74 replies to this topic

[grayson marik]

Dear PGI, fellow players,

during the last month's years we have seen multiple approaches to balance weapons, mechs and equipment to each other and then balance CLAN vs IS.
The newest iteration is currently running on PTS.

Many of you and also some of PGI staff have noted multiple times, that a real balance can never be achieved and that it will be an ongoing process with ever ongoing changes.

Now, let's put this number crunching to an end and finally balance the game with an auto scaling, self controlling system, that gives the player just one value to play with: Mech Battle Value.

Big words you say? Here is how it works:

Every Chassis, Weapon, Equipment, Module, Consumeable gets assigned a base BV value.
So the BV of a Mech is the sum of all those values.

Now here comes the candy:

All those values change dynamicly based on how often and how successful a wepon, chassis or any other equipment is used. PGI already collects all these data so there is no barrier to use it.

If something is used because its OP, the BV for it will rise. If something i.e. Urban Mech, Highlander, Awesome, Centurion, Single Heatsinks etc. is not used often, BV values would fall for this equipment. ( Even the skill tree and player rating system could be part of this system, as a very good player would also have a BV he adds to his Mech and thus would be challenged to use even more bad equipment in order to keep BV limits )

Now, if we go to matchmaking, we simply replace the weight based 3/3/3/3 for groups with a simple BV limit to keep. Or a BV limit for CW Dropdecks.

If you now want to field a very good Mech with a high BV, you will have to use some junk Mechs also in order to keep the limit.

The system would control and scale itself and PGI would be able to control it by the parameters:
Base BV, rate of BV rising and rate of BV falling.


Even the BV limit for matchmaking could be calculated dynamicly. MWO would simply calculate the "Average Joe BV" of all mech currently owned by players and would mutiply this with the number of mechs nedded. 4* Average Joe BV in CW and 12 * Average Joe BV in public queue.


What changes for the player?

• The Mechlab skill changes from " How to stuff as much current Meta Wepons into my Meta Chassis " to " How to build an efficient Mech, with as low BV value as possible"
  
• "Junk" Mechs and equipment would be used not only in beginners Matches but also in more competitive fields like CW
  
• With the more mixed equipment used, complete new tactics and roles would evolve and would bring more depth to the game
  
• beginners would not face the uber Meta Mechs all the time

What changes for PGI?

• The system is self scaling
  
• The system is dynamic
  
• The system provides balance without punishing Meta Configs at all
  
• Role warfare (or whatever other goals PGI has) can be promoted by PGI by changing base BV, rise and fall values for certein equipments

Approach used:







This system does not try to balance everything to each other it does NOT nerv any weapon or META config.It does not nerv CLANS. It does not limit players choice in Mechlab.
It only gives high META equipment a price in relation to not used equipment and finally ends the arms race wich happens after every balance pass. It consequently follows the high demand = high price, low demand = low price paradigm of economy.

And the very best of it all:
We could kick Ghost Heat, DHS with 1.4 and 2.0 values and set all weapons back to TT values the dynamic BV will balance all this stuff automaticly.


Where i got my inspiration from:
A similar system runs the economy on innerspherewars.eu.
For almost 8 month's now we use it to balance the troops you can buy/ build in the league by such a system and you know what?
Since the system started and the numbers regulated themselves from start values after the first 4 weeks, we have had not a single one complaint about equipment being to cheap/expensive used to often etc etc etc.

All the horrors of balancing a league wide economy have gone!
Only difference to this BV system is, that we also have maintenance consts of 5% of the units value. And as unit values rise and fall based on demand, maintenance consts also rise and fall...


testdata for clarification

Spoiler

Mechs 1 to 3 BV development with rising overall mech population and random usage and success changes:
MECH1:

MECH2:

MECH 3


Average BV calculation and droplimits for CW drop decks and 12 v 12 public queue

Now, this figure shows impressively, how the proposed BV system would react to the META or introduction of new equipment. META builds would not be nerved anymore and Mechs would not be gimped anymore by balancing passes month's after the player bought them.
Also, due to the drop limits there would not be such a thing as a "worthless" Mech, weapon System or whatever.

