On Sacking...

[Kobal2fr]

Following complaints that "blitzing is too easy - sacking gives more cash than economy !", I set out to figure out how sacking worked exactly. My initial assumption was that the game took the overall population of the captured city, substracted the dead, and gave the rest as florins, plus a bonus for buildings damaged/destroyed in the sack and/or settlement type.

I was wrong. Again.

Research Protocol

I used York as my testbed, because it's always captured by Rufus' army with autocalc. First I captured it on vanilla settings. Then I put a great_market in it and took it. Then I gave it walls and took it. Then I set it back to an empty village, gradually increased population and took, took, took it. All of those are in new campaigns each time obviously.

Results

Money gained from sacking is the same wether York is an empty village, a village with great_market, or an empty Town (with wooden walls). The Great_Market is destroyed each time.

Population-dependant results are a bit more complicated. Here's the data for sacks :

Population	Sack money	Pop. killed	%money/pop	%killed
800	240	168	30%	21%
1.500	800	315	53,3%	21%
3.000	2.000	630	66,6%	21%
5.000	3.600	1.050	72%	21%
10.000	7.600	2.100	76%	21%
15.000	11.600	3.150	77,3%	21%
20.000	15.600	4.200	78%	21%
30.000	23.600	6.300	78,6%	21%
60.000	47.600	12.600	79,3%	21%
100.000	79.600	21.000	79,6%	21%

and here's the date for exterminations :

Population	Cash made	Pop. Killed	%money/pop	%popkilled
800	120	300	15%	37.5%
1.500	400	1.000	26.6%	66.7%
3.000	1.000	2.250	33.3%	75%
5.000	1.800	3.750	36%	75%
10.000	3.800	7.500	38%	75%
15.000	5.800	11.250	38.7%	75%
20.000	7.800	15.000	39%	75%
30.000	11.800	22.500	39.3%	75%
60.000	23.800	45.000	39.7%	75%
100.000	39.800	75.000	39.8%	75%

From this, we learn 4 things :
- Extermination money is exactly half sack money, for 3.5 times the citizen killed
- killed citizen are linearish (the turning point is probably 2k)
- cash made by killing them isn't linear at all. Sacking converges towards 80%, Extermination towards 40%.
- thus, there's no relation between the two.

Further tweaking

As I said in another post, I just found out about two variables in the descr_campaign_db.xml, sack_money_modifier and exterminate_money_modifier. sack_money is set to 0.4 by default, exterminate_money is 0.5

I set sack_money to 0.8, and redid the test.
As expected, sacking now made twice as much cash, for all pop values. As not-expected-at-all, so did extermination.

I set it back to 0.4, set extermination_money to 1.0, redid the test.
This time, sack money was back to normal, and extermination money doubled, effectively making it on par with sacking.

In neither tests did the amount of dead people rise.

So despite their similar names, sack_money_modifier deals with how much money is made when taking a city, and extermination_money_modifier is in fact applied afterwards in the specific case of genocide. There doesn't seem to be a separate one for occupy, as is occupying gives the same amount genocide does.

But, and this is the important point here : it's perfectly doable to make blitzkrieg less viable financially, by tweaking a single variable. Setting it to 0.1 should be about adequate, I think. That said, I've been playing LTC for a long time now, and I don't really recall the population and pop.growth figures of the vanilla game, so 0.1 may be a bit extreme ?

Discuss.

EDIT : I just realized that cash for sacking is about equal to 2*population*sack_money_modifier. The difference must come from the same reason trade and taxes numbers don't exactly add up, i.e. lotsa rounding.
But this formula makes figuring out the right modifier easier.

[Askthepizzaguy]

Excellent post. Good info!

Note on exterminations:

The oddity in your percentages could be explained because there is a limit to how many citizens you can exterminate. Once you reach minimum, you cannot exterminate anyone! There will always be citizens in a province, in other words.

[FactionHeir]

Minimum is 400, as in RTW.

[SadCat]

From this, I learned 6 things:
- Extermination money is exactly half sack money, for 3.5 times the citizen killed.
- There is .5 relationship between sack and extermination.
- killed citizen are linearish (the turning point is probably 2k)
- cash made by killing them isn't linear at all. Sacking converges towards 80%, Extermination towards 40%.
- Occupying gives the same amount genocide does. Yes I had missed this.
- You do great work!

