus]

The basic idea and principle are sound. The rat is healthy.

However, The understanding of the Rat`s tail (the extended implications) is not.

Tabletop BV as already written is both inacurrate for MWO from a value /item perspective, and it was already an exploitable system when it was introduced. For ex. a Stalker bristling with MGs has a lower BV than a slashback, but will easily tear it to shreds F2F.

This produces 2 significant issues that only lots of development time and /or brainstorming can rectify, and a third that is insurmountable.:

1. Complete reallocation of all BV values to each and every item present and future. This part is actually the less daunting part as teh data could at least partly be gleaned from existing statistics.

2. Modifying the existing BV system to no longer be gameable. Ideas have been presented for this for almost 20 years now, and none have proven feasible.

Otherwise you will only push min/maxing to a new dimension, with teh added implication of people with a BV that is exceedingly high or too low (for ex. 6x[er]PPC stalker or 4MG locust) will never find a match in a reasonable amount of time until they change their loadout.

So you risk heavily and arbitrarily restricting loadout and mech choise on one side, while having a minmax scene on the other side complaining about weapon x being way too high /low in BV becasue weapon y has almost identical stats but (z), and choosing their loadouts so that they have the lowest possible score while retaining the highest possible firepower, exploiting any point that is "false", and thereby skewing the data collection until other weapons become more viably from a BV /damage perspective. At which time they switch to those weapons until the metrics say "these are overused and therefore have a high BV"

Side note: with an automated, use /kills based system, a Medium laser would have a BV of, like, 1000, and the flamer would have, like, 5, less than a MG. I dare say a flamer is a lot more dangerous than those numbers would imply... not as much as the medium laser, but certainly not 10000x worse, and probably equal to a machine gun... but people don`t take tons of them, so they don`t get many kills. On that note LRMs don`t generally get many kills, either, just lots of damage and a boatload of assists....

3. In a perfect world, the idea has merit. But last time I checked, ours is far from perfect, and we as a species are one of the primary engines behind that. I.e. It probably won`t work as long as humans are involved.

That is why BV will probably never work here on a broad scale, ESPECIALLY not one that worky dynamically based on kill rate and damage done by the community in mech x w/ weapon y. Most people will not willingly change their loadouts just becasue some arbitrary system essentially says (in the form of matchmaker fails) "I`m sorry, you`re too dangerous a pilot to be using that weapons system on that assault mech, please take something lighter and /or adjust the entire rest of your build to reflect that". Nor would a bad pilot be able to have fun in light mechs, because "You suck too much to run a mech that light, please ton up and be of use as a meat shield or at least mount a ppc"

That all said, nobody is stopping the 12 man /league scene from implementing such a system for their matches, and this is where it may have actual merit as opposed to the public queues.

Edited by Zerberus, 03 March 2014 - 01:51 PM.

[IceSerpent]

Zerberus, on 03 March 2014 - 01:30 PM, said:

This produces 2 significant issues that only lots of development time and /or brainstorming can rectify:

1. Complete reallocation of all BV values to each and every item present and future. This part is actually the less daunting part as teh data could at least partly be gleaned from existing statistics.

It's not necessary to allocate values to each and every item.

Quote

2. Modifying the existing BV system to no longer be gameable. Ideas have been presented for this for almost 20 years now, and none have proven feasible.

Suggestion doesn't include any existing BV system to be modified.

Quote

Otherwise you will only push min/maxing to a new dimension, with teh added implication of people with a BV that is exceedingly high or too low (for ex. 6x[er]PPC stalker or 4MG locust) will never find a match in a reasonable amount of time until they change their loadout.

6 PPC STK shouldn't have very high BV due to build being bad, meta HGN should have higher BV for example. 4MG Locust would indeed have very low BV, but its problems with finding a match can be rectified by allowing uneven teams.

[Zerberus]

IceSerpent, on 03 March 2014 - 01:39 PM, said:

It's not necessary to allocate values to each and every item.

Yes, it most definitely is, because there is not a single item that you can put on your mech that will not increase or lower it`s battlefield potential. Every HS, every ton of ammo, hell every point of armor increases your battlefield potential, as does every module (and even some camo schemes, but BVing those is getting a bit too **** about it). The whole point of BV is to have a guage of battlefield potential, and now things that directly influence that potential are suddenly non important /do not need a BV attached to them?

