Leathality vs attack trade off

[godsakes]

is there anyone who can give me an idea as to which is a better stat to look for in a troop?

for example:

ilergetan soldiers have 0.225 Leathality and 10 attack

Celtiberian Heavy Infantry have 0.13 leathality and 12 attack


(putting aside defense stats) presumably the ilergetan soldiers are more deadly? how much extra attack would the Celtiberian Heavy Infantry need for this to even out?

[Labrat]

Attack factor is compared with the target's defence factor to determine how likely it is for a strike to hit. Lethality is the chance that a hit translates into a kill; if it does not the target will be knocked-down.

If I understand the mathematics behind the attack-defence comparison (https://forums.totalwar.org/vb/showthread.php?p=1196063#post1196063) correctly, the defence factor also plays a role in whether attack or lethality are more important. The greater the discrepancy between attack and defence, the smaller the effect of an additional point of attack is. In EB's statting system, attack and lethality are also linked: high-lethality weapons are often less wieldy and therefore have a lower attack factor.

[econ21]

If I understand the mathematics (and Puzz3D is right):

Increasing your attack by 2, increases your chance of a hit by about 20%.

Increasing your lethality from 0.13 to 0.225, increases the chances of your hit being a kill by 73%.

If that is so, then +2 attack for the short sword does very little to "balance" the higher lethality of the longsword. Other things being equal, I would say the longsword is 44% better (1.73/1.2=1.44).

The difference seems too great so I wonder if my understanding is wrong.

[godsakes]

If I understand the mathematics (and Puzz3D is right):
Increasing your lethality from 0.13 to 0.225, increases the chances of your hit being a kill by 96%.

is it that much?

i though it'd be (0.225-0.13)/0.13 = 73% or did you apply something from Puzz3D's formula (i couldn't really get my head around it)

assuming you're right, you'd need around 19 attack before things balanced out?

[econ21]

is it that much?

i though it'd be (0.225-0.13)/0.13 = 73%

You are quite right, tx - I have corrected my mistake.


assuming you're right, you'd need around 19 attack before things balanced out?

With your correction, you would need +7.

[antisocialmunky]

There is also attack speed which increases the amount of rolls you get and knockdowns.

[godsakes]

There is also attack speed which increases the amount of rolls you get and knockdowns.

attack speed? there's a seperate stat for that?

from reading the link in the first reply i got the impression the attack essentially represents that, all the units go through a battle cycle the units with higher attack (say quicker shortswords), hit more often then say slower double handed swords (represented by the lower attack) but this is offset by the fact the double handed swords are more likely to do more damage (read: kill) represented by the higher leathality

[econ21]

There is also attack speed which increases the amount of rolls you get and knockdowns.

By attack speed do you mean "Min delay between attacks (in 1/10th of a second)" - i.e. the number before lethality in the EDU? I had assumed longswords would be slower, but both swords seem to have 0 for this stat.

Maybe the animations play a role here?

[Labrat]

from reading the link in the first reply i got the impression the attack essentially represents that, all the units go through a battle cycle the units with higher attack (say quicker shortswords), hit more often then say slower double handed swords (represented by the lower attack) but this is offset by the fact the double handed swords are more likely to do more damage (read: kill) represented by the higher leathality

Almost right. Attack speed determines how often a soldier will make strikes (IIRC the animation also plays a role in this). Strikes can be delayed by the soldier being hit, kocked-down or pushed. When the soldier strikes, his attack value is compared to the defence value of his target. A higher attack value and / or lower defence value increase the chance of a hit. Finally, the lethality value is the chance that a hit translates into a kill, if not the defender is simply knocked-down and can't strike or move for a second or two.


Increasing your attack by 2, increases your chance of a hit by about 20%.

Wouldn't a +2 attack improve hit-chance with 1.2^2 = 1.44 = 44%?

My above post was wrong, BTW: the greater the discrepancy between attack and defence, the greater the effect of an additional point of attack gets. The difference was capped at 20 in M:TW, but Puzz3D says the cap is far higher in R:TW.

[econ21]

Wouldn't a +2 attack improve hit-chance with 1.2^2 = 1.44 = 44%?

If I am understanding things right, yes, in STW and MTW. But the link in this thread to Puzz3D's post suggests the 1.2 constant was lowered to 1.1 in RTW.

