Information about developments at snowballing problem

Users who are viewing this thread

on a side note, reducing war declaration chance to happen less often would also be pretty welcome. I am even in favor of some long peace times (50 days or so), but maybe some players could complain about this.

Yes game should not be a war simulator. We should make peace times longer reduce avarage wars per year and introduce player different activities he can do during peace time like feasts or grand torunaments and we should increase issue variety which some can be hard to achieve and has good awards.

I would reduce the chance por decision trigger and not the clans war support, so in this way the player would be still able to declare new wars whenever he wants.

Will think this. Already when player enters a kingdom this probability reduces. You suggest to reduce more then ?
 
Last edited:
Yes game should not be a war simulator. We should make peace times longer reduce avarage wars per year and introduce player different activities he can do during peace time like feasts or grand torunaments and we should increase issue variety which some can be hard to achieve and has good awards.



Will think this. Already when player enters a kingdom this probability reduces. You suggest to reduce more then ?

Not exactly. I am suggesting to reduce the probability for new wars for everyone. I am especially interested to see a less aggressive AI when AI vs AI is involved.

Main problem with snowballing is that the player feels that he has to rush money, tier clan and influence in order to bring some kind of balance to the map (or at least I feel in this way). A less aggressive AI when the player is not involved would be great IMO.

On the other hands, this could bring other issues like some kingdoms getting tons of elite units while in peace (not sure if it would be the case after the perks and troops XP nerf). If it is something easy to change in DDLs, I could do it and run 10 tests or so with the changes to safe you time. Thanks!
 
Not exactly. I am suggesting to reduce the probability for new wars for everyone. I am especially interested to see a less aggressive AI when AI vs AI is involved.

Main problem with snowballing is that the player feels that he has to rush money, tier clan and influence in order to bring some kind of balance to the map (or at least I feel in this way). A less aggressive AI when the player is not involved would be great IMO.

On the other hands, this could bring other issues like some kingdoms getting tons of elite units while in peace (not sure if it would be the case after the perks and troops XP nerf). If it is something easy to change in DDLs, I could do it and run 10 tests or so with the changes to safe you time. Thanks!

We have one more problem by the way. Longer peace time means poorer clans currently. Because 35% of clan income is wars and 65% of clan income is taxes currently (as average). So if AI vs AI wars decreases this will increase poor clans especially if a kingdom lost half of their territory and if they have no enemy they will have big problems. Loot from looters do not bring that much gold. You can suggest making this ratio 20% loot / 80% taxes by increasing taxes and decreasing loots but this can make some players unhappy. I already reduced loots a bit in 1.5.7. Maybe can reduce more later and increase taxes a bit. However this time still kingdoms lost territories suffer much because losing territory means lower tax income also.
 
Last edited:
We have one more problem by the way. Longer peace time means poorer clans currently. Because 35% of clan income is wars and 65% of clan income is taxes currently. So if AI vs AI wars decreases this will increase poor clans especially if a kingdom lost halşf of theişr territory and if they have no enemy they will have big problems. Loot from looters do not bring that much gold. You can suggest making this ratio 20% loot / 80% taxes by increasing taxes and decreasing loots but this can make some players unhappy. I already reduced loots a bit in 1.5.7. Maybe can reduce more later and increase taxes a bit. However this time still kingdoms lost territories suffer much.

Yes, poorer clans is something that we have talked before related to peace time. Increasing fiefs income is something that would be pretty welcome for most of players IMO. On the other hand, it could be a problem for fiefless clans having to defect faster. Anyway, I would really like to run some tests with this change and post the results in the forum and maybe it could be helpfully.

Before I was able to reduce this random chance for peace/war declarations pretty easy:

@Dabos37 you could also try narrowing the range that the randomFloat variable can be to meet the conditions to consider war. Currently, the range is between 0.3 and 0.7. If you change those numbers to say, 0.5 to 0.6, then there should be ~4x fewer clans considering war each day.

