Information about developments at snowballing problem

Currently viewing this thread:

Apocal

Sergeant Knight at Arms
Best answers
1
I've been hoping for something like this since the game released. Would also be nice to have similar graphs over time for things like: individual village hearths, village production, # of nearby bandit parties, individual town prosperity, food stocks, tariff income, caravan and workshop income, quantities of specific resources in a town's market, etc. All of that would be useful information for the player to have.

Great write-up by the way!
Holy hell, you're back.
 

sniparsexe

Squire
Best answers
0
I've been hoping for something like this since the game released. Would also be nice to have similar graphs over time for things like: individual village hearths, village production, # of nearby bandit parties, individual town prosperity, food stocks, tariff income, caravan and workshop income, quantities of specific resources in a town's market, etc. All of that would be useful information for the player to have.

Great write-up by the way!
Our glorious and mighty lord has returned, bend the knee
 

D0c1

Sergeant
Best answers
0
maybe taleworlds should add mechanics that increases relations between AI lords?
like honor giving realtionship points so honorable lords will like each other, fighting together gives realtions and the resolve dispute quest from warband.
maybe they should remove the relation loss when raiding a village. since, in the eyes of the lords, they're just peasants and are meant to be preyed on. the relation hit with the notables is enough imo.
 

Carrion

Veteran
Best answers
0
Yea I really don't understand this approach. How come they're trying to balance these issues/symptoms isolated from other crucial mechanics that should be present? I really appreicate this kinda post from mexico though, thanks a lot!
 
Last edited:

Apocal

Sergeant Knight at Arms
Best answers
1
Yea I really don't understand this approach. How come they're trying to balance these issues/symptoms isolated from other crucial mechanics that should be present? I really appreicate this kinda post from mexico though, thanks a lot!
Because it is easier to balance one aspect in isolation than multiple. This way, if any new mechanics come along and upset things, you know where the issues probably lie without having to spend time testing every part.

Also, people complain more if you leave obviously unbalanced stuff unaddressed while waiting for a big release.
 

AJIexander

Subforum Moderator
M&BWBWF&SNWVC
Best answers
0
I want to give some information about upcoming developments about snowballing problem. All these developments are already done and will be ready at 1.5.6 beta. Unfortunately next beta 1.5.5 (which will be uploaded next monday or tuesday) will not have these changes.

Firstly lets determine problem :
In Bannerlord some kingdoms (usually Khuzait and Battania) get stronger and stronger each year and one or two factions can easily occupy 50% of map in late gameplay. In first 20 years usually 2 or 3 kingdoms be eliminated (usually N.Empire, S.Empire and Sturgia). It was worse in previous versions (1.5.0 and previous) and even in first 10 years 2 or 3 kingdoms were being eliminated. After developments at war / peace / tribute systems weakened factions can stay 10-15 years more by making more logical peaces and paying tributes to strong enemies and not facing with 2+ enemies mostly. However it was not enough to fix problem completely. Then with help of some of our players we realized problem is also related Khuzait's overpowered culture bonus. These upcoming developments for fixing snowball problem will not change faction bonuses. It is another issue. Even faction bonuses are still same and unbalanced after developments any kingdom cannot occupy 30-40% of map in first 40 years without player interaction.

When Bannerlord first released as EA late game was not well balanced. We and most players usually played first 10-20 years. However we have a feature in Bannerlord which you can continue playing from your adult children. So actually game should be balanced for centuries. We know 90% of players are playing at first 10-20 years but a good sandbox game could save its inside balance for 100+ years. There can be some surprising things like one faction getting half of world of course but they should be rare.

Secondly lets examine our current situation :
Last week several players (@Apocal, @Blood Gryphon) made their own tests to see situation at 1.5.4. They run game 20 years (without player interaction) and give us some tables. You can see these tables as below (here control points mean : 2 for town 1 for castle) :

In my tests I got similar results too. Khuzait and Battania got more than 60% of map at the end of first 20 years and 2-3 factions be eliminated or left with only 1-2 fiefs.

