33 replies to this topic

[Mercules]

BV will never be perfect... but what it IS good for is making sure one side isn't running 5 stock locusts again 5 Timberwolves. There might not ever be perfect parity, but it does help you get a bit closer. There are problems with the system as pointed out. PGI will have to do their own as what is good in TT is not what is good here, but at least something to account for the fact that an untouched Griffin is probably not as potent as say a customized Stormcrow.

[Xetelian]

Just do something like this

TBR 100
SCR 100
DWF 100

Everything else <100

Give quirks based on weight.

I have officially balanced the game.

[EGG]

It's 2015 not 1985, you don't need to invent numbers for anything to get a BV.You just datamine the massive database of mech battles PGI have. Thousands of battles are being added to this everyday. Mech chassis that win more battles = higher BV.

You then have the option of determining a BV per variant or per build. Per variant is much easier and means you don't have to care about 9-flamer Dire Wolves.

[Joseph Mallan]

Burktross, on 13 July 2015 - 04:15 PM, said:

When in tabletop, units-- more specifically components-- have battlevalues denoting their power. For instance, a locust has a BV of 432, but its 20t clan brother, the Phirana has a BV of 998. This is due to superior clantech equipment. Equipment upgrades in general, such as upgrades like endo/DHS raise BV. Bigger mechs also have bigger BV. So in essence, BV makes it such that you have many small, low power units, or few high power units.

Rule of thumb is that clanner mechs have higher BV than their equal weighted counterparts-- thus the IS was able to field a number advantage to counter the clan power advantage.

We don't have that here, so clantech and IStech must be balanced

And a single Piranha would completely wreck 2 Locusts.

[Impyrium]

The weird thing is that, while this BV thing of PGI's sounded like an interesting possibility at first, now it just seems like a mathematical way to calculate a 'mech's 'effectiveness' rather than the hand-picked tiers from before.

In other words, if it has a high BV, it's going to get negative quirked into the ground. Or the lower ones will get positive quirks.

Nothing really new... which saddens me a little, because I no longer see how IIC mechs are going to be balanced.

[Raggedyman]

MischiefSC, on 13 July 2015 - 04:15 PM, said:

Battle Value is a largely arbitrary point system where point values are assigned to components and mechs. 1,000 pts of one this is worth 1,000 pts of another. So a 1,000 pt mech that's 50 tons but endo and good weapons is equal to a 75 ton mech with crappy gear.

In reality it doesn't really work. There is no actual way to actually objectively assign values to things in this sort of environment. Synergy, positioning, the fact that an XL may be brilliant in one build and suicide in another, that on a Spider FF is worthwhile while on an Atlas it's stupid, etc. etc.

I will agree that a Battle Value/points system can never be perfect, however to describe them all as inherently arbitrary is somewhat disingenuous, as with work an analysis you can get them to the "pretty good" levels of usefulness, including allowing for synergy issues.

A lot of it comes down to setting a goal of accuracy, accepting outliers as inevitable, and going for the big numbers (say "90% of matches with a 95% points balance will have a 55/45 victory chance split, over a 3 month period") rather than focusing on singular situations with a lot of none-points related variables (say "I packed a 30ton mech with a million lasers, drank a bottle of bleach, and then lost against the DireWolf. System be broken!!").

It also involves having the points system fit around the game by going over the data of how things actually perform, rather than trying to shoehorn the game into the points system by going 'nope, drop that items range as it's still not an X point bit of kit'.

Edited by Raggedyman, 14 July 2015 - 04:10 AM.

[Ovion]

Flat points, yes will have flaws.
That's why you use equations.

Each thing gets pretty damn complex when, with a small sample:

Spoiler

In addition, all items may possibly want a small modifier for cost, but not entirely relevant in MWO.
Weapons: Damage, Heat, Cooldown, Range (+ Max Range), Slots, Tons, time to hit target, hits per shot, duration of shot, Ammo(if any), DPS, impulse, health, potential crit damage and any special modifiers (adjusted crit rates, need to get lock / hold lock, charge time, exponential heat, etc)
Ammo: Shots per ton, damage per ton, full or half ton, health, explosion damage, how many weapons it can service.
Modules: Various values and modifiers based on the effects of each module. These would be among the trickiest, with some major abstracts like Capture Accelerator, Radar Dep, etc in there.
Jump Jets: Weight, size, thrust/tonnage.
Heat Sinks: SHS/DHS, internal/external, slots, weight, dissipipation

Engines would have to have flat values based on weight, slots, STD/XL, internal Heat Sinks, then equations for mech weight, speed (for/rev), twist speed, torso pitch speed, torso yaw speed, arm yaw speed, arm pitch speed.

and so on.
Some things are dynamic too based on the mech and what else is on the mech.

Mechs would be a massive equation base.
Weight, class, ES, engine range, total hardpoints, ECM capable, JJ capable, MASC capable, internal structure, XL friendly are the base.
None/Basic/Elite status, Artemis and Quirks would apply flat bonuses (structure, armour, etc) and floating equations based on modifiers (twist, range, damage, heat dissipation, Artemis bonuses, etc) based on what they do with what, and if it's on the mech.

