54 replies to this topic

[SharDar]

Is there a formula that is used to achieve parity between different weapons? A weapon's effectiveness is based on:

• Damage
  
• Weight
  
• Range
  
• Heat
  
• Cooldown
  
• Accuracy
  
• Projectile speed, where applicable
  
• Guided/Non-Guided
  
• What else?

How does the Cauldron determine what values each weapon should have? To make sure a particular weapon is not OverPowered (OP), there are tradeoffs in these values. An energy weapon that shoots twice as far for the same damage should generate more heat. Is that or should that be twice as much heat or should it/does it use the inverse square rule where it generates 4 times as much heat. Likewise, a ballistic weapon should be less accurate at a greater distance.

Is it possible to create a formula with assigned coefficients that achieves this kind of parity? I'm not criticizing the efforts of the developers or the Cauldron; I'm just curious. Also, I used to work in a data intensive field, so I think about this kind of stuff.

[pbiggz]

I majored in Game Development in University. Design was my main area, so I can tell you with some authority that the sorts of formulas you're talking about can be useful tools, but they have their limitations. Reckless designers either ignore them completely, or overuse them to the point of letting their excel spreadsheets design games for them, and no design really survives contact with the player; its absolutely imperative that designers respond to the way players interact with their games. Just because a certain weapon has a similar DPS to another, doesn't mean some externality cant let players exploit it, or ignore it. In other words, you need to use your eyes. Math only gets you so far.

As for how the cauldron does its changes, I think they test rigorously, and pull feedback from a wide array of locations. I think they have some data access from PGI, but I can't speak to that authoritatively. I expect they're looking at some information, noting mechs and weapons that "underperform", as well as those that "overperform" and providing tweaks to edge the performance of those mechs and weapons to within an acceptable margin. It's how I would do it.

Edited by pbiggz, 19 January 2022 - 11:55 AM.

[ScrapIron Prime]

Heat, cooldown, damage output at a given range. You build your mech around the role you want it to have. Sniper, lurmer, brawler, generalist. Some folks like to build secondary weapons in so their snipers and brawlers won't have so much of an Achilles heel, but that dilutes their primary purpose.

Like pbiggz said, I can see using tables to figure this out, particularly with cooldown and range, but in reality... you build a mech, try it, and tweak it. My most consistent high scoring mech doesn't fit into any actuary table, and most people would consider it a potato. but it works for me. Find what works for you, even if you lose a lot of games in the process.

<wax on> </wax off>

[Zigmund Freud]

ScrapIron Prime, on 19 January 2022 - 12:00 PM, said:

My most consistent high scoring mech doesn't fit into any actuary table, and most people would consider it a potato. but it works for me. Find what works for you, even if you lose a lot of games in the process.

Build plz? I'm curious now.

[justcallme A S H]

How does a formula account for gameplay, player behaviour, maps, the things you are trying to kill etc etc.

Thus no formula exists.

If such a formula did exist, abd actually worked, no game would never need ongoing balance.

[ScrapIron Prime]

Zigmund Freud, on 19 January 2022 - 12:38 PM, said:

Build plz? I'm curious now.

Crom Dubh. I've posted it a few times here and there. Its designed to face off at 400m, getting its own locks by drilling with lasers, and having the heat capacity and ammo to not stop, even if the enemy breaks LOS. What you DON'T do with it is hang back at range, that's what pure LRM boats are for. This fellow shares armor and participates in pushes.

[LordNothing]

there are always unquantifiable variables. usually things with nonlinearity or randomness involved (like weapons with spread or dice roll jams). sometimes its not as obvious as that (ppfld is a good example) they tend to throw a monkey wrench into your math. you usually have to use testing to come up with fudge factors to compensate for those. and those are very fragile in lieu of changes.

id say balance everything first using your formulae.
determine which have known unquantifiable variables.
verify that the others do not have any hidden (non-obvious) unquantifiable variables.
do a comparative analysis in a real world test, use the difference in result to compute fudge factors.
apply fudge factors and then retest.
repeat last 2 steps until balance is obtained.

if you make broad changes to the quantifiable things, you just changed the baseline and need to re-determine all your fudge factors.
if you make small changes to the unquantifiable, you can re determine their fudge factors in isolation.
this can also be done with quantifiables but without the fudge factor step, so long as they are of equivalent performance to other quantifiables.
use deterministic mechanics whenever possible. it makes things easier to quantify.
the player is the biggest unquantifiable of all and will break your game more than any bug will.

Edited by LordNothing, 19 January 2022 - 01:01 PM.