Then you can see 2 problems here. 1-There are so many defections. 2-Minor factions are not joining weakened factions as much as it should be. When a faction lost their territories they should offer higher amounts to mercenary minor factions and recruit them to take lost settlements back. However if weakened factions are having financial difficulties they cannot offer high amounts to mercenary minor faction leaders. Then I made a detailed clan by clan examination to see also financial situation of clans. Because financial difficulties are also one of the biggest reason of defections. Then here is the result :

Clan datas in first 20 years :
Every clan has 4 datas for each year :
1st data is leader gold as 1000 (if <15 it has red font)
2nd/3th datas : towns / castles clan have
4th data : party size usage ratio (how much full their parties are as %, as example if clan can has 3 parties with 100 party size limit but they have 2 parties on map with 75 men this ratio is 50% (75+75 / 100 x 3))) :


As you see some factions have lots of defected clans (background colors are indicating faction clan belong to, change of background color means defection). When we examine factions of these defected clans (w.empire, s.empire, n.empire, sturgia and vlandia) they generally have several clans having financial difficulties. Defections usually happen because of combination of 2 main reasons : 1-Low relation with king 2-Losing all fiefs and having financial difficulties

If you examine above tables you can see there are two problems related to financial situations of clans. First is cumulative total of money in all clans are getting more and more each year. There is a money inflation however even there is a money inflation some clans are so poor and they are having difficulties in even recruiting men to their parties (they have less party usage ratios). Also probably some clans even have problems at feeding their men. Finally these clans defect to another kingdom. Also even some clans have nearly no money some clans have 1M+ money in same kingdom. There is a huge difference between clans as terms of financial situation even inside same kingdom. This happens because of several reasons but main reason is some clans remain without fief / income and there is no mechanism / feature for fief redistirbution. This shows our first and biggest problem. Money should distributed more equally among clans in same kingdom. Then several developments are done to fix this issue. Created a kingdom budget and any clan having 100K+ money start to give 0.5% of their (money - 100K) to kingdom wallet daily. For example if a clan leader has 200K he paid 500 denars daily to kingdom budget (200K - 100K) x 0.5 = 500. This collected money is used to help poor clans in same kingdom if any exists. Each day 500 gold money is given to any clan having less than 30K money, 1000 gold is given to any clan having less than 15K gold. Player is out of these calculations for now. Maybe later we can include player also. We are trying to create more balanced financial situations among clans in same kingdom. Also because there is a high money inflation loot gain of npc lords is decreased. For the ones who does not know they have a different mechanic, they are getting money rather than items when a battle ends. There were no trade penalty applied to these instant loot->money conversations. However as you know there is high trade penalty at weapons / armors, as player if you try to sell or buy same item you can see high price difference. When trade penalty added to this loot->money conversation money inflation is decreased and total clan leader money at mid/late game in all world is decreased. Unfortunately because of reduce in loot income of npc lords number of poor clans are increased first. With the help of kingdom budget addition I mentioned above these poor clans are then reduced. Then we get healthier and more balanced financial situations. Still of course we have poor clans but their number is decreased overall.

To make minor factions to join weaker factions more several formula changes are also applied. Now weak factions with less fiefs / lords / strength offer more money per influence compared to 1.5.4. So minor factions join weak factions more now. After first changes I collected data again to see effects of first developments, this time I made a 43 year test to see bigger picture with more data :

As you see from chart and graph Khuzait still get 40% of world at some point however they lost their areas after. Only 1 faction is eliminated at first 20 years. 2 factions are eliminated at first 43 years also 1 faction remained with only 1 fief. We still have one big problem. Weakened factions cannot make comebacks, you can see this from graph and datas. This make OP factions work easier when conquering half of world. When a faction go below 10 control points (something like 3 towns and 4 castles (3 x 2 + 4 x 1 = 10)) they cannot go up 15 again. This make game also a bit boring because you know that a weakened faction cannot save themselves from that bad situation. This is a bit related to financial problems they are having however when I deeply examined I see they also cannot form even armies. Because influence incomes are nearly all related to settlements they have. When a kingdom lose their settlements they also remain without influence. Then they cannot form even armies. To fix this problem I removed one condition at influence gain from supporter notables. Now supporter notables of a clan give influence to clans even owner of their settlement is another clan. So this is a good passive influence income for NPC clans. Even they lose their settlements they can gain influence from their supporters. After this change weakened factions become more resistant, they started to form armies and started to take get back their lost settlements.

