Information about developments at snowballing problem

Users who are viewing this thread

Dabos37

Sergeant Knight at Arms
@mexxico By the way, in terms of relevance, a kingdom having too much fiefs compared to clan numbers, has more weight than kingdom strength? I mean, let’s suppose that Khuzaits have a pretty favorable clans:fiefs ratio, something like 9:40, but this kingdom is also pretty strong, how this affects the war declaration chance?

I think that clans:fiefs ratio should be the most relevant thing and the kingdoms with too much fiefs should get a lot of hate from others.
 

mexxico

Sergeant Knight
@mexxico By the way, in terms of relevance, a kingdom having too much fiefs compared to clan numbers, has more weight than kingdom strength? I mean, let’s suppose that Khuzaits have a pretty favorable clans:fiefs ratio, something like 9:40, but this kingdom is also pretty strong, how this affects the war declaration chance?

I think that clans:fiefs ratio should be the most relevant thing and the kingdoms with too much fiefs should get a lot of hate from others.

(total clan tiers (except minors) x strengthPerTier + total strength of clans (including minors)) / total value of fiefs (this was only prosperity, problem i meant)

strengthPerTier = a value which slowly changes according to number of fiefs (fief count make it min 0.5x (if there are so less fiefs - due to recruitment problem), max x (normal case 90% this is used))

this is the basic of formula

of course I tried to simplfy 500 lines of code, there are tons of effects but this is main.

Main was only

total strength of clans (including minors)) / total value of fiefs (still value was prosperity which is problem today I realized, when I see value as variable name I did not checked where it come today I see its only prosperity :sad:)

+ 500 lines of side effects

before 1.5.x so this was also increasing number of war peace declerations because strength is something changing too much during time but number of clans is not. I made that formula like seen at above to make things more stable and realistic but because of adding trying to decrease number of wars faced for each faction at 1.5.1 war peace declerations are increased even this formula change. Because when a faction decleare war to other faction that faction try to get rid of one of existing wars. So there become too much war - peace declerations.
 
Last edited:

Dabos37

Sergeant Knight at Arms
Ok ok, I think that I got it. As smaller the formula result, bigger chance for war declaration right?

The mentioned change related to do not take into account just properity is really good, because besieged and conquered fiefs have usually lower prosperity and this is making some big kingdoms go under the radar while getting more and more fiefs.

Thank you for the hard work and explanation.
 

Blood Gryphon

Master Knight
WBVC
I do not know if I can add this fix to 1.5.7 or not because its testes are started.
Can't just slip this one in? :wink:

giphy.gif


Speaking of War/Peace declarations I don't think I ever posted the results from our 1.5.6 testing publicly. Y'all will notice that some changes in 1.5.6 originally had messed up war/peace declarations but Mex was able to fix it and get us back to a decent median war length before the patch actually dropped. Also how crazy is it that wars could last up to 4 years back in 1.5.0 :lol:

So yeah the data shows us that Mex was able to increase the average (median & mean) in 1.5.6 over 1.5.1 (last time there were changes here). He was able to decrease wars under a season (less than 21 days) from ~53% to ~30% of wars, increased the wars between one and two seasons (21 -42 days) from 16% to 43%, and kept wars 2 seasons and longer around the same at 31% and 27%.

It is good that we are now talking about why they are making these declarations.

s2Iwh.png


tBS13.png


IoM1c.png
 
Last edited:
Today during examining war peace codes I see that factions evaluate enemy settlements by only their prosperity. Basics of these codes are not mine I am just going over them since 1.5.1 (https://forums.taleworlds.com/index.php?threads/diplomacy-developments.429850/) to make war / peace declerations more logical which also effects snowballing positively. I see too many weird things since going over these codes and fixed some. This only caring prosperity as a value measurement is another problem which causes factions to decleare more wars towards Empire factions because of their high prosperity towns compared others. Okay prosperity is something important but caring only prosperity without any addition make things worse. Enemy settlements should be counted as X + prosperity instead of directly using prosperity otherwise if a faction have 6 towns with 4000 prosperity has a value of 24000 while Khuzaits have 6 towns with 3000 prosperity has a value of 18000. So Khuzaits attract other facitons less and they are attacked less similarly Aserai / Sturgia too also castles are nearly not counted because their maximum prosperity is mostly 1000. If a faction has bigger prosperity / strength ratio generally they are attacked more. I will make this calculation (prosperity * 0.25 + (town -> 2000 , castle -> 1000)) per each fief. I do not know if I can add this fix to 1.5.7 or not because its testes are started.
Yeah thanks, that is some Ingo👌👌
 

lunelune2413

Recruit
M&BWBWF&SNWVC
Today during examining war peace codes I see that factions evaluate enemy settlements by only their prosperity. Basics of these codes are not mine I am just going over them since 1.5.1 (https://forums.taleworlds.com/index.php?threads/diplomacy-developments.429850/) to make war / peace declerations more logical which also effects snowballing positively. I see too many weird things since going over these codes and fixed some. This only caring prosperity as a value measurement is another problem which causes factions to decleare more wars towards Empire factions because of their high prosperity towns compared others. Okay prosperity is something important but caring only prosperity without any addition make things worse. Enemy settlements should be counted as X + prosperity instead of directly using prosperity otherwise if a faction have 6 towns with 4000 prosperity has a value of 24000 while Khuzaits have 6 towns with 3000 prosperity has a value of 18000. So Khuzaits attract other facitons less and they are attacked less similarly Aserai / Sturgia too also castles are nearly not counted because their maximum prosperity is mostly 1000. If a faction has bigger prosperity / strength ratio generally they are attacked more. I will make this calculation (prosperity * 0.25 + (town -> 2000 , castle -> 1000)) per each fief. I do not know if I can add this fix to 1.5.7 or not because its testes are started.
Sounds amazing. It would be a shame if it's not included in 1.5.7, then 1.5.8/1.6.0 at the very least :smile:
 