Discuss?
sack_money is set to 0.3 should be 60% and Extermination/occupation 30%
sack_money is set to 0.2 should be 40% and Extermination/occupation 20%
sack_money is set to 0.1 should be 20% and Extermination/occupation 10%

I spend my life short of cash so vote for Sack_money no lower than .2

You would receive next to nothing for taking a small to mid size city.

So please move exterminate_money to say 0.75?

That way Sacking would give only 25% more than extermination/occupation.

Yet would move occupation up to 30% for Sack_money of .2?

SadCat

[sapi]

Nice work

[Kadagar_AV]

Not much to discuss...

Helps a lot, thanks mate

[Didz]

It would be interesting therefore to switch this financial reward system off and see if it makes any difference to the behaviour of blitzer's. With presumably less cash incentive would they be forced to play the game, if only to raise the cash needed to continue their rampage of destruction.

Or is the aditional tax revenue from their conquests enough to sustain the blitzkrieg.

[WhiskeyGhost]

I'd personally would want to see someone make Exterminate the bigger pay-off as apposed to sacking. That way, you can make quick bucks if you want, at the cost of reputation, and of course the economy your damaging by killing all those people. Then again, i don't use any of the mods that are out and about, so there may be one which already does it, and if thats the case, could someone post a link towards which one it is? (sorry, i have problems finding my shoes, let alone something specific, even with search engine)

edit: actually, i mean to add that i hoped they do what i stated above as part of next patch, like Wishlist for 1.3

[Kobal2fr]

I'd personally would want to see someone make Exterminate the bigger pay-off as apposed to sacking. That way, you can make quick bucks if you want, at the cost of reputation, and of course the economy your damaging by killing all those people. Then again, i don't use any of the mods that are out and about, so there may be one which already does it, and if thats the case, could someone post a link towards which one it is? (sorry, i have problems finding my shoes, let alone something specific, even with search engine)

edit: actually, i mean to add that i hoped they do what i stated above as part of next patch, like Wishlist for 1.3

Well that's easy : if you set the exterminate_money_variable to say 2.0, then genocide'll pay twice as much as sacking.

And thanks, because I tested this just to be doubly sure (it works), and realized I was misguided : Occupy is in fact not affected by the exterminate variable. I suppose it just so happened that my combination of sack*exterminate variables at the time was equal to the hardcoded occupy one when I checked the occupy numbers (I admit I didn't pay any attention to them during my tests, and just checked them once as an afterthought).

Occupy seems in fact to converge towards 10% of pop, no matter the variables.

[danfda]

Nice job, kobal2fr.

I turtle, but always sack enemy towns. Agreed that something should be adjusted.

[John_Longarrow]

Kobal,

Great job! I will have to give you the most sincere form of flattery I know of, I'm going to swipe your research for the mod I'm starting on!

[Silverhawk]

By looking at the numbers you posted, the formulae for cash appear to be linear (pop*exterminate_money_variable)-200 = cash gained from extermination and (pop*sack_money_variable)-400 = cash gained from sacking.

You could theorize that the - variable is dependant upon the money_variable as well, as the limited data shows it correlates to an increase in the variable in a linear fashion. If the value were set at 1.0 I'd suspect the value would be -500.

Thoughts?

[alpaca]

A bit more number crunching:
There's a perfect linear fit for a formula for sacking money of:

m = s + (p-40000)/100 = p/100 + s - 400 = p/100 + p*0.79 -400 = p*0.80 - 400

where s is the surviving population and p is the starting population.

However I wonder where the modifier kicks in... You said you got twice as much for a modifier of twice the initial value. Could you post your data for that or even better a modifier of 1.3 or so?
Did you only make twice as much money or were twice as many people killed, too?

Theoretising, the formula could be m = p*modifier*2 - modifier*1000 or m = p*modifier*2 - 400

[Kobal2fr]

Exactly twice as much money, exactly the same amount of people killed, in all cases.

Did only three of those tests before figuring the pattern was verified : 480fl at 800, 1600fl at 1500 and 23.200 for a pop of 15.000

23.200 = 15.000*2*0.8 - 800, or m=p*2*var-1000*var

My current modifier is set to 0.15, sacking York (800 souls) nets me 90fl, or 2*800*0.15-150. Fits.

