Author Topic: A possible systematic approach to balancing the fleet ship types  (Read 7730 times)

Offline Wingflier

  • Core Member Mark II
  • *****
  • Posts: 2,753
  • To add me on Steam, click the little Steam icon ^
Re: A possible systematic approach to balancing the fleet ship types
« Reply #15 on: December 06, 2010, 02:35:22 pm »
Without critisisng that proposal, I am an absolute disbeleiver in deriving balance using mathmatical calculation:

  • it is impossible to realisticallly factor in everything
  • it is very difficult to even get a formula that represents the reality of balance
  • it results in less contrast between unit types
  • it makes for dull balance
  • it gives hugely inferior results to the process of incrementally balancing and nerfing based on player feedback

That oppinion is a product of 5 years involvement in the balance of total annihlation derivative mods via TA spring, and it is impossible to list how many balance models for multiplayer RTS balance have been tried in that enviroment (so many).

The most successful method in spring by a huge margin just followed a 'make units each really cool, then refine them between releases until everything is useful'

in RL, the balance of fleetships in last 3.0 version was pretty good, and I have played that last release to death. OK, zenith polarisers were dominating- but no unit besides acid sprayers & neinzul ships, raptors & microfighters were actually non-viable choices of starting unit.

imo, the best way to balance an RTS is just make every unit cool to use from a players perspective, then make incremental changes accordingly. I think acting upon the feedbac  in the ARS unlock voting thead, and making changes based upon your own intuition as the developers and people who know the game best would give infitiely better results than some equation you've made in excell.

So yeah: the objective imo is just balance based upon feedback and intuition in a way that every unit is cool, fun and effective to use in the hands of the player, which can be acheived by changes of shipcap and cost as well tweaking as unit stats and roles. every unit should be considered by the criteria: 'is this a cool idea? do I want to get out there and use this ship?'

the zenith charmelon for example, isnt cool. its fragile, weak attacker with an ability that is unsuited to to both its role- aggression- and the purpose of building ships instead of turrets as a player- aggression. imo, it should be rebalanced as a defensive ambusher unit that functions like a toad- slow moving, then stops to camoflage itself, and when somthing comes close, BAM
cool idea balance method gives much more interesting unit rosters than mathmatical balancing.

just my 2cents, I support any direction AI war goes, I just have alot of misgivings about balancing methods that involve excell.. I really think that by now you guys & this community knows what balance changes to be made without scrapping it in favour of impartial maths :(

edit: I have an established history of being somthing of an opposition party on the AI war forum, and I just want to say, I realyl hope I dont come accross as too much of a jerk when I make these kind of posts... thats not my intention.
Honestly I think this kind of attitude is just as extreme, radical, and detrimental as those people who say that you can't possibly balance a game using only what I like to call "guess and check".

I have actually been there through a lot of the Total Annihilation mod balancing process (I was actually around during UberHack, the first mod that ever attempted to balance the game), and I have seen both sides of the argument, and the effects of each method.

Player feedback is simply too biased to be the only factor in balance equations.  In Spring, this is doubly so because most games are fought 8v8 style, which is balanced nightmarishly different than 1v1 games would be, and in my opinion contributes greatly to the balance problems even BA has had over the many years.

Is it impossible to balance a game completely on mathematics?  Maybe.  But you can't argue that it doesn't help, because I know for a fact that Blizzard is a HUGE proponent of using math to balance their games, they don't even keep it a secret.  It comes as no surprise then, that the Starcraft series are, without a doubt, the most balanced RTS games on earth; try and prove me wrong.

So what I'm saying is that you're both right.  Both Keith's and Chris' method of balancing are correct, when correctly analyzed, and used in tandem.
"Inner peace is the void of expectation. It is the absence of our shared desperation to feel a certain way."

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: A possible systematic approach to balancing the fleet ship types
« Reply #16 on: December 06, 2010, 02:36:44 pm »
I firmly believe there is a formula that can give a broad overview about unit strength but it might very well be extremely time consuming to find it.

