How is combat effectiviness calculated?

[Duke John]

The Shogun Cafe has a Unit Comparision tool in which you can see the combat effectiviness calcuated on the stats of the unit.

Does anyone know how this is done? I want to avoid overpowering/unbalanced units as much as possible and I think that I can be a great help on balancing my Wars of the Roses mod. I would really appreciate any information on this.

Thanks in advance,
Duke John

[CBR]

You cant use the Combat rating for that much as it uses morale and armor value too and these values doesnt have anything to do with melee combat power (as long as the unit doesnt rout)

The basic formula for calculating a chance for one soldier to hit an enemy soldier is: 1.9% x 1.2^(attack-defense)

The (attack-defense) value can max be 20.

A simple way of looking at it is to add a unit's attack and defense for overall combat power. We just call it combat points. A FMAA has 3/4 in attack/defense so thats 7 combat points(CP).

For each difference in combat power you will overall kill 1.2^(difference in combat power) more than the enemy.

Example: one FMAA with 7 CP faces a CMAA with 8 CP so the CMAA will kill 20% more than the FMAA will.

Using the 1.2^(difference) formula:

CP difference/real killing power
1/120%
2/144%
3/173%
4/201%
5/249%
6/299%
7/358%
8/430%
etc

So a valor 3 CMAA (6 extra CP) facing a valor 0 CMAA will kill about 3 times as many as it will lose.

And the 3:1 ratio will even be worse as the high valor CMAA soon will outnumber the low valor CMAA.

Edit: I will add some more later..

CBR

[Duke John]

Yes I kinda understood that, but how do you get from that to a combat value? Do you think that these formulas can be used and transformed into some kind of tool to help balancing mods?

Now I'll just wait for your edit.

Thanks, Duke John

[CBR]

Sorry about this late answer. Got held up by other stuff..

One thing I worked on was to leave the formula that CA uses to calculate unitcosts and go to a more simpler system based on miniature rulesets.

CA's system looks good on the surface as increase in killingpower is about the same as increase in cost. It uses 1.2^(attack + defense + 1/5 charge + 1/3 morale + 1/3 anticav stats + some more) for calculating costs.

But when you play with limited money the very good units are also terrible expensive. CA even made a fudge so units around 450+ were cheaper.

An example for the differences is that Chivalric knights were 850 and Lancers 1100 florins in 1.0 but now only 675 and 850. Even with the fudge such units are costly in 5k games.

For the Community Mod I changed the cost formula to use only 1.11^(etc) That fixed a lot of problems although I still have to do some manual rounding up/down.

---

But enough of pointless spam http://www.totalwar.org/forum/non-cgi/emoticons/biggrin.gif

As it is now I havent come up with the perfect formula yet and I dont know if it can be done. There will always be rounding off issues and some stats like speed and charge (that does cost something) cant always be used in a battle (because of circumstances, terrain, skill etc)

But the most important stats are the attack/defense stats and then of course number of men in a unit.

One way is to give one medium quality infantry unit a price (6-700 florins) and then base all others on that unit. For each CP difference add/remove say 100 florins. Look at armourpiercing weapon as just another CP and same thing with having anti-cav ability. Of course this is proportional with unitsize. And cavalry is multiplied with 1.5-2.0. If high morale then maybe add another 100..depends.

That is a pretty simple system that should give ok results. Then you can always take a look at the overall picture and see if its either too cheap/expensive and modify all units with +/- 100. Then look at different army setups and do the min/max tests: What if I take super units only, is there a counter to that?

If you try and keep it simple and use nice big blocks of 100 florins its easier to do IMO.

I hope this helps a bit heh.


CBR

[Puzz3D]

Duke John,

You start with a concept of the gameplay you want to achieve and some idea of how each unit will function in the overall picture, and I assume you have those ideas already established. Then set the stats of the unit you envision as the core fighting unit. For instance, in we/mi v1.02 we selected the yari samurai as the core unit around which everything else was built. Then you create the second unit's stats and run tests against the first one adjusting the second unit's stats until you get the battle results you want. That means which unit wins, by what margin and how long it takes to win. The analytical formula simply tells you that each attack or defend combat point provides 20% advantage in chance to kill. My rule of thumb is that a 20% advantage will get 6 wins out of 10, and a 44% advantage will win about 9 out of 10. Roughly, about 4 or 5 points of charge equals 1 point of melee. You'll also want to test flank and rear attacks to make sure the results are close to what you envision meaning how quickly the unit collapses if hit in the flank or rear. This early test is where you choose the initial morale level, but keep in mind that in a full 16 unit battle there are a lot of morale penalties flying around and single unit tests don't take this into account. So, I would suggest to do some full blown 16 unit battle tests with the first few units in case a morale adjustment is required because altering the core unit later on means going back and retesting everything. Lastly, you would balance the ranged units which is probably the most difficult thing to do. As long as you do periodic full battle tests, the end result should come out to be as you originally envisioned. I'd be careful about making units too slow because fatigue is time based and was optimized for the smallest map size. So, on large maps slow units are going to be functioning at high fatigue levels most of the time which will dramatically affect their combat power, and their morale level.

Any mod can only function correctly at the designed valor level because morale is tied to valor upgrades which affects how long units fight, and excessive upgrading also tends to swamp out special bonuses such as the anti-cav bonus which can alter a unit from it's intended role. CBR has addressed that in his mod by raising the price of units so that upgrading isn't economical, and this also servers to limit the number of stronger units because you don't have enough money to buy lots of them. No matter how balanced the mod is, you'll find that many players won't like it. In fact, what seems to be most popular is tons of units and lots of upgrading.

[Duke John]

CBR and Puzz3D
Thanks alot for your replies http://www.totalwar.org/forum/non-cgi/emoticons/gc-2thumbsup.gif

Sounds like I need to alot testing...or would you be so kindly?

I have already in my mind how the units should react and I made some initial stats for them. It seemed kinda good, but after reading Puzz3D's method I think the stats will need to be redone from start.

After some hints from you early on I decided that the Wars or the Roses mod will be designed for 10K. The high cost of the units will hopefully discourage upgrading although you can never stop the 99K'ers.

I hope that I can release a beta for playtesting today or tomorrow and if you two can also help then that would be great. Perhaps I will see you some time soon in a WotR battle http://www.totalwar.org/forum/non-cgi/emoticons/wave.gif

Cheers, Duke John