I feel that after mech, engine, base heatsinks, and base upgrades, potentially a value applied for total crit slots and weight available would be beneficial.

Hardpoint specifics would all need doing individually, based on what it is, where it is, what can fit there, and if it's used or not (as it isn't just a case of if a hardpoint is good or not, if it isn't used it's irrelevant in this part, and covered in the flat hardpoint numbers in the base mech equation)

Armour is seperate and frankly the simplest part, with a value per point plus a modifier for FF.

Then, for game terms you'd have a modifier based on ELO to account for pilot skill - probably *0.5 to 1.5 or something.
Or preferably, a Total Pilot BV, based on total games, game time experience points earned, c-bills earned, total gamescore, overall accuracy AND Mech specific BV (by chasis, not variant), based on total games in that mech, game time in that mech, experience points earned in that mech, c-bills earned in that mech, and total gamescore in that mech.
Those two combined (averaged or something?), would either give a flat value to add to the BV, or a percentage modifier to add.

This shows how you can make a comprehensive and practical BV that actually does what we want.
But is also VERY complicated.
It wont be perfect, but it'd be as close as you could possibly get.

Edited by Ovion, 14 July 2015 - 05:32 AM.

[Roadkill]

The easiest way to use all that data that PGI has would be to implement Elo for every component.

Seriously.

Elo works fine in multiplayer games, but it needs a lot more data than it does in chess. Imagine the possibilities if every single game by every single player could be used to rate the Medium Laser. Or the PPC. Or every weapon and every Mech and every hardpoint location all at once.

Sure, a medium laser on a Timber Wolf will seem "more powerful" than a medium laser on an Adder. The effect on the medium laser will be a wash between those two Mechs, but the Timber Wolf's rating would go up and the Adder's would go down. Eventually things would settle down and you'd have objective and accurate ratings for everything in the game.

But yeah, never gonna happen. Too much work to implement it, and then they'd have to let it run in the background without using it for a while to let the ratings settle before doing anything with them. PGI do work on something they can't make money off of for months? Yeah... right...

[MischiefSC]

Raggedyman, on 14 July 2015 - 04:07 AM, said:

I will agree that a Battle Value/points system can never be perfect, however to describe them all as inherently arbitrary is somewhat disingenuous, as with work an analysis you can get them to the &quot;pretty good&quot; levels of usefulness, including allowing for synergy issues.

A lot of it comes down to setting a goal of accuracy, accepting outliers as inevitable, and going for the big numbers (say &quot;90% of matches with a 95% points balance will have a 55/45 victory chance split, over a 3 month period&quot;) rather than focusing on singular situations with a lot of none-points related variables (say &quot;I packed a 30ton mech with a million lasers, drank a bottle of bleach, and then lost against the DireWolf. System be broken!!&quot;).

nope

It also involves having the points system fit around the game by going over the data of how things actually perform, rather than trying to shoehorn the game into the points system by going 'nope, drop that items range as it's still not an X point bit of kit'.

Nope, largely arbitrary. The values are not consistent and heavily inter reliant. Beyond which their value is relevant only in relation to the skill of the player using them. Some people rock lights some don't. Beyond which things like hsr updates can dramatically affect the viability of mechs and weapons.

An xl in a spider is great. Mgs in some spiders are good. Xl in a atlas and Mgs in an atlas are terrible.

You are better off using something akin to Elo but drilled down to components and chassis as a modifier to a players overall Elo. That could be very useful. A flat bv system though is broken out of the gate, even more so in mwo than tt.

Edited by MischiefSC, 14 July 2015 - 07:37 AM.

[Summon3r]

dont hold your breath on the rebalance if its as great as the highly touted hit reg fix is then add it to the list of huge let downs.

[Jack Corban]

DjPush, on 13 July 2015 - 04:11 PM, said:

What is battle value and how is it used to balance a game?

Read the Background infos or gtfo.

[RedDevil]

BV (Battle Value), is ELO for mechs and their loadouts.
Better mechs have a higher BV. Those same mechs with optimal loadouts have even higher BVs.

It's superior to balance by tonnage, and PGI has said they are going to use their own customized version, and not the Table Top version.

They have much more data to backup their numbers.

Edited by reddevil, 14 July 2015 - 09:02 AM.

[Wronka]

My understanding of this new "BV" like system re-balance was not to be used for matchmaking at all. It is going to be used to figure out which mechs need the most help based on hardpoints, weapon locations, etc. Though I am not sure how that data is not already obvious. I can take about 30 minutes and tell you exactly which mechs are not good and which ones are good. There are already tier lists out there that do this that are very accurate.

[RedDevil]

Wronka, it will probably be a bit of both. It won't be used in matchmaking, but could be implemented in CW. Instead of tonnage you could have a max BV value that you can mix and match mechs to reach.