Furthermore, the more wide spread usage of different equipment and weapon systems would require more skill from grouped players to find tactics that work with such a mixed loadout.

Also MM would have a much more easy live, as the PSR rating and the statistics from equipment, weapons and Mechs would be combined to just one value, that would reflect the true combat value of a pilot/Mech combination much closer than currently PSR and weight class.

And this would do 2 things to MM: Firstly it would improve overall match quality and secondly could reduce wait times, as there is only one value left to calculate with, not 2.

Want to play with the numbers yourself?
Here is your place to go:
http://innerspherewa...VStatistics.php

Edited by grayson marik, 27 October 2015 - 10:46 PM.

[Tarogato]

I've always thought that PGI should do exactly this. In fact, I see no reason not to.

[grayson marik]

Tarogato, on 16 September 2015 - 08:29 AM, said:

I've always thought that PGI should do exactly this. In fact, I see no reason not to.

Thx for your backup. I already thought no one would even read it ...

[Kirtanus]

I just dont' get how it will work in usual group or solo matches? Could you provide a few examples?

[BladeXXL]

Kirtanus, on 16 September 2015 - 12:33 PM, said:

I just dont' get how it will work in usual group or solo matches? Could you provide a few examples?

In the current system a non-skilled player with a non-skilled stock atlas, single heat sinks, without modules has the same wight for the matchmaker like a full skilled fine tuned dire full of modules in a advanced skilled hand (around same PSR like a new player)... it ist fair? is it fair when 3 of that atlas have to play against 3 that dires? NO it is not! But current matchmaker will match them because it mess on PSR and wight only.

So about the BV:
Each team is limited to lets say 3000 BV points means 1000 per lance (to visualize with smaller numbers)

You love your skilled 4xCML/4xUAC5-Dire? ... fine, but it has a BV of 400 build on weapons, armor, heat sink count, mech skill level, modules, your skill, etc.
Now the rest of your lance has to share 600 points.
With the left 600 points there is no way you lance can ever get 3 of YOU. :-)
This system works for solo aswell as for groups. Private matches could set own drop limits for the BV.

If you love to play your locust with BV of 80 - no problem to worry about being ball and chain for your team.
If you love to play your 6xMG/2SL-Jager - no problem it has a BV of 150 and won't disable your team completely. :-)
The BV system will compensate your team with more powerfull mechs so being near equal to your opponent.

Edited by BladeXXL, 16 September 2015 - 01:13 PM.

[Tsangdhori]

This is way too exploitable. To give an example, the original incarnation of the MMNET megamek server (before Mekwars) ran a similar system that would weight mechs' BV(this was before 2.0) based on win rate. Theoretically this allowed mechs' bv to arrive at what their 'true' BV actually was. What actually happened was that a number of mechs, especially rare ones, were hilariously undervalued by the BV system. DRGs for 650 BV, FS9-Fs for 900, and even really common stuff would undergo large BV swings as good players flocked away from overvalued mechs while flocking to mechs whose wide availability ensured that bad players would eventually undervalue them.

I see a similar situation arising where those players with large stables and an ear to the ground flock from one mech and weapon system to the next as the BV system adjusts. Newer or less dedicated-to-meta players would be, in essence, digging their own BV graves as they'd be playing just that over-valued equipment that the power-players just abandoned and in doing so eventually undervalue it again.

Basically you'll get stuck in a cycle where the common player is at such a huge disadvantage because they don't keep track of BV fluctuation and the power-player has an even greater advantage, before even dropping into the match.

[McGrizzled]

I was thinking of an example more like this (if this is what you're proposing):

The system takes your mech and puts together it's value based on how it's setup. It then factors in your PSR and ends up with a total Points Rating for pilot skill and mech build. Lets say you have the newer pilot in a newer Atlas and they have a point rating of 200. You have a skilled pilot in a well built Dire Wolf that is valued at 400 points. Sucks to be Atlas.