Daily-Tick-Clan-Kingdom-Decision.png

Buw I am seeing now that this code has changed and I am not sure how to modify this:

private KingdomDecision GetRandomWarDecision(Clan clan)
{
KingdomDecision result = null;
Kingdom kingdom = clan.Kingdom;
if (kingdom.UnresolvedDecisions.FirstOrDefault((KingdomDecision x) => x is DeclareWarDecision) != null)
{
return null;
}
Kingdom randomElement = (from x in Kingdom.All
where x != kingdom && !x.IsAtWarWith(kingdom) && x.GetStanceWith(kingdom).PeaceDeclarationDate.ElapsedDaysUntilNow > 20f
select x).ToArray<Kingdom>().GetRandomElement<Kingdom>();
if (randomElement != null && this.ConsiderWar(clan, kingdom, randomElement))
{
result = new DeclareWarDecision(clan, randomElement);
}
return result;
 
@mexxico

Uhmm, something is weird here. I am not seeing exactly what you have posted, maybe because our version is a bit different than yours:


As you can see, I have modified to pretty low values (I started using your suggested numbers but anything changed), and still I have seen 5 new wars being declared in less than 30 days as you can check in this short video (using multiply x20):



I have deleted all files from dnsSpy and re-upload them to check that the values are actually modified.
 
@mexxico

Uhmm, something is weird here. I am not seeing exactly what you have posted, maybe because our version is a bit different than yours:


As you can see, I have modified to pretty low values (I started using your suggested numbers but anything changed), and still I have seen 5 new wars being declared in less than 30 days as you can check in this short video (using multiply x20):



I have deleted all files from dnsSpy and re-upload them to check that the values are actually modified.


I do not know reason, maybe you can try disabling that line (by using * 1, *1.5, * 2) completely and you can see if they are triggered somewhere else.

If it completely disables declerations then you can leave decleration so small probability like using *1.0x, *1.5x, *2.0x (x can be 1-2-3-4-5)
 
Yes game should not be a war simulator. We should make peace times longer reduce average wars per year and introduce player different activities he can do during peace time like feasts or grand tournaments and we should increase issue variety which some can be hard to achieve and has good awards.
Thing is there is no mobilisation/demobilisation mechanic.
AI should reduce size of its armies in peace time to reduce costs but should have ability to somewhat store excess units somewhere.
Maybe adding ability to store units in villages? And in all fiefs(Castles/villages/towns) ability to store some units for free(zero upkeep)?
And reduce garrison upkeep for the rest of the garrison?
Maybe reduce upkeep for units on friendly territory and increase upkeep when in enemy one?
Maybe increase amount of mercenaries available in Inns(instead of one group in small number make it 3-4 groups with bigger numbers) so AI(and Player) is able rapidly increase number of troops under command but at the cost(with these troops being first to go when peace happen)
 
@mexxico what would the downsides be of letting AI cheat to stop from going broke? Will it make it too hard for player to recruit AI lords?

I’m starting to think this may be an acceptable solution. I’ll be honest the wealth of enemy clans is not really something I think about while playing and I only see the effect when they are so poor they can’t field parties. If they never went bankrupt no one would notice that you had changed anything, since we have no indication of their wealth anyways. If this make gameplay bad for the player give us a cheat that makes it easier to recruit fiefless wealthy clans (maybe scale difficulty with total clans in player kingdom).

If the merc bug fix doesn’t make a big impact, I would test this out.
 
Last edited:
I do not know reason, maybe you can try disabling that line (by using * 1, *1.5, * 2) completely and you can see if they are triggered somewhere else.

If it completely disables declerations then you can leave decleration so small probability like using *1.0x, *1.5x, *2.0x (x can be 1-2-3-4-5)

Ok I think that maybe the random number is sometimes too low and for this reason the condition < happens and new wars are getting declared. I have modified the values to this:


And not new wars have been declared at day 50. Some peace declarations are still happens though, probably for this line:

if (randomFloat < num2 && clan.Influence > (float)diplomacyModel.GetInfluenceCostOfProposingPeace())

Going to continue making some tests with these numbers to try to find a good value for me. Thanks!
 