[Labrat]

If I am understanding things right, yes, in STW and MTW. But the link in this thread to Puzz3D's post suggests the 1.2 constant was lowered to 1.1 in RTW.

:oops:
-2 for my reading comprehension. I need some caffeine, apparently.

[CBR]

Yes the test we did showed RTW more or less halved the effect so its about 1.1.

Not sure if lethality is linear. But the best approach is to ensure that all weapons have the correct lethality so equal attack/defense also means equal melee strength (except for spears of course). AFAIK the AI does not consider lethality but only the attack/defense points as well if a unit is a spear v non spear unit.


CBR

[bobbin]

By attack speed do you mean "Min delay between attacks (in 1/10th of a second)" - i.e. the number before lethality in the EDU? I had assumed longswords would be slower, but both swords seem to have 0 for this stat.

Maybe the animations play a role here?

Attack speed is only used for cavalry spears and ranged weapons in EB (all others are set to 0) and so isn't a factor.

From what I've heard model animations do affect a units combat performance.

[antisocialmunky]

The only notable example I can think of are that Phalangites have the rapid poke that does alot of pushback/knockback

[econ21]

Not sure if lethality is linear. But the best approach is to ensure that all weapons have the correct lethality so equal attack/defense also means equal melee strength (except for spears of course). AFAIK the AI does not consider lethality but only the attack/defense points as well if a unit is a spear v non spear unit.

Interesting point about the AI not considering lethality. Personally, I don't understand why the EB team made lethality vary between ordinary weapons like long swords, short swords, spears etc. Leaving lethality constant and just varying the attack stat alone can achieve any desired balance and would make the stats so much more transparent to the human, let alone the AI.

Lower lethality in general to get back to STW/MTW combat speeds was, of course, a neat design decision.

[CBR]

There has to be a difference in lethality for the basic weapons used in RTW. Maybe modded weapon animations in EB screws it up even more, can't say. Modders have found the basic numbers needed to make weapons like daggers, swords, spears and two handed weapons equal and the numbers are found in the modding guide:



fs_slow_swordsman / fs_swordsman / fs_semi_fast_swordsman / fs_fast_swordsman : lethality = 0.5
fs_slow_2handed / fs_2handed : lethality = 0.41
fs_slow_spearman / fs_spearman / fs_semi_fast_spearman / fs_fast_spearman : lethality = 0.41
fs_dagger / fs_semi_fast_dagger / fs_fast_dagger : lethality = 0.57
fs_2handed_berserker (scale 1.1) : lethality = 0.38


If one wants slower or faster combat then just multiply the lethality with whatever number one wants and it should still work fine I think.


CBR

[antisocialmunky]

If anyone cares to calculate this out IE-hits before a kill, this is the distribution you should use:

http://en.wikipedia.org/wiki/Geometric_distribution

We just need to figure out the probabilities involved. Then we can make a % kill chart.

We can get a 'how many pokes does it take to pop a unit' chart with this distribution:

http://en.wikipedia.org/wiki/Negative_binomial_distribution

[econ21]

Modders have found the basic numbers needed to make weapons like daggers, swords, spears and two handed weapons equal and the numbers are found in the modding guide:

fs_slow_swordsman / fs_swordsman / fs_semi_fast_swordsman / fs_fast_swordsman : lethality = 0.5
fs_slow_2handed / fs_2handed : lethality = 0.41
fs_slow_spearman / fs_spearman / fs_semi_fast_spearman / fs_fast_spearman : lethality = 0.41
fs_dagger / fs_semi_fast_dagger / fs_fast_dagger : lethality = 0.57
fs_2handed_berserker (scale 1.1) : lethality = 0.38




Wow - where is this stuff reported? It looks like pure gold for serious modders and should not get lost in the ether. :bow:

[CBR]

It is part of "The Complete EDU Guide for RTW" in the Scriptorium: https://forums.totalwar.org/vb/showthread.php?t=88859

It certainly saved me lots of time when working over unit stats for Ran No Jidai. I did manipulate it in such a way that the AI did not always do the frontal suicide charge into pikes. Or at least it seemed that the AI was not totally brain dead when using stats :beam:

edit: pikes in RTW have very low attack stats but Phalanx mode is still a killer and the AI does not see pikes as the tough opponent as it should. So one can increase pikes attack value but lower its lethality. Its overall melee power stays the same but the AI now sees a much more powerful unit.


CBR