Mathmatical Balence Model?

[BackTrak]

Hello dev team!

I was kicking around the idea that it should be possible mathmatically model an allegiance core so that each factions tech could be balanced with some degree of certainty. Based on what's going on in other core related news, it sounds like a new core could be coming to plan shortly. The time is certainly ripe to attempt to deliver a core that would not have a long revision cycle before it becomes playable. As is well known, an initial impression is vital to success, and a good first core release is money in the bank. (well, happy player feelings in the bank, anyway /wub.gif" style="vertical-align:middle" emoid=":iluv:" border="0" alt="wub.gif" /> ).

Would it be possible to model basic features like:

* Income over time
* Total damage over time
* etc.

So that given an average game duration of 1 hour (or some other agreed upon average duration of time), at the end of the hour all factions would have achieved the same amount of income, caused the same amount of damage, etc.

Essentially, I would like to see a model that could accomodate fast tech factions, late blooming factions, and strong tech path factions and have all key performance indicators align at a set point in time. If this convergance of statistics can occur at a point in time, then the game should be squarely based on the players' (and commander's) skill alone.

Does it have merit? Has it been tried before?

Keep up the great work guys, it's excellent fun!!

[holtonster]

I like the idea, but the biggest problem I have with coming up with a mathematical model for cores is how to take into account many of the non-mathematically based balance methods. Things like hit boxes and speed can make a big difference in game, yet I cannot think of a way to mathematically determine their effect on the game. However, I wouldn't mind helping out with the research on this, if you want to give it a shot.

[BackTrak]

QUOTE (quail @ Jul 4 2007, 02:08 AM) I like the idea, but the biggest problem I have with coming up with a mathematical model for cores is how to take into account many of the non-mathematically based balance methods. Things like hit boxes and speed can make a big difference in game, yet I cannot think of a way to mathematically determine their effect on the game. However, I wouldn't mind helping out with the research on this, if you want to give it a shot.

Hi Quail, good point.

How about if we started by considered hit boxes as a single number representing the cubic pixels of the hit box. We might also have include an adjustment for the smallest box face (phoenix) if that has an impact, which I think it might.

What might be a good goal would be an overall number that can represent all the ships' capicity at a given point in time such that:

Scout Magic Number
+ Fighter Magic Number
+ Bomber Magic Number
= Total Offensive/Defensive Capacitiy

Miner Magic Number
+ Support base Magic Number
= Total Econ Capacity

etc.

So, we would have to locate things about each ship type than could be used to commonly describe them.

-- Speed
-- Hitbox area
-- Average Effective Range in sectors (boost2 adds a sector to an int, so forth)
-- Damage Capacity Magic Number (This is a function of how much OVERALL damage this ship type can do based on time, as some ships loose steam over time while others gain it)
etc...

These magic numbers will be the very tricky parts. One should be able to adjust time as a variable, and have the magic number always hold true. Once this equation is determined, balance changes to any measured part of the core should be able to be accuratly proven as balanced.

I have more travelling to do, so I will noodle more, suggestions needed!

[Barrager]

if this were a solo game its easy to calculate balance. if this were a game where you had 2 sides and a simple plan of killing the other side and capture a point it gets harder but it;s still doible to balance using math.

The biggest fact is the more teamwork you add into a game th harder it is to balance, the variables get larger. not saying it isn't possible but it would take some one way too much dedicated time and much more to the point it's much harder to calculate for a smaller community. you have more experienced soldiers being noticed and the cannon fodder getting pretty much ignored.

[BackTrak]

QUOTE (Barrager @ Jul 6 2007, 09:11 AM) if this were a solo game its easy to calculate balance. if this were a game where you had 2 sides and a simple plan of killing the other side and capture a point it gets harder but it;s still doible to balance using math.

The biggest fact is the more teamwork you add into a game th harder it is to balance, the variables get larger. not saying it isn't possible but it would take some one way too much dedicated time and much more to the point it's much harder to calculate for a smaller community. you have more experienced soldiers being noticed and the cannon fodder getting pretty much ignored.

There's no ignoring that the team element makes or breaks any given allegiance game. I say ignore the skill problem completly, and focus on making the playing field level. One makes a pool table as flat and level as possible, but one does not try to pre-determine the skill of a pool player and adjust the table accordingly.