Thing is there is no mobilisation/demobilisation mechanic.
AI should reduce size of its armies in peace time to reduce costs but should have ability to somewhat store excess units somewhere.
Maybe adding ability to store units in villages? And in all fiefs(Castles/villages/towns) ability to store some units for free(zero upkeep)?
And reduce garrison upkeep for the rest of the garrison?
Maybe reduce upkeep for units on friendly territory and increase upkeep when in enemy one?
Maybe increase amount of mercenaries available in Inns(instead of one group in small number make it 3-4 groups with bigger numbers) so AI(and Player) is able rapidly increase number of troops under command but at the cost(with these troops being first to go when peace happen)
This sounds quite reasonable in my opinion.
 
We have one more problem by the way. Longer peace time means poorer clans currently. Because 35% of clan income is wars and 65% of clan income is taxes currently (as average). So if AI vs AI wars decreases this will increase poor clans especially if a kingdom lost half of their territory and if they have no enemy they will have big problems. Loot from looters do not bring that much gold. You can suggest making this ratio 20% loot / 80% taxes by increasing taxes and decreasing loots but this can make some players unhappy. I already reduced loots a bit in 1.5.7. Maybe can reduce more later and increase taxes a bit. However this time still kingdoms lost territories suffer much because losing territory means lower tax income also.
I think loot being a smaller share of income is a good thing in general. If you wanted to cheat it you could simply make it so AI pay less wages durring peace time.

Or the Lords collect a portion of town caravan icome during peace given caravans do better in peace? Just a spit ball if you want to do something inside the economy rather than my suggested cheat.

Also
Tests 1.5.7
FactionClans YR0Fief Score YR0Clans YR20Fief Score YR20
Aserai9251122
Battania8181144
Khuzait9211239
N Empire9211016
S Empire92230
Sturgia922810
Vlandia112486
W Empire9201536

Below is some observation of Clans movements at end game
Aserai + 2 clans : 1 vland 1 emp
Battan +3 clans : 2 emp 1 sturgia
Khuzait +3 clans: 2 vland 1 vland rebel
N empire +1 clan: 1 vland rebel
S empire -6 clans
Sturgia -3 clans +2 sturgia rebel
Vlandia -3
W empire +6 2 Sturgia 3 empire 1 emp rebel

5 Total Rebel clans consolidated

Broad observations: Clan defections didnt seem to be the deciding factor or further nudge in snowballing any given conflict. The majority of defections occur when a faction is majority wiped out. These defections would however make a comeback more difficult, but they dont appear to add a huge amount of momnetum in face of what is established by the already rolling snowball. The deffections just secure the decline of already pummelled factions.

The wars in this one were quite fun to watch. Sturgia pushed into the North Empire early then got double teamed and swept out. Khuzaits swooping like vultures. Battania and vlandia traded sargot 6 or so times then vlandia actually won out and pushed into battania taking 2 of their towns, Vlandia then got on a multifront war and ended up losing to Battania despite almost having them on the ropes. Western empire went no where in a war with asarai early, sat at peice for a while, then found its success fighting enemies that were already at war. Securing territory in the north and south that had been heavily traded.
 
Last edited:
I think loot being a smaller share of income is a good thing in general. If you wanted to cheat it you could simply make it so AI pay less wages durring peace time.

Or the Lords collect a portion of town caravan icome during peace given caravans do better in peace? Just a spit ball if you want to do something inside the economy rather than my suggested cheat.

Also
Tests 1.5.7
FactionClans YR0Fief Score YR0Clans YR20Fief Score YR20
Aserai9251122
Battania8181144
Khuzait9211239
N Empire9211016
S Empire92230
Sturgia922810
Vlandia112486
W Empire9201536

Below is some observation of Clans movements at end game
Aserai + 2 clans : 1 vland 1 emp
Battan +3 clans : 2 emp 1 sturgia
Khuzait +3 clans: 2 vland 1 vland rebel
N empire +1 clan: 1 vland rebel
S empire -6 clans
Sturgia -3 clans +2 sturgia rebel
Vlandia -3
W empire +6 2 Sturgia 3 empire 1 emp rebel

5 Total Rebel clans consolidated

Broad observations: Clan defections didnt seem to be the deciding factor or further nudge in snowballing any given conflict. The majority of defections occur when a faction is majority wiped out. These defections would however make a comeback more difficult, but they dont appear to add a huge amount of momnetum in face of what is established by the already rolling snowball. The deffections just secure the decline of already pummelled factions.