[pbiggz]

justcallme A S H, on 19 January 2022 - 12:42 PM, said:

How does a formula account for gameplay, player behaviour, maps, the things you are trying to kill etc etc.

Thus no formula exists.

If such a formula did exist, abd actually worked, no game would never need ongoing balance.

They can be valuable tools for establishing a system, in other words, they're a great starting point for a new game; in the case of mwo, their value is far far far more limited.

[LordNothing]

pbiggz, on 19 January 2022 - 11:53 AM, said:

I majored in Game Development in University. Design was my main area, so I can tell you with some authority that the sorts of formulas you're talking about can be useful tools, but they have their limitations. Reckless designers either ignore them completely, or overuse them to the point of letting their excel spreadsheets design games for them, and no design really survives contact with the player; its absolutely imperative that designers respond to the way players interact with their games. Just because a certain weapon has a similar DPS to another, doesn't mean some externality cant let players exploit it, or ignore it. In other words, you need to use your eyes. Math only gets you so far.

As for how the cauldron does its changes, I think they test rigorously, and pull feedback from a wide array of locations. I think they have some data access from PGI, but I can't speak to that authoritatively. I expect they're looking at some information, noting mechs and weapons that "underperform", as well as those that "overperform" and providing tweaks to edge the performance of those mechs and weapons to within an acceptable margin. It's how I would do it.

im one of those people who think that live testing is good. lots of gamers see that as being "too cheap to hire professional testers", but i think even with a huge budget for testers, you still get more eyes on it in the live environment. they might not be as good at filing bug reports, but having a lot of eyes, people recording things and screencapping things, and sending in error logs, can generate a huge amount of data in a very short period of time. your internal testers are better off sifting through that data for issues, confirm them first hand, and propose tweaks for the dev staff to implement, than they are to play the game in test environment and report anything they see out of whack. why pay a dozen people to do a job that gamers pay you for.

Edited by LordNothing, 19 January 2022 - 01:09 PM.

[FupDup]

pbiggz, on 19 January 2022 - 11:53 AM, said:

I majored in Game Development in University. Design was my main area, so I can tell you with some authority that the sorts of formulas you're talking about can be useful tools, but they have their limitations. Reckless designers either ignore them completely, or overuse them to the point of letting their excel spreadsheets design games for them, and no design really survives contact with the player; its absolutely imperative that designers respond to the way players interact with their games. Just because a certain weapon has a similar DPS to another, doesn't mean some externality cant let players exploit it, or ignore it. In other words, you need to use your eyes. Math only gets you so far.

As for how the cauldron does its changes, I think they test rigorously, and pull feedback from a wide array of locations. I think they have some data access from PGI, but I can't speak to that authoritatively. I expect they're looking at some information, noting mechs and weapons that "underperform", as well as those that "overperform" and providing tweaks to edge the performance of those mechs and weapons to within an acceptable margin. It's how I would do it.

As I would put it: Game balance is qualitative, not quantitative.

[1453 R]

ScrapIron Prime, on 19 January 2022 - 12:44 PM, said:

Crom Dubh. I've posted it a few times here and there. Its designed to face off at 400m, getting its own locks by drilling with lasers, and having the heat capacity and ammo to not stop, even if the enemy breaks LOS. What you DON'T do with it is hang back at range, that's what pure LRM boats are for. This fellow shares armor and participates in pushes.

Off-topic, but I remember you talking to me about that Highlander of yours before and how it just fits. Heh...just struck me, looking at your fit, how eerily similar it is to Spicy Lady, my "it just works" mainstay. Heh...two large lasers, two medium lasers, pair of LRM-15s, crammed full of as many heat sinks as fit and with a smattering of support equipment, and both generally angling for the same midrange whoosh-and-burn engagements. Lady's significantly faster and hits harder overall, but Dubh's tougher and runs a lot cooler even with all the cDHS in Lady. Be interesting to see how they performed together, or even head to head, heh.

Robot fistbump, Scrapiron. Heh, never let comps tell you a 'Mech doesn't work when your own-a$% experience says otherwise.

[ScrapIron Prime]

1453 R, on 19 January 2022 - 04:10 PM, said:

Off-topic, but I remember you talking to me about that Highlander of yours before and how it just fits. Heh...just struck me, looking at your fit, how eerily similar it is to Spicy Lady, my "it just works" mainstay. Heh...two large lasers, two medium lasers, pair of LRM-15s, crammed full of as many heat sinks as fit and with a smattering of support equipment, and both generally angling for the same midrange whoosh-and-burn engagements. Lady's significantly faster and hits harder overall, but Dubh's tougher and runs a lot cooler even with all the cDHS in Lady. Be interesting to see how they performed together, or even head to head, heh.

