Hexnibbler
Recruit
I think the current formula for peasants/caravans to "give" (bribe/take everything is broken.
The "ratio" applied to the army makes it so that the decision is made based on proportions rather than difference :
I have this results with the formula ((to maximize chances)
power of units :
T5 cav = 2.1*1.2 = 2.52
T1 peasan = 0.66
peasants < (power * 0.4 * 0.5) / (1->2)
y < x * 0.4*0.5 *0.5
y/x < 0.4*0.5*0.5 = 0.1
x >= y *10
(x*3.81) >= y * 10
x >= y * 10 / 3.81
x >= y * 2.62
-> at ratio 0.4, it takes 2.62 times more T5 cav than there are peasants holding 10k wealth.
ratio 0.1
y/x = 0.1*0.5*0.5 = 0.025
(x*3.81) >= y * 40
x >= y * 40/3.81
x >= y * 10.5
-> 10.5 times more T5 cav than T1 peasant at ratio 0.1
ratio 0.05
y/x = 0.05 * 0.5 *0.5= 0.0125
x >= y * 80 / 3.81
x >= y * 21
the issue is that the relation of proportions gives weird logic : if there is one peasant, it needs 21 T5 cav to consider it certain he will give in. But if there are 2 peasants it multiply that by 2, to 42 T5 cav.
now imagine 10 peasant. or 50. So the ideal course of action is to kill everyone but one peasant/caravan dude.
maybe my computation is wrong, but I think it's due to the "acceptablePowerRatio" being multiplied to the player troops (instead of computing a difference in power), or because that ratio is too low (0.4, 0.1 are big factors). Or a mix of both (higher ratio value, and raw difference in power level)
the other issue I found is that the "wealth" of the peasant used to push them to resists more is used to decrease the player troop power rather than to increase their own power level. Ideally, having that wealth factor the power of peasant would allow to greatly affect their power level by killing them.
and there is no "rogue skill" impact on those chance to bribe/take everything (aside from two face perk). There should be.
I think at 5k wealth it's 2/3 of the numbers shown, and at 1k wealth its 1/2. But this proportion thing is weird.
The "ratio" applied to the army makes it so that the decision is made based on proportions rather than difference :
I have this results with the formula ((to maximize chances)
power of units :
T5 cav = 2.1*1.2 = 2.52
T1 peasan = 0.66
peasants < (power * 0.4 * 0.5) / (1->2)
y < x * 0.4*0.5 *0.5
y/x < 0.4*0.5*0.5 = 0.1
x >= y *10
(x*3.81) >= y * 10
x >= y * 10 / 3.81
x >= y * 2.62
-> at ratio 0.4, it takes 2.62 times more T5 cav than there are peasants holding 10k wealth.
ratio 0.1
y/x = 0.1*0.5*0.5 = 0.025
(x*3.81) >= y * 40
x >= y * 40/3.81
x >= y * 10.5
-> 10.5 times more T5 cav than T1 peasant at ratio 0.1
ratio 0.05
y/x = 0.05 * 0.5 *0.5= 0.0125
x >= y * 80 / 3.81
x >= y * 21
the issue is that the relation of proportions gives weird logic : if there is one peasant, it needs 21 T5 cav to consider it certain he will give in. But if there are 2 peasants it multiply that by 2, to 42 T5 cav.
now imagine 10 peasant. or 50. So the ideal course of action is to kill everyone but one peasant/caravan dude.
maybe my computation is wrong, but I think it's due to the "acceptablePowerRatio" being multiplied to the player troops (instead of computing a difference in power), or because that ratio is too low (0.4, 0.1 are big factors). Or a mix of both (higher ratio value, and raw difference in power level)
the other issue I found is that the "wealth" of the peasant used to push them to resists more is used to decrease the player troop power rather than to increase their own power level. Ideally, having that wealth factor the power of peasant would allow to greatly affect their power level by killing them.
and there is no "rogue skill" impact on those chance to bribe/take everything (aside from two face perk). There should be.
I think at 5k wealth it's 2/3 of the numbers shown, and at 1k wealth its 1/2. But this proportion thing is weird.