Bad dice

[tabanli]

I also notice that Scouts and Crows always loose to Archers in forest, Orcs in Swamp, Light Cavalary in plain, and Dwarfs in mountain. They are always under 10% so they loose. I think it is a bit too much but I guess you take good with the bad.

[KGB]

I've noticed that also. Same goes with Dragons vs certain solo units.

I think the 90% rule should be amended to allow 1 unit to always be lost regardless of how unlikely. This would enable those scouts/crows in 1v1 battles to occasionally make a kill and allow a lone Orc to have a shot at a super Barbarian.

KGB

[nagrogin]

I just suffered an attack loss where the odds were 84.3% and the total value on both
sides was well over 400.

There was ambush on both sides as well, but no conditional bonuses (anti-air, etc).
Just straight-up leadership and morale bonuses, which I understood to be included
in the total-power tally. Is this wrong? (e.g., only leadership is counted and not morale?)

The opposing sides were:

Att : 49,49,49,74,74,75 = 370
Def : 52,75(3life),42,42,72 = 283

If I have not botched the math, the total is 653.
Last I saw about the bad-dice implementation, early in this
thread, is that the battle is not lose-able at >80% if the total
strength is >400.

Is this no longer the way it works? If not, how does it currently work?

I can provide the mods with a screencap of the battle summary (or just the
gameID) if helpful.

[piranha]

I answered you message but will answer here to so others can see.
When deciding if using 90, 85 or 80 as the rule only personal bonuses are counted such as terrain, no army bonuses.
So in this case the total value is 365. Have to say it was one of the unluckier battles I've seen.

[Chazar]

Question? 90% Rule, does it also apply to Ruins and UL ? I send a message to Solo saying that it does but now I am not sure.

No, it only applies to actual combat. Ruins are simply fought 'as is'

Ruins offer something similar though, as a high enough UL guarantees a win - I think this is enough.

Furthermore, due to the way warbarons is rounding floating point numbers, a hero with 10UL is already save in L2 Ruins versus Skeletons, etc. (this was already reported a while ago)

[tabanli]

I think the value calculation does not reflect the real value.

25+25+25 is not equal to 75 as 75 strength dominates 3 units of 25.

If a 75 strength unit attacks a city full of 25 strength units, it will kill ON AVERAGE 8 units. So we need to look at the power,

S=Strength
H=Hit
P=Power

P = H*(100-S)/S

The true value of a stack is the sum of individual powers (without ambush).
Ambush makes calculations a bit messy but we can make an approximation. Remember ambush can take one or two hits since most units are two hit units. If A is the average final average ambush of the enemy stack,

V=Value
V= Sum (H-1.5A)*(100-S)/S

Basically for each unit you need to calculate the possibility of ambush and subtract that from the hit, but for a 3 hit unit u need to subtract 2 (Ambush has possibility of taking 1 hit, 2 hit or 3 hit so the average is 2) for a 4 hit unit subtract 2.5 etc. If a unit has personal ward, it's ambush can be modified. Fight order also matters but we can't derive a general formula for each and every case.

Again this is an approximation that better reflects winning chances of two stacks with equal number of units in it e.g 8 vs 8 rather than adding the strengths. It may not work if 8 attacks 32. Believe it or not, I make this calculations before engaging in a critical battle.

[KGB]

Tabanli,

But Piranha is not trying to calculate the 'real' value as a combat simulator.

He's just trying to determine whether this is an 'important' battle between key stacks or a routine one that happens lots of times per game. Summing the strengths of units without stack bonus's is a quick way to determine whether there are lots of higher level units involved and when there is, that's a pretty good indication this is an 'important' battle.

KGB