The wars in this one were quite fun to watch. Sturgia pushed into the North Empire early then got double teamed and swept out. Khuzaits swooping like vultures. Battania and vlandia traded sargot 6 or so times then vlandia actually won out and pushed into battania taking 2 of their towns, Vlandia then got on a multifront war and ended up losing to Battania despite almost having them on the ropes. Western empire went no where in a war with asarai early, sat at peice for a while, then found its success fighting enemies that were already at war. Securing territory in the north and south that had been heavily traded.
Great observations! I think its clear that defections are what stop a faction comeback, but is not the reason they got weak in the first place.

The behavior to gang up on weaker factions is something I've noticed too, normally its the khuzait doing it to wild success instead of the western empire. Thats a hard one because the AI is just noticing that they have an opportunity to win and get good tributes, but what I think we really want is for them to ignore those situation UNLESS its a stronger faction and your advantage has come from you not being as weak. So the kingdoms that should be getting two wars declared on them are the stronger factions not the weaker factions (although this is not necessarily the logical choice for an easy win). Even without alliances if this was flipped on its head I think it would have our desired behavior and kingdoms wouldn't get trampled at the first sign of weakness
 
Last edited:
Guys I do not want to give anyone false hopes but this is how the map look at day 300 in my test campaign with modified values (1.12, 1.35, 1.75 - I just put some random numbers here decreasing what we had before):


Wars are still being declared each 30-40 days or so in every kingdom, and most of kingdoms are still usually at war but just against one enemy. Maybe a lucky run but it is looking pretty promising. I will provide more information later, and perform a lot of tests this weekend about this.
 
Yeah but they escape with how many men?
most of their troops. ~75%.
Will these mean that there is never deciding battle?
sieges will be the most important and kind of deciding because the garrison can't rout.
it can be done so when a lord dies or gets captures, his troops become deserters/bandits or there can be a chance for them to join a garrison if a friendly fief is nearby. so there will be a chance for a faction to lose much power in a single battle.
When defeated defender would respawn with free troops and could mount counterattack
currently, lords get ~20-25% of their party filled for free (or do they get 20-25 troops? can't remember which). the rest they have to go about and recruit like the player.
if they retained most of their troops, it'd be better for them than the free troops.
 
Guys I do not want to give anyone false hopes but this is how the map look at day 300 in my test campaign with modified values (1.12, 1.35, 1.75 - I just put some random numbers here decreasing what we had before):


Wars are still being declared each 30-40 days or so in every kingdom, and most of kingdoms are still usually at war but just against one enemy. Maybe a lucky run but it is looking pretty promising. I will provide more information later, and perform a lot of tests this weekend about this.
I've been noticing the first 10 years are much much better, so might be a bit too early to call. Give us another update around 600-800 days please :grin:
 
I've been noticing the first 10 years are much much better, so might be a bit too early to call. Give us another update around 600-800 days please :grin:

Not in my experience, Khuzaits are always able to take Onira or Amprela pretty soon, because NE and SE get tons of wars pretty early. I have to stop now to make some job stuff and will continue this afternoon. Going to try to run at least two 20 years runs. And I am going to record a video in the second run to let you see war/peace declarations timing (I doubt that someone would be interested on see that 1 or 2 hours video but who knows xD).
 
Not in my experience, Khuzaits are always able to take Onira or Amprela pretty soon, because NE and SE get tons of wars pretty early. I have to stop now to make some work stuff and will continue this afternoon. Going to try to run at least two 20 years runs. And I am going to record a video in the second run to let you see war/peace declarations timing (I doubt that someone would be interested on see that 1 or 2 hours video but who knows xD).
In my games Khuzait take Amprela, Myzea and Argoron lose Argoron to a rebellion lose Baltakhand retake Baltakhand and the do their usual thing but a bit slower
 
Back
Top Bottom