So OK, sacking really is linear. But in a fiendishly cunning way.

[alpaca]

Exactly twice as much money, exactly the same amount of people killed, in all cases.

Did only three of those tests before figuring the pattern was verified : 480fl at 800, 1600fl at 1500 and 23.200 for a pop of 15.000

23.200 = 15.000*2*0.8 - 800, or m=p*2*var-1000*var

My current modifier is set to 0.15, sacking York (800 souls) nets me 90fl, or 2*800*0.15-150. Fits.

So OK, sacking really is linear. But in a fiendishly cunning way.

Fiendishly cunning? Never seen a matrix didya? They are linear, too...

[Kobal2fr]

I was being ironic . I merely realized I sucked at basic math even worse than I thought for not figuring it out

Then again, it's been a while since i've stared a first degree equation in the face, nevermind a matrix. So I'm allowed to suck.

[alpaca]

I was being ironic . I merely realized I sucked at basic math even worse than I thought for not figuring it out

Then again, it's been a while since i've stared a first degree equation in the face, nevermind a matrix. So I'm allowed to suck.

Sucking at maths is not excusable under any circumstances whatsoever

[slippybee]

Interesting research.

Any thoughts on how extermination/Sacking/Occupy impacts on Generals traits?

I've been deliberately avoiding sacking/exterminating settlements over my Spanish campaign and notice that not only does it increase Chivalry it also increases the Loyalty of a General.


Is chivalry and loyalty directly related?

[Kobal2fr]

That's another matter but yes, Occupying will give chivalry through StrategyChivalry, and genocide gives StrategyDread.

here's the triggers :

;------------------------------------------
Trigger extermination1
    WhenToTest SackSettlement


    Affects Despoiler  1  Chance  20 

;------------------------------------------
Trigger extermination2
    WhenToTest ExterminatePopulation


    Affects Genocide  1  Chance  100 
    Affects Bloodthirsty  1  Chance  10 
    Affects StrategyDread  1  Chance  100 
    Affects HarshJustice  1  Chance  20 
    Affects Authoritarian  1  Chance  3 

;------------------------------------------
Trigger extermination3
    WhenToTest OccupySettlement


    Affects StrategyChivalry  1  Chance  100 
    Affects NonAuthoritarian  1  Chance  3 

;------------------------------------------
Trigger extermination4
    WhenToTest OccupySettlement

    Condition Trait StrategyChivalry > 0

    Affects LenientJustice  1  Chance  20 
    Affects NonAuthoritarian  1  Chance  3

I don't think loyalty is affected at all, but attacking with high odds in the general's favor increases loyalty, that may be what you're witnessing.

[Seabourch]

Interesting! I used to think extermination killed the entire populace, same for sacking!

[Fußball]

Good work on the testing and research.

Tschüß!
Erich

[Odin]

Great post, I have a few thoughts.

1. This sacking and extermination variables do they take into account the religion of the area?

2.Isnt there a file/tweek (forgive my ignorance of the files) that allows you to obtain cash from destroying buildings?

3. If yes to 1. wouldnt it be logical to have a christian army kill more non christians? So should there be seperate modifiers for religion predomance in a region?

4. if yes to 2., can it be linked to the sacking modifier? It seems to me a bit more cash should be given for destroyed buildings and the money derived not soely from the population.

Just something that came to mind, great post.

[Kobal2fr]

1. I shouldn't think so, but I haven't explicitly tested for this either. I'll try to do a quick test for this sometime by making York 100% Muslim

2. You mean buildings destroyed when you sack/genocide a city ? There's no such variable that I've seen - yet. My knowledge of the different .txt files is none too shabby, and still I find new stuff all the time

3. Again, haven't seen any defined modifier or setting relating to this, and if there were I'd assume they would be in descr_campaign_db.xml or maybe descr_settlement_mechanics.xml , both of which I could by now almost quote verbatim by heart

4. I'm sure alpaca or some other scriptomancer out there could whip out a code snippet that would do this if asked nicely, but I don't think a) such mechanism is in the vanilla game b) it'd be as easy/straightforward to put in as copy/pasting a line here and there or fiddling with existing variables.