What if we just assume that each team is playing at 100% efficiency. Then, if we set our sights on a convergence every 15 minutes, it would be possible to say that if both team x and team y are making no mistakes, and operating at 100%, then they will be at a stalemate in score and advantage every 15 minutes. Thus, if one side gets a leg up, but the other side hangs in there then after 15 minutes, the game should be back up for grabs. That gets those big swings in team control that make for excellent games.

IE: balance factors get set up for a giga scout rush in the beginning. If the bios team can hold onto all thier cons and miners for 15 minutes, some bios tech would arrive that would neutralize the scouts. Next, bios gets light gun ships, and giga needs to hold for 15 minutes until they get thier next balancing tech + an advantage tech, so forth and so on.

Would this mean calculations for every viable strategy? I don't know, but I'm leaning towards no, because you should be able to assign a value to anti-utility type tech vs anti-small ship tech vs anti-base tech, etc.

You are right, to make these formulas would require more time that building the core itself, but after they are made they could be reused to balance every core after.

[spacejocky]

Maybe I'm not understanding the 'thought' behind the idea. Original post I thought was a "Core" balance of sorts...to make the factions even based on tech path's/development after a period of time. (If a Comm fails to develop properly then his team would/should be at a dis-advantage still right ?)

But then it seemed to turn into an "In-Game" balancing idea ?

I can see where "balancing" a game in progress can be a good thing, like say against an obvious stack/greater number of players...

But I thought also, A Team works hard to gain an advantage...either in Tech or KB, and a great deal of strategy involves pressing the advantage's gained by teamwork and development. "Evening" out the playing field during the game would seem to negate that effort somewhat...(Would lesser weapons become more powerfull, or vice versa ?)

If a team knows at some point the field will be leveled, what purpose trying to gain advantage early on ?. Better perhaps to expand slowly and fortify than explore and conquer...Unless perhaps to end game early B4 the balancing...

(Will go stand in corner if I missed idea behind post totally)

[Andon]

I think you missed the point. I think it is more of a pre-game leveling, not even in Allegiance but in the core itself. It seems like a good idea, but I'm not sure how well it will work.

[holtonster]

I doubt that if such a program/equation were ever written, it wouldnt get serious use. People have a hard enough time deciding balance with a their computer telling them what to do. However, I do think itd be fun just to see if it can be done well and see the results.

BackTrak, I think your suggestion above is on the right path, especially breaking it down into different parts (econ/offense/defense/ect). However, Im a bit hesitant about combining the numbers at the end, as I feel it would be comparing apples to oranges. However, a modifying factor could be use to make it closer.

A good place to start might be with the faction stats that have numerical modifiers (He yield/starting money/miner cap/ect). Ill try to take a look at some of them this weekend.

[BackTrak]

QUOTE (quail @ Jul 6 2007, 11:39 PM) I doubt that if such a program/equation were ever written, it wouldnt get serious use. People have a hard enough time deciding balance with a their computer telling them what to do. However, I do think itd be fun just to see if it can be done well and see the results.

BackTrak, I think your suggestion above is on the right path, especially breaking it down into different parts (econ/offense/defense/ect). However, Im a bit hesitant about combining the numbers at the end, as I feel it would be comparing apples to oranges. However, a modifying factor could be use to make it closer.

A good place to start might be with the faction stats that have numerical modifiers (He yield/starting money/miner cap/ect). Ill try to take a look at some of them this weekend.

Excellent. It looks like TigerEye has already figured out how to read the core files for his kneeboard. With some tweeking, we could probably make a core file analyzer that can apply the formlae, and then tell which factions are +/- some balence points. This was one could tune the core, and then check it for balance right away.

[holtonster]

QUOTE (BackTrak @ Jul 6 2007, 09:56 PM) Excellent. It looks like TigerEye has already figured out how to read the core files for his kneeboard. With some tweeking, we could probably make a core file analyzer that can apply the formlae, and then tell which factions are +/- some balence points. This was one could tune the core, and then check it for balance right away.

Theres one already. Look on the acadamy for ICE.