But what the matchmaker can do, then, is take a mid ranked pilot in a good Timberwolf worth 300 points and a decent pilot in a stock Grasshopper worth 150 points and make a match up of Dire/GHR 550 pts and TBR/AS7 500 pts and probably end up with a decent match. Where as the current system may do a Dire/TBR vs AS7/GHR and it becomes a stomp. Heck, it might even be able to take a great Firestarter pilot worth 100 points and call it an 'even' match at Dire/FS9 500 pts and TBR/AS7 500 pts, which if it works would ease the burden of 3/3/3/3 weight class setups.

Its not that the matchmaker will necessarily need to stick to only certain values, but the main thing is that it will figure in power of the build with power of the player to try and even out matches, where before it was just looking at power of the player and size of the mech. In this new system, someone leveling a new mech or playing a more funsy build should be natually devalued by the calculated mech value verses a build running meta stuff.

[grayson marik]

Tsangdhori, on 16 September 2015 - 02:05 PM, said:

This is way too exploitable. To give an example, the original incarnation of the MMNET megamek server (before Mekwars) ran a similar system that would weight mechs' BV(this was before 2.0) based on win rate. Theoretically this allowed mechs' bv to arrive at what their 'true' BV actually was. What actually happened was that a number of mechs, especially rare ones, were hilariously undervalued by the BV system. DRGs for 650 BV, FS9-Fs for 900, and even really common stuff would undergo large BV swings as good players flocked away from overvalued mechs while flocking to mechs whose wide availability ensured that bad players would eventually undervalue them.

I see a similar situation arising where those players with large stables and an ear to the ground flock from one mech and weapon system to the next as the BV system adjusts. Newer or less dedicated-to-meta players would be, in essence, digging their own BV graves as they'd be playing just that over-valued equipment that the power-players just abandoned and in doing so eventually undervalue it again.

Basically you'll get stuck in a cycle where the common player is at such a huge disadvantage because they don't keep track of BV fluctuation and the power-player has an even greater advantage, before even dropping into the match.

Ahh well well, this is going to happen for the first 3 or 4 weeks, when all Mechs are starting from their base values. After this time, the gaps would close.

As mentioned above, this system is already running on a planetary league and works perfectly there.

And this is not MMNET: there is no such thing like a rare Mech in here. If there is a rare one, because it was utterly crap in the current system, it will rise in numbers because players start using it more for its low BV value.

And then it will rise in BV and find its true value after a while. And yes, you are right, for a short period of time, seldom used Mechs will be underrated and there will be a shift from one Mech to another from time to time, especially when new Mechs arrive.


But this can be easily countered, if new Mechs get a decent base BV value at start, which puts them into the average area, not in the low Bv area.

In the end, the described shif will by far not be so hard as the current 3 to 5 viable Mechs META is.


And just in case, there is really a rare Mech, lets say a hero Mech, of couse the usage numbers must be factored to the numbers this particular Mech exists in MWO.

If there are only 10 Mechs of a type but those 10 are used in every match by their owners, they need to rise appropriate. But this is one of the parameters, I already described in my initial post:

BV rise rate, BV fall rate : could be set for every Mech/ Equipment and would make sure that rare stuff will not be unterrated, as in your MMNET example.

Edited by grayson marik, 16 September 2015 - 11:01 PM.

[Kirtanus]

Current PSR is looking at your match score and not your mech. That made my matches in solo completely better balanced, than before!
Your idea is based on mechs only, and i imagine 12-0 matches every second drop, because good players in pug matches are not always plays "meta", but there are plenty of good configs which are played rare but in right hands - this will lead to stomp...

[slide]

This game so needs something like this that it is not funny anymore.

I dropped in a game last night with players on the other side that are so far above my skill level that it wasn't even remotely funny.