On that note, a ton on ammo for a weapon you don`t have should actually lower it a bit, as it`s an unnecessary explosion risk.

Quote

Suggestion doesn't include any existing BV system to be modified.

Then one has to be created entirely from scratch, which is even more work than modifying an existing one.

Quote

6 PPC STK shouldn't have very high BV due to build being bad, meta HGN should have higher BV for example. 4MG Locust would indeed have very low BV, but its problems with finding a match can be rectified by allowing uneven teams.

And who tells the system what is "good" and what is "bad", or above what number a weapon suddenly drops so radically in BV that it actually lowers the BV of the basic chassis? You? Me? PGI? The community? Are you starting to see the problem?

Becasue otherwise it sees
Stalker = value x
+ 6ppcs = value Y
STK + 6ppc = value z

Uneven teams in public matches are an unnecesary cost factor for the same basic technical reasons that private matches w/ uneven teams require 2 premium accounts to start and may eventually be pay by use. And what you do do if the locust would have been number 12? Magically create a 13th slot and watch the other team either A: disconnect because of perceived imbalance or b: start drooling becasue they know there`s 2 locusts on the team.

Or taken to it`s logical conclusion: What do you do if 3 friends group upo with 3 locusts? Make a 15vs 12 match or force 12 vs 9, making teh team of 12 that has been sitting in the matchmaker for 5 minutes fail because the MM has to create a whole new special match for the locusts and fill it with other special needs cases? and once again, who`s paying for the additional server overhead, or are they all free players?

See the very tip of the resulting problems yet? Money being one of the largest, with Human nature still holding the lead....

Make the game into a subscription model and some of those problems go away or at least become smaller, but by far not all of them, much less teh actually important ones.... but then a lot of freeloaders will go away, killing teh playerbase for everyone...

So any rational solution for how to implement BV into MWO has to include a solution for solving all of those issues and more. I neither have one nor do I have the time to dream one up that is less flawed than the one BT already has. maybe someone else does.... But I haven`t see a single suggestion between closed beta and now that could realistically be made to fly in the public queue.

Edited by Zerberus, 03 March 2014 - 02:31 PM.

[Varent]

IceSerpent, on 03 March 2014 - 07:28 AM, said:

Varent, you are overcomplicating things a bit except for JJs that definitely should be included in BV calc. Weapons themselves are fairly easy - just check what damage you do at various ranges (i.e. at 100m intervals all the way to max range in game), throw in a few modifiers for guided/unguided and spread/pinpoint, and add everything together to get BV of your weapon setup. That way you don't end up with wonky values.

not really... mechs function very differently and play differently if you don't build them right and when you throw together certain combination. The hit boxes of mechs in general also make them more well suited to different tasks. This needs to be accounted for. Otherwise you will just replace one broken system for another broken system. No one wants that.

[Illuzian Pryde]

Because it's exactly the same as weight with launch module.

We get long waits as is. Implementing anything as refined as either BV or Weight only, especially when bringing a selection of 4 mechs would severely limit choice, fun and diversity.

You'd either see serious min/maxing issues where lights/heavies are the only mechs your see or you'd see constant cheese.

As explained by Bryan, having l/m/h/a means you can choose 4 mechs of any one class and choose any one of them. It also means that all of that class get deployed for diversity.

The game is competitive to a degree but considering HST etc it won't ever be that competitive without local servers, so why implement a hardcore mechanic like that that severely limits what load outs people bring?

[Roland]

I'm working on putting together a detailed version of my earlier suggestion of a dynamic, market-based BattleValue system.

As it stands, I pretty much have it fully laid out, and should be able to post up exactly how it can work pretty soon. I'm just having a few folks check my math before I throw it out there.

[Zerberus]

^^ Interesting... but out of curiosity, what do you mean by "market based" in this context?

Edited by Zerberus, 03 March 2014 - 03:24 PM.

[IceSerpent]

Zerberus, on 03 March 2014 - 02:24 PM, said:

Yes, it most definitely is, because there is not a single item that you can put on your mech that will not increase or lower it`s battlefield potential. Every HS, every ton of ammo, hell every point of armor increases your battlefield potential, as does every module (and even some camo schemes, but BVing those is getting a bit too **** about it). The whole point of BV is to have a guage of battlefield potential, and now things that directly influence that potential are suddenly non important /do not need a BV attached to them?