Robot fistbump, Scrapiron. Heh, never let comps tell you a 'Mech doesn't work when your own-a$% experience says otherwise.

Heck yeah.
May our unlikely mechs meet on the field some day!

[justcallme A S H]

pbiggz, on 19 January 2022 - 12:53 PM, said:

They can be valuable tools for establishing a system, in other words, they're a great starting point for a new game; in the case of mwo, their value is far far far more limited.

Yep could for sure be useful when doing initial designs before you get onto the field to play and see how it all feels/works in the game itself.

Right now for MWO - not needed. Most of the values have been, apart from the initial bulk changes, just tweaks here/there based on mostly gameplay/feel and community feedback.

Some changes of course we math/test in lobbies/training grounds out pretty hard and in depth some probably too much. That isn't needed in every case though.

We didn't need a formula to know cLPL was overperforming slightly in QP. Now is the little nerf enough, again gotta let it play out for a month or two to see.

[feeWAIVER]

No formula? Yikes. Kinda explains the cycle of mudflation that's bringing us around to buffing Blood Asps and Battlemasters.

Yes, game development uses formula's and math to seek balance.
Just like good players use formula's and math to understand balance.
Just look at Filthy Robot, who breaks every game he plays down into a spreadsheet to analyze the data and find the best opportunity costs in his games.

Even DATA here in the MWO community openly talks about using formulas and math to guide his decisions.

Edited by feeWAIVER, 19 January 2022 - 05:24 PM.

[justcallme A S H]

Yes no formula was needed to know cLPL was overperforming. It was pretty much unanimously agreed by all in Cauldron and also various areas of feedback. That might come as a surprise to the odd person, as always.



You can look at the details stats and all the comparables in the spreadsheets at any time with the more details and some numbers and calculations that are taken into consideration.

https://www.mwocomp.com/patches.html

Edited by justcallme A S H, 19 January 2022 - 05:25 PM.

[feeWAIVER]

justcallme A S H, on 19 January 2022 - 05:24 PM, said:

You can look at the details stats and all the comparables in the spreadsheets at any time with the more details and some numbers and calculations that are taken into consideration.

https://www.mwocomp.com/patches.html

So..
Formulas are being used, just by the smarter people.
And then the go-along-gang can eyeball it and agree.

[FupDup]

feeWAIVER, on 19 January 2022 - 05:17 PM, said:

No formula? Yikes. Kinda explains the cycle of mudflation that's bringing us around to buffing Blood Asps and Battlemasters.

Yes, game development uses formula's and math to seek balance.
Just like good players use formula's and math to understand balance.
Just look at Filthy Robot, who breaks every game he plays down into a spreadsheet to analyze the data and find the best opportunity costs in his games.

Even DATA here in the MWO community openly talks about using formulas and math to guide his decisions.

Formulas can only do so much. In terms of whether something is OP or not, they can tell you how many people use a weapon, the win rate of players using it, etc.

But math doesn't tell you why you got there or what to do about it. The formula won't tell you if you need to nerf heat, damage, cooldown, range, or whatever else. That's where the qualitative human element comes in.

[feeWAIVER]

FupDup, on 19 January 2022 - 05:32 PM, said:

Formulas can only do so much. In terms of whether something is OP or not, they can tell you how many people use a weapon, the win rate of players using it, etc.

But math doesn't tell you why you got there or what to do about it. The formula won't tell you if you need to nerf heat, damage, cooldown, range, or whatever else. That's where the qualitative human element comes in.

I think you're really underestimating the power of math.

Star Trek Discovery Season 2 In a Nutshell - YouTube

Edited by feeWAIVER, 19 January 2022 - 05:42 PM.

[justcallme A S H]

FupDup, on 19 January 2022 - 05:32 PM, said:

Formulas can only do so much. In terms of whether something is OP or not, they can tell you how many people use a weapon, the win rate of players using it, etc.

But math doesn't tell you why you got there or what to do about it. The formula won't tell you if you need to nerf heat, damage, cooldown, range, or whatever else. That's where the qualitative human element comes in.

Spot on!

[Hobbles v]

A formula would give flawed results because there is too many variables. The relative balance is affected by mech size, variety and number of hardpoints, speed, available tonnage and slots.

For example. Ermicro lasers are great for pirhranas but utterly useless for Dire wolves.

Guass rifles are really good weapons, but not on a firestarter.

Variables like this cant accounted for in simple formulas. Its why changes are tested in actual combat with a variety of mechs