The current matchmaker is so borked by low player numbers in my TZ that it may as well not even be turned on. None of the systems PGI has implemented has worked and won't work unless there is a minimum number (no idea what) of players which is a lot higher than what we have.

A radical new approach to game balance is required and it needs to happen soon. Hand adjusting weapons values and quirks across the 200+ mechs we have now will only become harder and harder as time goes on and the number of chassis grows and eventually new weapons are added as well.

PGI has millions of games worth of data, if they mined it properly I would bet they could give every chassis a pretty accurate starting BV (based on current use and K/D ratio) and then refine it from there with your system. They could keep the quirks to give flavor to mechs or drop them all together.

Powercreep might even become a thing of the past as people will use mechs they like because they can compete. TTK might even go up a little as the overall BV of matches may drop a little over time.

+1000 from me

[grayson marik]

Kirtanus, on 16 September 2015 - 11:09 PM, said:

Current PSR is looking at your match score and not your mech. That made my matches in solo completely better balanced, than before!
Your idea is based on mechs only, and i imagine 12-0 matches every second drop, because good players in pug matches are not always plays "meta", but there are plenty of good configs which are played rare but in right hands - this will lead to stomp...

No, i have also stated, that pilot ranking could also contribute to BV values, as well as the skill tree for Mechs.
This would connect PSR AND Mech value to grant even better matches than now with only PSR .

And to find effective Mech builds with low BV would be a new skill to develop in Mechlab, a good one for the game as I see it.

Edited by grayson marik, 16 September 2015 - 11:58 PM.

[The Witt]

The thing I don't like about a BV system is that you still have a problem of some mechs are way better than others, like way way better.

I've been thinking on a similar thought path, but applying mech popularity to the magnitude of quirks. So a mech being used less often gets +1% to it's quirk to encourage more popularity and a mech that is very popular gets -1% to its quirk. This would eventually balance out the whole mech ecosystem.
Of course, this is something that should have been done for the last 12 months....

[grayson marik]

The Witt, on 17 September 2015 - 12:42 AM, said:

The thing I don't like about a BV system is that you still have a problem of some mechs are way better than others, like way way better.

I've been thinking on a similar thought path, but applying mech popularity to the magnitude of quirks. So a mech being used less often gets +1% to it's quirk to encourage more popularity and a mech that is very popular gets -1% to its quirk. This would eventually balance out the whole mech ecosystem.
Of course, this is something that should have been done for the last 12 months....

This is just the system with an other name... and it has a backdraw:
Mechs are changed, after people bought them, just as now. I do not like my Mechs being changed every 4 weeks with jet another quirk pass.
I do not like the endless struggle of CLAN vs IS balance, where numbers are crunched and weapon, armor, structure values changed, just to see that nothing changed in the end.

The suggested BV system would make it possible, that Clan Mechs can be left untouched with superior tech. They will simply acquire a higher BV as they are indeed more valuable on the battlefield. And thus with the same BV limit on both sides, we could reach balance without the endless " jet another balance pass nightmare" we have now. If even PSR would contribute to the system, even better PUG matches would be possible.

[SlugShot]

sounds good

[grayson marik]

The Point is this: There are always Mechs that have more value on the battlefield than others. But if you don't have any advantage by using the Mechs that dont have the same high Battle Value, you simply don't use them. This is what happens for years now and which will stay this way, without the BV system I suggested.

Only change will be, this month it is a Thunderbold, after next Patch it might be a Centurion that is the Mech with the highest usage because of its value on the battlefield.

Now, with the BV system, you simply reflect this value to the actual Mechs and achieve balance by managing the added battle value that is on the field, instead of managing the weight, that is on the field.

The BV is much more exact to the true value of a Mech and with the dynamic calculation, new equipment, Mechs and skill tree items can flow in easily without leading to the next balancing marathon nightmare.