Also some settlements were having prosperity penalties from issues not fixed. When I examine the code I see only npc clan leaders were solving issues with 10% probability when they enter settlements. However some clan leaders were not able to form a party for long times if they were get prisoned and released and if during that time another party is formed from that clan. Then clan leader wait any of these actual parties to be destroyed. During that time some unresolved issues were decreasing prosperity of some settlements to a very low level like 1000-. Especially negative effect of clear bandit hideout issue was very high. Owner of these towns cannot get enough tax and having also financial difficulties. These problems are also fixed. Now all npc lords can resolve issues when they enter a settlement with 5%. Also if a town has very less prosperity like 1000- now they get a small daily prosperity bonus. This helps balancing game a bit.

Then while testing situation after first developments I see another problem which usually weak factions suffer from. Some armies were not buying food even they are starving. After examining this problem I see this was happening when an army is far from all settlements they can buy food from. This problem also fixed. So even an army is far to all neutral towns they stop what they are doing and return one of closest town when they start to have foods problems. Also sometimes I see some lords in weak factions which are forming army even they have so less troops like 10-20 men and success ratio of these armies were low. As you know lords only join armies when they have at least 40% of their party size is full. This condition was not applied to army leaders. Now it is also applied (except player). These small fixes started to make thing better for weaker factions.

One more development is clans do not care only king relation now while defecting they also care how is their relation with other clans inside that kingdom.

After all these developments (actually there are more but I am not mentioning half of developments / bugfixes, next week I can add more changes) I got below results :

As you see from above data, any kingdom did not be eliminated in first 43 years. Also Vlandia goes below 10 control points at some years then they make a come back and reach 20 control points at the end of 43 years test. So there will be more surprise come backs when you play 1.5.6.

If you examine clan tables shared above you can see party usage ratio data (4th data for each clan) is getting less and less every year. Party usage ratio is 100 if all parties in a clan uses all party size limits completely (fe, if clan can have 3 party however if they have 2 party on map with 75% of party size usage clan's party size ratio become 50% because (75+75) / (3 x 100) is 0.5) This ratio starts with 70-80% average for all kingdoms and it become 50-60% average at 20th year of tests. I examined its reason and see that if a lord escape from prison as wounded they are not healed entire game (except player, because player get its party right after escaping). Healing system were only applied to heroes in parties. If a hero is even a prisoner in a dungeon healing is applied but if he escapes when he is wounded he cannot be healthy again. This bug was one reason of this party usage ratio reduce over years. Also new born heroes were not being active when they reach 18 years. They were not forming armies. This bug is also fixed. This problem also effect weak factions more. Because they lost more heroes compared to strong factions and after a point they have so less parties on map. Game should continue with same 70+% party size usage ratio all time. After fixing these bugs this problem is also fixed. You can compare before fixes and after fixes :


As you see after these 2 bug fixes we have higher party size usage ratio averages at late game. In this table there are datas (not per 1 year but per 5 years). I want to spend less time so collected data once per 5 years. Here are control point datas and graph after these bug fixes :

As summary :
There are much more bug fixes and developments to fix this issue but I could give details from some of them. Next week I can give more data and tables. This is a bit huge issue. Now kingdoms are more balanced, there are less poor clans, there are less defections, there is no big money inflation, even weak factions can make come backs and at late game there is not less parties on map. With more balanced kingdoms and even comebacks of weakened factions gameplay will be probably better. Also another team member is working on rebellions currently, we are in coordination with him. When rebellions are added game will be hopefully more balanced, because rebellions will happen in mostly captured towns with low loyality and new rebel clans will want to join previous kingdom (which has same culture with settlement) mostly.

Sorry for my English if I used wrong in some parts, because I am not good at English maybe I selected wrong words in some sentences. Again I need to underline that these developments will take place at 1.5.6 not one next one. Have a good weekend. I will answer all questions and discuss with you at a spare time in weekend. Now need to go dinner.

And here is world situation from final situation :
You are an amazing person and developer. It's hard for me to remember that for all the time, someone talked in such detail about the process of their work and gave so much interesting information with such attention to detail. It's great that you always share information with us and make the Bannerlord development process more transparent. Thanks for your work.
 