[Odin]

2. You mean buildings destroyed when you sack/genocide a city ? There's no such variable that I've seen - yet. My knowledge of the different .txt files is none too shabby, and still I find new stuff all the time

Well yes, If you destroy a building you get cash (if I recall), when you own a city so there must be some file that does that. (you know calculates the cash you get). Your test showed that sacking is exclusive to population, and while thats okay, it would be ideal to get the buildings income as well.

3. Again, haven't seen any defined modifier or setting relating to this, and if there were I'd assume they would be in descr_campaign_db.xml or maybe descr_settlement_mechanics.xml , both of which I could by now almost quote verbatim by heart

Well can it be modded? forget the game mechanics for a minute, its logical to me that a christian army is more apt to kill more non christians when sacking. Granted not in all cases, but I think you know what i mean, so back to game mechanics, if I sack alexandria and its a muslim held city and a muslim region I would most likely be killing more of the inhabitants.

The mechanism of population as the exclusive trigger to sacking and income derived is okay, but if there is a possible modification that can take into account religion then it should be considered.

Heck, even a high dread general opposed to a high chivalry general should get more from sacking right? Cant we assume that the later would kill more? The test you did was awesome, thanks for the time you put it, its revealing in both its application and its limitations.

[Kobal2fr]

Well can it be modded? forget the game mechanics for a minute, its logical to me that a christian army is more apt to kill more non christians when sacking. Granted not in all cases, but I think you know what i mean, so back to game mechanics, if I sack alexandria and its a muslim held city and a muslim region I would most likely be killing more of the inhabitants.

The mechanism of population as the exclusive trigger to sacking and income derived is okay, but if there is a possible modification that can take into account religion then it should be considered.

That's a good point, but you could also contend that the decision to sack/genocide Alexandria instead of occupying it represents this bias against people of other religions. Many people use the "always occupy non x-commed catholics" house rule for this very reason (or its opposite, "always butcher the infidel")
In fact, if you haven't sent priests to convert the Levant before your crusaders get there, you sometimes HAVE to genocide to keep a decent public order rating because of combined religious unrest and distance to capital factors (well, in LTC at least)

Heck, even a high dread general opposed to a high chivalry general should get more from sacking right? Cant we assume that the later would kill more?

Already in there. Check the trait triggers I posted further up : Despoiler gives bonus to sacking money, as does Genocide (only Genocide gives dread on top of the bonus to sacking). In fact, you can get both traits, for a max +60% loot. (70% with the Mercenary Captain ancillary on top, if you really want to Get$$$Quick ).
I don't think they kill more, though, or that a general's Dread/Chiv affects the population killed. That's one more test to run I guess. There's always more tests

[Odin]

That's a good point, but you could also contend that the decision to sack/genocide Alexandria instead of occupying it represents this bias against people of other religions. Many people use the "always occupy non x-commed catholics" house rule for this very reason (or its opposite, "always butcher the infidel")
In fact, if you haven't sent priests to convert the Levant before your crusaders get there, you sometimes HAVE to genocide to keep a decent public order rating because of combined religious unrest and distance to capital factors (well, in LTC at least)

I use LTC exclusively, great mod. There are a lot of variables to this discussion when we add in religious factors, that might be more then the engine can handle. Its worth discussing, but I dont think the current system is that bad anyway.... However there is always some one who is ambitous and wants to mod, so if this theory could be applied then I think its a reasonable addition to the game.

Already in there. Check the trait triggers I posted further up : Despoiler gives bonus to sacking money, as does Genocide (only Genocide gives dread on top of the bonus to sacking). In fact, you can get both traits, for a max +60% loot. (70% with the Mercenary Captain ancillary on top, if you really want to Get$$$Quick ).
I don't think they kill more, though, or that a general's Dread/Chiv affects the population killed. That's one more test to run I guess. There's always more tests

Well Im not trying to create work for you mate, honest . However the entire calculation of money gained is based on 3 choices, and the population of the city. So your trait gives you a monetary bonus, but you dont kill more population.

Thats certainly a viable way around the issue, for me its all about the link to killing. You kill the same amount of people with exterminate, if your a 10 dread, or 10 chivalry general (if I am reading your test correctly) and if you happen to not have the trait triggers you would get the same amount of income.