Wanna bet? I'll give you a hint - if you make mech speed increasing BV and heat efficiency increasing BV, you've covered all engines and heatsinks. Same goes for armor. It's not that things are unimportant, it's that you can combine most of those things into a few parameters that increase BV.

Quote

On that note, a ton on ammo for a weapon you don`t have should actually lower it a bit, as it`s an unnecessary explosion risk.

You can certainly do it this way, but personally I would skip ammo altogether, as its effect on the actual performance is minimal assuming that you didn't forget to add ammo for weapons you have and remove ammo for weapons you don't have. That being said, adding ammo-based bonus/malus is definitely doable.

Quote

Then one has to be created entirely from scratch, which is even more work than modifying an existing one.

We don't have an "existing one" for MWO, so creating a new one is what we are talking about.

Quote

And who tells the system what is "good" and what is "bad", or above what number a weapon suddenly drops so radically in BV that it actually lowers the BV of the basic chassis? You? Me? PGI? The community? Are you starting to see the problem?

Becasue otherwise it sees
Stalker = value x
+ 6ppcs = value Y
STK + 6ppc = value z

No, I don't see the problem. The value is based on speed of that STK + armor + sum of damage done by those 6 PPCs at various ranges (let's say with 100m intervals) + heat efficiency (which is stupid low) + bonus for high monted hardpoints if you want to include it.

If you apply the same approach to HGN jump sniper, you'll get much better heat efficiency, bonus for JJs, more armor, but less overall weapon damage (although AC5s not having minimum range would mitigate it somewhat).

Quote

Uneven teams in public matches are an unnecesary cost factor for the same basic technical reasons that private matches w/ uneven teams require 2 premium accounts to start and may eventually be pay by use.

I guess we should all go tell PGI that their idea of 10v12 balance for Clantech is an "unnecessary cost factor".

Quote

And what you do do if the locust would have been number 12? Magically create a 13th slot and watch the other team either A: disconnect because of perceived imbalance or b: start drooling becasue they know there`s 2 locusts on the team.

I don't even understand what you're saying - why do you want MM to number the mechs? There's no such thing as "mech number 12" or "mech number 3", or at least there shouldn't be (it's PGI we're dealing with, so one never knows what weirdness they come up with).

Quote

Or taken to it`s logical conclusion: What do you do if 3 friends group upo with 3 locusts? Make a 15vs 12 match or force 12 vs 9, making teh team of 12 that has been sitting in the matchmaker for 5 minutes fail because the MM has to create a whole new special match for the locusts and fill it with other special needs cases? and once again, who`s paying for the additional server overhead, or are they all free players?

Again, not sure whtat you're talking about. 3 Locusts would most likely end up in an uneven match if such matches are allowed (or in an even match with 3 equally crappy builds on the other side - not necessarily Locusts). Why would it cause some mystical team of 12 to fail?

Quote

See the very tip of the resulting problems yet? Money being one of the largest, with Human nature still holding the lead....

Nope, have yet to see a single one. Money (server overhead) is not a huge problem, as you don't need to go all the way to 1v1, you can limit it to 8v12 at most (probably 10v12 will already do the trick).

Varent, on 03 March 2014 - 02:32 PM, said:

not really... mechs function very differently and play differently if you don't build them right and when you throw together certain combination. The hit boxes of mechs in general also make them more well suited to different tasks. This needs to be accounted for. Otherwise you will just replace one broken system for another broken system. No one wants that.

What do you mean by "suited for different tasks"? High mounted hardpoints alowing for hill-humping? I think it's a completely separate thing from BV - it's only useful if you have hills to hump, doesn't give you much of a benefit in an urban fight.

[Roland]

Zerberus, on 03 March 2014 - 03:24 PM, said:

^^ Interesting... but out of curiosity, what do you mean by "market based"?

It's a system where the BV of a given mech variant is determined through it's relative usage by the playerbase.

That is, the more a mech is used, the higher its Battlevalue... Thus, it's effectively using a simulated "market" where using a mech constitutes a transaction.

The approach has a number of huge advantages.. Mainly, it does not require anyone to define what the "correct" value of any given mech is. The values are automatically determined by the usage patterns of the actual players. It requires very little maintenance, is easy to implement, and will automatically balance itself over time, even as the metagame changes and usage patterns adjust.