Sarejo

Sergeant Knight at Arms
M&BWB
Best answers
0
Fascinating. This weekend I ran a few long term experiments as well, but didn't have the time to examine the results. I did notice the army-food issue, and the settlement issue as well, but did not think it'd affect mid/late-game as much as it does. Looking forward to these changes. Checking the old Warband Beta bugs (http://bugs.taleworlds.com/edit_bug.aspx?id=126), I remember Treason Indictments being a cause for imbalance in mid/late game at the beginning: many lords were landless, and kept changing sides.
 

Carrion

Veteran
Best answers
0
Because it is easier to balance one aspect in isolation than multiple. This way, if any new mechanics come along and upset things, you know where the issues probably lie without having to spend time testing every part.

Also, people complain more if you leave obviously unbalanced stuff unaddressed while waiting for a big release.
I understand what you are saying but I said "crucial mechanics" and not accompanying systems. They should just unleash everything, imho, I am certain the community would help identify problems and potential solutions faster then this piecemeal approach.
 

Kazet

Recruit
Best answers
0
@mexxico It was super interesting to see all your process. I hope we will have more of this in the future !
I would like to see the Kingdom Budget as a feature, with the player taking part in it as you suggested and maybe even with related policies.
(honorable characters or characters with weak friends would want policies increasing Kingdom Budget, while powerful greedy characters won't)
 

Ganeic

Recruit
Best answers
0
Its nice to see that this.

According to my speculations: 4 factions should have the biggest advantage depending on map placement. Sturgia, Vladia, Aserai and Khuzait. Which all start in one 'corner' of the map giving them a huge advantage in terms of not having to usually be threatened from multiple sides at once compared to empire factions that start in the middle for example.

And out of all factions Aseri should have the best position as they literally have 2 bottle necks leading in and out of their realm. even if it can be a dual edged sword at times it usually so far what I've seen it led to them being able to recover quite well unless they face a vastly superior enemy.
 

CrazyElf

Regular
Best answers
0
The current marriages and children being activated at 18 fixed probably extended the amount of nobles still around for his test. But you are right after that first generations kids are all dead its over, so probably around year 60 (~5,000 days) in the game without the player killing anyone. I am definitely interested if his version had marriages in yet


I would like to see this resolved, especially because there is already a pretty decent mod out there that addresses the problem.


Essentially the game is in decline by the third generation because of the lack of children.

We also have problems in relation mechanics. Currently relations between different faction’s clans 95% be worse and worse by time passes because we do not have any mechanics for giving positive relations to npc clans which are in different factions. So by time hostile actions happen and they have more and more enemies, I will focus on this problem in future to make relations more balanced over time.

Currently the only real way to build clan relationships is to fight alongside the clan in battle or voting for them in fiefs, which angers 2 other clans or their laws when they are proposing it, which angers the opponents of the proposal.


I think that there should be a function of what decisions are made by the clans that helps out relationships. Maybe being economically dependent on each other should help.
 

Old-Bull

Sergeant at Arms
Best answers
0
1.5.5 removes autocalc on sieges for Khuzait and gives a -20 percent damage/reload penalty at 0 riding skill that scales up to normal damage/ reload at 100 riding.

Mexxico, I'm assuming the tests in these threads were on an earlier patch.

Do you have any numbers on the 1.5.5 patch because those changes should really help as well.
 

mexxico

Developer
Best answers
0
1.5.5 removes autocalc on sieges for Khuzait and gives a -20 percent damage/reload penalty at 0 riding skill that scales up to normal damage/ reload at 100 riding.

Mexxico, I'm assuming the tests in these threads were on an earlier patch.

Do you have any numbers on the 1.5.5 patch because those changes should really help as well.
I will collect data from 1.5.4, 1.5.5 and 1.5.6 and compare all and will give you information. By the way you can make your own tests after all patches and see situation. @Blood Gryphon can give you information how you can do these tests. Its good this post is pinned. If after 1.5.6 any patch damages world balance and triggers snowballing you can alert me here with your 20 year speed run results.

Lets define how you can degree value of snowballing :
You should run game in 10x for 20 years without player interaction and count fief numbers for each faction, count towns as 2, castles as 1, villages as 0. If number of fiefs in any faction is out of 20-35 range add difference from 20-35 for all factions. Assume khuzait and battania are 60 and 50 then you will get 25 and 15 points for khuzait and battania. If northern empire is eliminated add 20 for northern empire, if southern empire left with 1 castle and 1 town (3 fief points) add 17 for them. Then collect all these values : 25(khuzait) + 15(battania) + 20(n.empire) + 17(s.empire) = 73

If this total is bigger than 60 this means we have a problem. 50-60 is not so good also. 30-50 is ok. Below 30 is good. This total was mostly between 80-140 for 1.5.4 tests.
 