I think the main thing right now is just for finding outliers.  My methods use actual simulation data (fight a group of unit X against a group of unit Y and see what happens), and combining this plus my method plus player feedback should give us a pretty reasonable framework.
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Wingflier

  • Core Member Mark II
  • *****
  • Posts: 2,753
  • To add me on Steam, click the little Steam icon ^
Re: A possible systematic approach to balancing the fleet ship types
« Reply #17 on: December 06, 2010, 02:51:30 pm »
Okay now looking at the spreadsheet, here's what I would do:

The individual hull types (for example light, heavy, neutron, etc.) are going to be extremely hard to find a "number" for.  In general, some hull types are used by a lot of ships, and others are used by very few.  However, the number of ships that use each individual hull type does not necessarily translate into how much having a multiplier against that type is "worth", because some hull types (say turret or shield) are extremely important to destroy for obvious reasons.

In my opinion, the best thing you can do is rebalance the ships (wasn't Chris going to do this anyway) to where they are relatively "good" against everything, and especially good against a few hull types.  If you can make each ship to where it is only good against 3-4 hull types, then you don't even have to include DPS vs. Hull Types into the equation, since every ship has some bonus damage against something.  What you CAN give a balance "price" to are things like "armor penetration", range, and "area of effect".  Those are very quantifiable elements that can be more easily given a mathematical quantity of usefulness.

Another thing I would stop doing is balancing "per cap".  Ships should be balanced by their individual cost (In Metal, Crystal, and Energy), as compared to what the ship actually gives you.  In actuality, Metal, Crystal, and Energy Costs should offset the result of the final "cost" of the other ship attributes.

For the sake of balance, we can say that every 2667 energy a ship costs to sustain is worth 1 metal and 1 crystal.  The reason we do this is because a MKII Energy Generator (the most efficient type) creates 40,000 energy for -15 metal and crystal a piece.

So if each point of metal or crystal is worth 1000 "points" in the equation, then each 2667 energy a ship costs to maintain is also worth that.

Cap balance is tricky.  Ships that have a higher cap are definitely worth more, but how much more?  I think the easiest way to balance ship caps is to simply add a multiplier at the end of the equation that makes that particular ship type worth more because you can have more of them.
« Last Edit: December 06, 2010, 03:24:41 pm by Wingflier »
"Inner peace is the void of expectation. It is the absence of our shared desperation to feel a certain way."

Offline Suzera

  • Full Member Mark III
  • ***
  • Posts: 220
Re: A possible systematic approach to balancing the fleet ship types
« Reply #18 on: December 06, 2010, 03:04:57 pm »
I guess they're simple enough, but I think you should use equations instead of case switch type things where possible. Granularity is great. Many of these need to be non-linear though, which you did to some extent.. Armor gets more powerful per point as it gets higher, so it should be less than linear (armor point spent ^ .9 * constant as an example). Speed is offset to 22, since that's what missile frigates are, so the equation is a bit more complex, but still fairly easy to get an exponential form. It should count as less beneficial as it goes up, and more penalizing as it goes down in linear increments, so maybe a doubleing or halving of speed is worth one ship point (I'm doing these forward calculating though, from points to stats as examples instead of the reverse like you are right now, but some simple algebra would net the inverse):

(2^(speed points spent) * 22)

at 0 (for 22 speed) this is

2^(0) * 22 = 1 * 22 = 22 speed

For one point it will get

2^1 * 22 = 2 * 22 = 44 speed

So you'll get a chart like:
Code: [Select]
Points  Speed
-3      2.75
-2      5.5
-1      11
0       22
1       44
2       88
3       176

Range should probably be done similar to speed, unless you want to encourage things to be micro heavy or people to not use fast or slow ships. If you want it micro-heavy, count them linearly because with extreme micro even a few points bonus to range on a high speed ship is huge.