[Karl Streiger]

I have tried to Powerpoint this idea.... maybe its a different interpretation of Graysons idea -

currently we should have 5 different player skill levels - PSR - if they work properly
it should have 5 pools of different sizes:

excuse my bad power point style -

i don't know if it is exactly what Grayson did suggest - but i would also use the "known" Tier rating for Mechs and combine this with the PSR rating: intstead of 5 bigger pools you get fluctuating 25 minor pools.

why does the aquariums have different sizes?

My reason behind: is that based on player numbers per pool you create MatchMaking groups:

The player have to know his PSR as well as the Tier of his Mech Type. Combined with different player behaviours:
I play to win - play because of challenge etc.
the choice of the Mech may place him in different Pools.

a Tier 2 player - will run a Tier 1 Mech - to compete with the TopDogs in the First Group - or he will run a Tier 3 Mech to be the TopDog in the second group - it is unlikely that he will run a Tier 2 Mech - because he has a big disadvantage by PSR and Mech Tier
So different player skills will run different Mechs - that means PGI should get really good Server Data - how a Mech performes.

[grayson marik]

Alright then, here it comes. I have made up a testing line with 3 different mech types and rising mech population as BV making variable I used the number of times the mech was used in battle.
Of course, PGI could and should use more variables, as I described in my initial post.
What I did to calculate it, so that a rare mech cannot be used to ride the BV system is the following:
I calculate the average mech number per mech type. Then I calculate the gap between the current mech model and that average number. If the gap is positive - means the mech type is only available in small numbers, this gap will be used to calculate the usage number up to how many times would the mech have been used, if it were of average mech type population.
If the gap was negative, I also toned the usage numbers down as if the mech type was of average mech population.
The result is of course that the BV level could not spike out for example gold mechs used obsessively by their owners.
ok, enough text, here are the numbers: http://innerspherewa.../isw2/mechs.htm
And here you can produce test data on your own: http://innerspherewa...VStatistics.php
Enjoy.

[grayson marik]

Ok, following my statisitcal test example from yesterday, here the BV limits for public 12 man and CW queues (the glowing lines ). As you can see, MECH 2 is our most META mech in the example and could be used maybe 2 or 3 times in a 12 Man and once in a CW Dropdeck, while MECH 3 ( this would be our little URBANMECH :-D ) can be used more than often.

Now, this figure shows impressively, how the proposed BV system would react to the META or introduction of new equipment. META builds would not be nerved anymore and Mechs would not be gimped anymore by balancing passes month's after the player bought them.
Also, due to the drop limits there would not be such a thing as a "worthless" Mech, weapon System or whatever.

Furthermore, the more wide spread usage of different equipment and weapon systems would require more skill from grouped players to find tactics that work with such a mixed loadout.

Also MM would have a much more easy live, as the PSR rating and the statistics from equipment, weapons and Mechs would be combined to just one value, that would reflect the true combat value of a pilot/Mech combination much closer than currently PSR and weight class.

And this would do 2 things to MM: Firstly it would improve overall match quality and secondly could reduce wait times, as there is only one value left to calculate with, not 2.


So well, forgive me my arrogance PGI, but here is the salvation for your stressed MM and poor usage of many many Mechs as well as too your never ending introduction of arbitrary "balance" systems. take it or leave it PGI but don't cry if I will say " I told you so" after your next try on balancing Mechs and CLan vs IS.

[grayson marik]

Edit, just sent this email to PGI:

Dear PGI staff,

please take the time to go over my thesis here: http://mwomercs.com/...ost__p__4695529

I am sure, such a system would solve the current balancing nightmare you face every day and would also be future save, when new mechs and equipment arrive on the battlefields.
The best the system would do for you would be to connect your good player skill rating with an transparent and very exact rating for the Mech used. The sum of these 2 ratings, would be the only variable needed for CW dropdeck and PUG queue team calculations.

A very similar system is currently running my planetary league at innerspherewars.eu successful for about 8 month's now and has proven robust and fair.


If you have any questions, I would be happy to answer.

For the best Mechwarrior ever created!

Gerald aka Grayson Marik

[Wingm3n]

Great idea, good work!