Last edited:

Old-Bull

Sergeant at Arms
Best answers
0
I will collect data from 1.5.4, 1.5.5 and 1.5.6 and compare all and will give you information. By the way you can make your own tests after all patches and see situation. @Blood Gryphon can give you information how you can do these tests. Its good this post is pinned. If after 1.5.6 any patch damages world balance and triggers snowballing you can alert me here with your 20 year speed run results.

Lets define how you can degree value of snowballing :
You should run game in 10x for 20 years without player interaction and count fief numbers for each faction, count towns as 2, castles as 1, villages as 0. If number of fiefs in any faction is out of 20-35 range add difference from 20-35 for all factions. Assume khuzait is 50 then you will get 15 points for khuzait, if northern empire is eliminated add 20 for northern empire, if southern empire is 12 add 8 for them. Then collect all these values : 15 + 20 + 8 = 43

If this total is bigger than 50 this means we have a problem. 40-50 is not so good also. 30-40 is ok. Below 30 is good. This total was about 120-140 for 1.5.4 in average.
Thanks Mexico. From the data you already posted it's apparent that you guys are working hard to fix this.

I'm curious to see how this will work with the changes in 1.5.5 because the loss of autocalc bonus and the damage/speed reduction for lower tier Khuzait troops should have a pretty significant effect on it's own.

Keep up the great work.
 

Flesson19

Sergeant
Best answers
0
@mexxico one thing I noticed you mentioned was running out of money, most of my games the S empire gets rolled over and they all sit in Volstrum so seeing your post this time I talked to them in the keep and they all had 0 denars so they couldn't recruit, and also to this point they were paying tribute, and when I joined them I had 6 million and started paying over 60,000 a day in tribute, now over time that number came down to around 10k so if the tribute is being calculated incorrectly that could help cause why they are running out of money. I went to each of the lords and gave them denars and they started recruiting again but not sure if that is enough to get them going strong again. The Kingdom screen said we should be paying around 4,000 a day so not sure why I had to pay 60,000 msyelf. Let me know if you want my save file when I first joined them so you can see the high tribute being paid.
 

mexxico

Developer
Best answers
0
@mexxico one thing I noticed you mentioned was running out of money, most of my games the S empire gets rolled over and they all sit in Volstrum so seeing your post this time I talked to them in the keep and they all had 0 denars so they couldn't recruit, and also to this point they were paying tribute, and when I joined them I had 6 million and started paying over 60,000 a day in tribute, now over time that number came down to around 10k so if the tribute is being calculated incorrectly that could help cause why they are running out of money. I went to each of the lords and gave them denars and they started recruiting again but not sure if that is enough to get them going strong again. The Kingdom screen said we should be paying around 4,000 a day so not sure why I had to pay 60,000 msyelf. Let me know if you want my save file when I first joined them so you can see the high tribute being paid.
In some very rare cases if any of faction members cannot pay tribute then tribute debt of kingdom is increasing. This is normal because money could not be paid to other kingdoms and this is hold as debt. So if you join a kingdom you should pay this if there is no money in anyone.

However there was a bug in this scenario, in this case everyday all member's debt to kingdom was also increasing. For example if a lord could not pay its tribute share this is also hold at their debtToKingdom variable and whenever they get money this was taken to pay debtToKingdom. Even we increase this debt also Tribute debt of kingdom was increasing too. So assume 5 lords have to pay 1K tribute each. First day any of them cannot pay and tribute debt was 5K. And all their personal debtToKingdom was being 1K. Second day they should pay (5K+5K) / 5 = 2K each. They cannot pay and tribute debt be 10K. Each get 2K as debt to kingdom. Third day 3K. Fourth day 4K. This debt value (debt to kingdom) was even go over 100Ms+ in 1-2 years for all clans in that kingdom. So after this happen this kingdom's lords cannot have money after that. Whenever they get money their money was going to paying debt. It was a bug (wrong calculation of debts) and fixed.