Morale is king in 1.8?

[Pornek]

Tactics/Discipline were the most important military stat in 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6 and 1.7.

In 1.8 the most important military stat is ... Tactics/Discipline.

Yep.
Have 0.8 tactics advantage can attack into mountain over river for -3. Win battle with 800 to 4000 casulties.

 

[zodium]

* Imagining zodium working at the combat math on Christmas day *

I'm a methodologist statistician, it's what I do. Deal with it.

The combat algorithm is actually an interesting optimization problem!

 

[unmerged(2833)]

With the new changes to damage calculations (exponential values from dice rolls replaced by linear), battles are much more of a grind in 1.8. Big roll differentials don't move the dial like they used to, dampening the difference from general, pips, terrain, and plain old fashioned luck.

Decisive battles are much harder to achieve because of this (at least in their old form).

My take is that the tradeoff between staying power (morale) and killing power (pips, and to a lesser extent discipline) has swung decisively in favour of morale. It's enough to rethink Defensive again as your first military idea set to take, rather than Offensive.

What are everyone's thoughts/observations?

do you know how damage is calculated in first place? Because the difference between linear and exponential (assuming it was that way, it wasnt said in patch note) can be small.

Is there some documentation about combat anywhere?

/Edit. Oh, wiki. Well, if its not this patch but previous, the curve was not that different from linear, and actually linear at higher die rolls.

 

[zodium]

do you know how damage is calculated in first place? Because the difference between linear and exponential (assuming it was that way, it wasnt said in patch note) can be small.

Is there some documentation about combat anywhere?

/Edit. Oh, wiki. Well, if its not this patch but previous, the curve was not that different from linear, and actually linear at higher die rolls.

Remember that the resultant casualty difference is per battle step, and that this difference propagates through Regiment Strength (if for even rolls, army A takes more casualties than army B in step 1, then the casualty difference in step 2 will be even larger than in step 1, and larger still in step 3, etc.) Straightening out the base die tables is huge.

 

[nouli]

do you know how damage is calculated in first place? Because the difference between linear and exponential (assuming it was that way, it wasnt said in patch note) can be small.

Now it's 5 * roll + 15 (at least for casualties) (graphs)

 

[Magean]

Now it's 5 * roll + 15 (at least for casualties) (graphs)

So you've done the maths for the new system. Awesome

What do you mean by "mean damage" ? Mean as to what ?

Edit : also, with a formula such as this one, it still seems better to increase pips rather than combat ability. +1 pip gives you 5 more damage, multiplied by the rest of the casualties formula. +X% combat ability increases damages by base casualties*rest of the formula. If you can get good rolls and high base casualties, combat ability will be very efficient, but it's more reliable to choose a consistent increase.

 

[Squirrelloid]

I'd say the king is 'number of units'. It's a lot harder to win now when outnumbered (although a significant tactics / unit performance advantage can achieve it, but it's not as dominant as it used to be).

 

[Magean]

I'd say the king is 'number of units'. It's a lot harder to win now when outnumbered (although a significant tactics / unit performance advantage can achieve it, but it's not as dominant as it used to be).

In my Hansa game, it's indeed harder to win when I'm outnumbered. But when I outnumber Branderburg, they still do very well

 

[Oorlog]

On a related note. Does anyone know when two ally armies fight an enemy army, if the one with the best tech (morale, discipline, tactics) is used? or is it averaged?

 

[TheMeInTeam]

With linearization of the base die tables, there is no longer an obvious 'king'. If I had to take a guess, the variable to optimize now will be a ratio or derivative of some kind, but I can't look at the math till around Christmas.

I think "numbers" (as in troop count) took a larger role than it used to have, and it was already significant.

 

[nouli]

What do you mean by "mean damage" ? Mean as to what ?

Mean of damage for given roll modifier (generals + pips etc.)
1/10 Sum(i=0 to n)( based_damage(total_modifier + i) )
Seems like the easiest way to compare damage without random factor.

Edit : also, with a formula such as this one, it still seems better to increase pips rather than combat ability. +1 pip gives you 5 more damage, multiplied by the rest of the casualties formula. +X% combat ability increases damages by base casualties*rest of the formula. If you can get good rolls and high base casualties, combat ability will be very efficient, but it's more reliable to choose a consistent increase.

1 pip gives you 5 more dmg only for 10% of rolls. In 90% of cases your damage doesn't change.

The general problem here is propagation of damage throughout battle (as zodium mentioned). That's what complicates the whole thing. If not for that, we could simply solve mean equations at roll modifier 0 and analyze variables influence (I'm actually not sure if I'm right or wrong here, so I will leave this equation here: 15/2 (-1 + (d^2 mod mt2)/(d2^2 mod2 mt)) == b2, where mod is modifiers, b is pip advantage, mt is military tactics, _2 is second army. It may give some general guideline until someone does full analysis).

 

[Magean]

1 pip gives you 5 more dmg only for 10% of rolls. In 90% of cases your damage doesn't change.

Thanks for the reply.

Also, why the quoted part ? Is it that common to roll less than -2 ?