Speed + range can have a funky relationship though, so the faster things should probably tend to not have the longest ranges. The only thing I can think of violating this rule to a troublesome degree is raptors, but those only baaarely edge out frigates in range, the AI doesn't micro them, and AI guardians have more range than Raptors.


Armor points would NORMALLY multiply points in health, but since you're aiming for linear armor being hard crushed by linear high damage from certain counter ships, it is less of an issue, as long as you make sure the counter ships are numerous, focus on their targets only they can take out, and aren't taken out easily by the armored target in question. Right now fighters achieve this against equal mk bombers as an example. Mk1 fighters do not achieve this against mk 3+ bombers, but that is a minorish issue in my opinion. Raid ships currently violate this rule. They are extremely tough in the early game due to their armor, and nothing the humans start with my default that is mobile can go way over raid starship armor. Frigates damage just doesn't cut it against them.

Rate of fire would normally multiply damage and that's the end if this were a % damage reduction game, but since it is a linear damage reduction game, singular big shots tend to be favored to overcome armor. This one is tricky with linear armor because REALLY BIG shots are wasteful because of overdamage, but REALLY SMALL shots are completely negated by armor. See: cruisers blowing up infiltrators (huge damage waste each) and cruisers trying to blow up raid starships (damage to small overcome armor). It may have to be step functioned per ship based on whatever you want to consider "high" armor for that ship's targets to be to determine the point at which raw damage below that line is considered worth less penalty per additional point below, and what you want to consider "high" health or "high" damage for whatever the favored target type of the ship will be, above which point the damage would be worth less per point above, and linear between.

Each default ship hull type bonus damage amount should probably add a value of (number of bonus ships in the game/(constant between 1 and 2)) to the damage multiplier point value category, each ARS hull type should add 1 to it's relevant hull type category. Command grade should not appear as a damage multiplier type (or at least not have value because humans generally get no benefit from it really and thus assigning it value and balancing on that makes it less worthwhile to humans), turrets shouldn't appear as a damage multiplier type (or again at least not have value because it is AI only and has 0 value to the players).

These are all mostly example numbers since I am just donating text in my spare time. The point values assigned to generate values would in reality be much higher integers so they could be balanced across each other without huge floating points all over the place. I'm sure armor would have a much weaker than I specified function too.
« Last Edit: December 06, 2010, 03:12:22 pm by Suzera »

Offline superking

  • Hero Member Mark III
  • *****
  • Posts: 1,205
Re: A possible systematic approach to balancing the fleet ship types
« Reply #19 on: December 06, 2010, 03:11:38 pm »
Player feedback is simply too biased to be the only factor in balance equations.  In Spring, this is doubly so because most games are fought 8v8 style, which is balanced nightmarishly different than 1v1 games would be, and in my opinion contributes greatly to the balance problems even BA has had over the many years.

in rl though, BA is balanced by and for 1v1 by a small group of good players, and was balanced via a long history incremental 5%/10% buffs and nerfs... uberhack was balanced by cadyr, who was relatively clueless and just did complete lolz based on forum suggestions by noobs ;D

Starcraft 2 was indeed balanced mathmatically, but they had an effectively unlimited budget over a 10 year development period, not to mention the vast amount of theory generated by SC1... and most of said maths was just copying starcraft 1 values and then working from them. SC1 was not balanced mathmatically as the difference in balance between the 1.0 release version and latest current patch is pretty much absolute.

I'm getting a little involved in achron, the time travel rts (http://www.achrongame.com) but they suffer from a lead dev who beleives in balancing completely from a maths formula- when an imbalance is found he changes the formula, not the unit- and the result is both really bland and still fairly imbalanced after many years (where I'm pretty sure a few dedicated and skilled testers with the ol' incremental changes would've worked in a matter of months).

there was a spring mod, BA derivative, balanced from a maths formula that lasted 6 months- 'tired annihlation' that actually wasnt bad.. it just wasnt much fun.

anyway, as you say the best is to use both

Offline Suzera

  • Full Member Mark III
  • ***
  • Posts: 220
Re: A possible systematic approach to balancing the fleet ship types
« Reply #20 on: December 06, 2010, 03:13:18 pm »
Also, power becomes more and more costly as it goes up, not less. 10000 to 20000 power is not as bad as 20000 to 30000. You only get so many efficient mk 1 power plants.