Indeed, it will actually instigate changes in the metagame, by providing incentive for using the least used mechs, and a disincentive for using the most popular mechs.

[Varent]

IceSerpent, on 03 March 2014 - 03:24 PM, said:

Wanna bet? I'll give you a hint - if you make mech speed increasing BV and heat efficiency increasing BV, you've covered all engines and heatsinks. Same goes for armor. It's not that things are unimportant, it's that you can combine most of those things into a few parameters that increase BV.



You can certainly do it this way, but personally I would skip ammo altogether, as its effect on the actual performance is minimal assuming that you didn't forget to add ammo for weapons you have and remove ammo for weapons you don't have. That being said, adding ammo-based bonus/malus is definitely doable.



We don't have an "existing one" for MWO, so creating a new one is what we are talking about.



No, I don't see the problem. The value is based on speed of that STK + armor + sum of damage done by those 6 PPCs at various ranges (let's say with 100m intervals) + heat efficiency (which is stupid low) + bonus for high monted hardpoints if you want to include it.

If you apply the same approach to HGN jump sniper, you'll get much better heat efficiency, bonus for JJs, more armor, but less overall weapon damage (although AC5s not having minimum range would mitigate it somewhat).



I guess we should all go tell PGI that their idea of 10v12 balance for Clantech is an "unnecessary cost factor".



I don't even understand what you're saying - why do you want MM to number the mechs? There's no such thing as "mech number 12" or "mech number 3", or at least there shouldn't be (it's PGI we're dealing with, so one never knows what weirdness they come up with).



Again, not sure whtat you're talking about. 3 Locusts would most likely end up in an uneven match if such matches are allowed (or in an even match with 3 equally crappy builds on the other side - not necessarily Locusts). Why would it cause some mystical team of 12 to fail?



Nope, have yet to see a single one. Money (server overhead) is not a huge problem, as you don't need to go all the way to 1v1, you can limit it to 8v12 at most (probably 10v12 will already do the trick).



What do you mean by "suited for different tasks"? High mounted hardpoints alowing for hill-humping? I think it's a completely separate thing from BV - it's only useful if you have hills to hump, doesn't give you much of a benefit in an urban fight.

I was actually reffering to the way many mechs are laid out with most of there weapons on one side, so you can block fully with one side and fight with the other and torso twist more effectively. Also many mechs have larger or smaller arms and larger or smaller side torso wich make them more effective at torso twisting. In addition to this you have mechs with higher mounted weapons yes and arms that are better at firing overall over things. You also have to consider the downside of certain chasis like the dragon with its massive center torso of death that everyone will target incredibly fast. These are things that make chasis good or bad. In addition to wether or not some of them can mount jump jets or ecm to give them added advantages in the battle field. Youneed to take all these things into account.

Roland, on 03 March 2014 - 03:27 PM, said:

It's a system where the BV of a given mech variant is determined through it's relative usage by the playerbase.

That is, the more a mech is used, the higher its Battlevalue... Thus, it's effectively using a simulated "market" where using a mech constitutes a transaction.

The approach has a number of huge advantages.. Mainly, it does not require anyone to define what the "correct" value of any given mech is. The values are automatically determined by the usage patterns of the actual players. It requires very little maintenance, is easy to implement, and will automatically balance itself over time, even as the metagame changes and usage patterns adjust.

Indeed, it will actually instigate changes in the metagame, by providing incentive for using the least used mechs, and a disincentive for using the most popular mechs.

this could have potential provided it accounts for three things and gives a player an overall value. It needs to account for 1) the value of the chasis, what your mentioning above, pretty much to a degree. 2) the value of weapon systems in the same degree. 3) the skill of the player and win loss in that chasis of mech.

[Willard Phule]

Zerberus, on 03 March 2014 - 02:24 PM, said:

Becasue otherwise it sees
Stalker = value x
+ 6ppcs = value Y
STK + 6ppc = value z

That's the cool part of min/maxing...at least from where I see it. Granted, there's beer involved at this point, so my brain isn't the same as it is in the morning after coffee..

So...some cretin puts 6 PPCs on a Stalker. Cool. He's artifically inflating his BV...at his and his team's detriment.