Edit: Oh wait, right. You're doing this the other direction. Nevermind this post. I'm used to thinking things from scratch to implementation, not implementation to point amounts. You should continue it into geometrically larger negatives though in my opinion.
« Last Edit: December 06, 2010, 03:23:50 pm by Suzera »

Offline Suzera

  • Full Member Mark III
  • ***
  • Posts: 220
Re: A possible systematic approach to balancing the fleet ship types
« Reply #21 on: December 06, 2010, 03:19:21 pm »
You do look like you are probably way overvaluing attack range and movement speed though, due to treating it as linear.

Offline keith.lamothe

  • Arcen Games Staff
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 19,505
Re: A possible systematic approach to balancing the fleet ship types
« Reply #22 on: December 06, 2010, 03:26:00 pm »
You do look like you are probably way overvaluing attack range and movement speed though, due to treating it as linear.
Right, like I said in the OP I'm way off.  And I should probably be using fancier relationships (exponential/logarithmic) etc.  The suggestions you're giving me are exactly the sort of thing I wanted in terms of feedback :)  That was assuming the framework as a whole seems at least worth pursuing but it seems that it is, though the words of caution that this can realistically only go so far are also helpful).

At least, I think it's safe to say this won't make balance demonstrably _worse_ ;)
Have ideas or bug reports for one of our games? Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Lancefighter

  • Core Member
  • *****
  • Posts: 2,440
Re: A possible systematic approach to balancing the fleet ship types
« Reply #23 on: December 06, 2010, 03:27:19 pm »
Yeah, but BA was boring anyway. NOTA is where the fun is at ;)
Ideas? Suggestions? Concerns? Bugs to be squashed? Report them on the Mantis Bugtracker!

Author of the Dyson Project and the Spire Gambit

Offline Suzera

  • Full Member Mark III
  • ***
  • Posts: 220
Re: A possible systematic approach to balancing the fleet ship types
« Reply #24 on: December 06, 2010, 03:31:03 pm »

I'm getting a little involved in achron, the time travel rts (http://www.achrongame.com) but they suffer from a lead dev who beleives in balancing completely from a maths formula- when an imbalance is found he changes the formula, not the unit- and the result is both really bland and still fairly imbalanced after many years (where I'm pretty sure a few dedicated and skilled testers with the ol' incremental changes would've worked in a matter of months).

If he is changing the math model, that should change the implementation, if his model -> implementation was any good. If he is changing the model, but not implementation, that means he is NOT actually implementing the model. Unless you're ACTUALLY going to fullsim the thing in math (which is sort of what Chris is doing, but I mean in a much more encompassing fashion to include many types of combined arms) so that everything always matches the implementation, the math model isn't flawless. AI War kind of game would be really really hard to mathematically model perfectly without some radical, hard to program cpu-intensive changes to the ship maneuver mechanics to make them perfectly optimal per ship every time. Fleetblobbing is very very far from optimal, as is probably anything anyone else is doing. You would need to spend hours per game minute of combat right now to get close. It would look really really really weird and mindblowing to watch too, I assure you. The fudge factor with that is where the playtesting comes in, to change the fudge numbers in the model, and thus the implementation.
« Last Edit: December 06, 2010, 03:33:47 pm by Suzera »

Offline Wingflier

  • Core Member Mark II
  • *****
  • Posts: 2,753
  • To add me on Steam, click the little Steam icon ^
Re: A possible systematic approach to balancing the fleet ship types
« Reply #25 on: December 06, 2010, 03:37:04 pm »
Player feedback is simply too biased to be the only factor in balance equations.  In Spring, this is doubly so because most games are fought 8v8 style, which is balanced nightmarishly different than 1v1 games would be, and in my opinion contributes greatly to the balance problems even BA has had over the many years.