He's got tonnage and whatever for his base BV...and he added 6 PPCs. He can't fire them all together without shutting down...so, cool. He's got a GREAT BV.....and when he ends up using that crap build, he dies.

BV works real well for balanced mechs. Not so much for crap builds. That guy will quickly find himself in over his head.

[Zerberus]

Varent, on 03 March 2014 - 03:31 PM, said:

this could have potential provided it accounts for three things and gives a player an overall value. It needs to account for 1) the value of the chasis, what your mentioning above, pretty much to a degree. 2) the value of weapon systems in the same degree. 3) the skill of the player and win loss in that chasis of mech.

I agree entirely, but I think that Elo can handle number 3 well enough. Weapons most definitely need to be tracked as well, though.

Edited by Zerberus, 03 March 2014 - 03:35 PM.

[Roland]

Varent, on 03 March 2014 - 03:31 PM, said:

this could have potential provided it accounts for three things and gives a player an overall value. It needs to account for 1) the value of the chasis, what your mentioning above, pretty much to a degree. 2) the value of weapon systems in the same degree. 3) the skill of the player and win loss in that chasis of mech.

Turns out, it really doesn't need to account for what you actually put on the mech. Simply by looking at the usage of a given chassis (Note, each VARIANT is considered separately. This is key.), you can derive a value which will effectively achieve the desired goals of increasing mech variety... and since one of the primary differentiators of variants is the hardpoints, this ends up achieving a fairly good analog for a more complex system which accounted for individual weapons.

The weapons themselves end up being inherently accounted for by the system because weapons balance ends up being one of the driving forces in mech selection.... The reason why a certain variant is used, is because it is capable of carrying optimal weapons loadouts. The D-DC, for instance, gets more usage than other atlases... mainly because it carries ECM. The same goes for any of the mechs carrying ECM. The ECM variant inevitably gets used more than the alternatives. Thus, there's really no need to worry about the weapons' values themselves, as it all ends up coming out in the wash.

The only thing that really ends up being missed is that you don't really get any incentive to bring the best chassis and load it up with garbage weapons.. but that's really not important anyway.

Likewise, this isn't account for player skill at all. A battlevalue system isn't a replacement for a matchmaker. Instead, a battlevalue system provides a measurement of the mech's value itself, which can then be INCLUDED within a larger matchmaking system.

Although, that being said, the system I'm laying out actually DOES help provide some balancing of teams of various sizes, by providing bonuses to smaller teams and solo pilots, and imposing stricter BV limits on larger teams.

But again, a BV system isn't a replacement for a matchmaker. It's just one metric which can be used in matchmaking.

Edited by Roland, 03 March 2014 - 03:44 PM.

[IceSerpent]

Varent, on 03 March 2014 - 03:31 PM, said:

I was actually reffering to the way many mechs are laid out with most of there weapons on one side, so you can block fully with one side and fight with the other and torso twist more effectively. Also many mechs have larger or smaller arms and larger or smaller side torso wich make them more effective at torso twisting. In addition to this you have mechs with higher mounted weapons yes and arms that are better at firing overall over things. You also have to consider the downside of certain chasis like the dragon with its massive center torso of death that everyone will target incredibly fast. These are things that make chasis good or bad. In addition to wether or not some of them can mount jump jets or ecm to give them added advantages in the battle field. Youneed to take all these things into account.

You are mixing a few different things together here. All weapons on one side is a mixed blessing - you get that "shield" side, but if you lose that "good" side torso, you're done. Shielding arms / high mounted hardpoints are pretty much mutually exclusive - you get either one or the other (the only exception that comes to mind is Cataphract) and usefulness depends on the situation. Large CT is also kind of mixed thing - on one hand it's easy to hit, on the other hand you can use XL engine without increasing your vulnerability too much. All this stuff is too reliant on what you're fighting and where you're fighting to be included in matchmaking. Not to mention that quite a bit of it comes from weapon balance currently being messed up.
ECM and JJs definitely should be include in BV calc, but as mounted equipment, not a feature of a given mech. It doesn't matter that you can mount ECM or JJS, what matters is that you actually did mount them (or not).

[verybad]

Beacause battle values are considered innacurate even by the TT makers.Catalyst. They were going to remake them themselves...If they don't work well for the TT game, then how are they going to work for the video game?