in rl though, BA is balanced by and for 1v1 by a small group of good players, and was balanced via a long history incremental 5%/10% buffs and nerfs... uberhack was balanced by cadyr, who was relatively clueless and just did complete lolz based on forum suggestions by noobs ;D

Starcraft 2 was indeed balanced mathmatically, but they had an effectively unlimited budget over a 10 year development period, not to mention the vast amount of theory generated by SC1... and most of said maths was just copying starcraft 1 values and then working from them. SC1 was not balanced mathmatically as the difference in balance between the 1.0 release version and latest current patch is pretty much absolute.

I'm getting a little involved in achron, the time travel rts (http://www.achrongame.com) but they suffer from a lead dev who beleives in balancing completely from a maths formula- when an imbalance is found he changes the formula, not the unit- and the result is both really bland and still fairly imbalanced after many years (where I'm pretty sure a few dedicated and skilled testers with the ol' incremental changes would've worked in a matter of months).

there was a spring mod, BA derivative, balanced from a maths formula that lasted 6 months- 'tired annihlation' that actually wasnt bad.. it just wasnt much fun.

anyway, as you say the best is to use both
Uberhack was not balanced by Caydr, it was balanced by BraveSirRobin.  Caydr made AA which came after that, and I will agree was a joke.

And granted, Starcraft 2 had plenty of playtesting, but originally it was all balanced mathematically (as with the 2nd one as well).  The changes they make are all mathematic changes as well, they don't just simply "guess" how much to change a number; they use math combined with high-level playtesting to determine what should be done.  As far as Tired Annihilation goes, the reason it didn't become popular had little or nothing to do with how "balanced" or "fun" it was.  The fact of the matter is, Balanced Annihilation will always be the most popular Spring mod, because the playerbase is only big enough to support 1 mod fully, while having small niche playerbases for the other mods.  Absolute Annihilation, with all its horrible balance, ruled the scene until Cadyr quit back in '06 I think?  And I have no reason to think it wouldn't still be the most popular if he was still around today.  It has nothing to do with how 'fun' or 'balanced' any of the mods are.

But enough on that, let's get on topic.
"Inner peace is the void of expectation. It is the absence of our shared desperation to feel a certain way."

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: A possible systematic approach to balancing the fleet ship types
« Reply #26 on: December 06, 2010, 03:42:51 pm »
If it makes anyone feel any better, professional game developers (AAA or indie) aren't any more agreed on this issue than the players are. :)

But I agree, back to the topic at hand... which I pretty much leave to Keith, as I have my own foci at the moment.
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline keith.lamothe

  • Arcen Games Staff
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 19,505
Re: A possible systematic approach to balancing the fleet ship types
« Reply #27 on: December 06, 2010, 03:51:30 pm »
I do find it amusing that this approach was partly prompted because I could ask players what they thought of the balance of unit XYZ and get a mix of responses including "massively overpowered" and "totally useless".

And when I ask about this kind of approach to balancing I get a mix of responses including "the only way to go" and "totally useless".

;D
Have ideas or bug reports for one of our games? Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: A possible systematic approach to balancing the fleet ship types
« Reply #28 on: December 06, 2010, 03:54:06 pm »
Ask any 10 people their opinion, get 11 different opinions, right?  ;D
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Mánagarmr

  • Core Member Mark V
  • *****
  • Posts: 4,272
  • if (isInRange(target)) { kill(target); }
Re: A possible systematic approach to balancing the fleet ship types
« Reply #29 on: December 06, 2010, 03:55:12 pm »
Isn't it lovely being a balancer? :D I'm more of the opinion that maths are good for laying the groundwork and playtesting is for ironing issues out and doing the finaly touch up.
Click here to get started with Mantis for Suggestions and Bug Reports.

Thank you for contributing to making the game better!