Sarissofoi

Veteran
M&BWBWF&S
This is all tiresome.
Does anyone supervise the code in TW or its like a pot where anyone can throw anything?
What is the point of even doing balancing in current state when most components are probably don't even work properly?
I appreciate Mexxico fixing stuff but it look like he is only one trying to put down fire using bucket when rest of brigade slack off.
 
Last edited:

sniparsexe

Knight at Arms
Can't just slip this one in? :wink:

giphy.gif


Speaking of War/Peace declarations I don't think I ever posted the results from our 1.5.6 testing publicly. Y'all will notice that some changes in 1.5.6 originally had messed up war/peace declarations but Mex was able to fix it and get us back to a decent median war length before the patch actually dropped. Also how crazy is it that wars could last up to 4 years back in 1.5.0 :lol:

So yeah the data shows us that Mex was able to increase the average (median & mean) in 1.5.6 over 1.5.1 (last time there were changes here). He was able to decrease wars under a season (less than 21 days) from ~53% to ~30% of wars, increased the wars between one and two seasons (21 -42 days) from 16% to 43%, and kept wars 2 seasons and longer around the same at 31% and 27%.

It is good that we are now talking about why they are making these declarations.

s2Iwh.png


tBS13.png


IoM1c.png
this stuff is really good, great job Gryphon!
 

Dabos37

Sergeant Knight at Arms
Can't just slip this one in? :wink:

giphy.gif


Speaking of War/Peace declarations I don't think I ever posted the results from our 1.5.6 testing publicly. Y'all will notice that some changes in 1.5.6 originally had messed up war/peace declarations but Mex was able to fix it and get us back to a decent median war length before the patch actually dropped. Also how crazy is it that wars could last up to 4 years back in 1.5.0 :lol:

So yeah the data shows us that Mex was able to increase the average (median & mean) in 1.5.6 over 1.5.1 (last time there were changes here). He was able to decrease wars under a season (less than 21 days) from ~53% to ~30% of wars, increased the wars between one and two seasons (21 -42 days) from 16% to 43%, and kept wars 2 seasons and longer around the same at 31% and 27%.

It is good that we are now talking about why they are making these declarations.

s2Iwh.png


tBS13.png


IoM1c.png



Great info mate. I am just confused about this:

“Y'all will notice that some changes in 1.5.6 originally had messed up war/peace declarations but Mex was able to fix it and get us back to a decent median war length before the patch actually dropped”

As far I know, we will get the changes for increasing war length in 1.5.7, and 1.5.6 is the same as it was at release (24th December), or maybe I am missing something here?
 

Blood Gryphon

Master Knight
WBVC
Great info mate. I am just confused about this:

“Y'all will notice that some changes in 1.5.6 originally had messed up war/peace declarations but Mex was able to fix it and get us back to a decent median war length before the patch actually dropped”

As far I know, we will get the changes for increasing war length in 1.5.7, and 1.5.6 is the same as it was at release (24th December), or maybe I am missing something here?
Remember that 1.5.6 spent a long time in testing, so long that mex was able to send me data to analyze and make the necessary adjustments (77% of wars were less than a season) before 1.5.6 actually dropped on the 26th.

He has now made even more improvements so 1.5.7 should be even better, although I haven’t looked at any data yet.
How you even get you hands on this data?
Sadly we don’t have an easy way to collect this data, mexxico was the one to provide it.

Great stuff as always @Blood Gryphon, thank you for all these tables.
this stuff is really good, great job Gryphon!
giphy.gif
 
Last edited:

mexxico

Sergeant Knight
Actually what @Blood Gryphon posted as 1.5.6 data changed before 1.5.6 is released and war times get shorter again, some latest additions to solve other problems effected them badly. Probably his 1.5.6 results will be similar to 1.5.7. Thats why I could not posted these tables. These were results we get before 1.5.6 is released. We made some work with him but I could not find time to collect data and give him after latest additions. So 1.5.6 median war durations are shorter than table’s data.

When 1.5.7 is out we can make new measurements.
 
Last edited:

Blood Gryphon

Master Knight
WBVC
Actually what @Blood Gryphon posted as 1.5.6 data changed before 1.5.6 is released and war times get shorter again, some latest additions to solve other problems effected them badly. Probably his 1.5.6 results will be similar to 1.5.7. Thats why I could not posted these tables. These were results we get before 1.5.6 is released. When 1.5.7 is out we can make new measurements.
That makes more sense. I'll mark it up accordingly in a bit
 

Dabos37

Sergeant Knight at Arms
Yes, that makes more sense to me because I am noticing shorter wars in 1.5.6 as discussed days ago. Nice that war length is getting increased again in 1.5.7.

Thanks for the info.

BTW, reading again some older posts, I have noticed that the change for making more units dying in simulated battles (instead of getting injured) is not included in this list:


Here all test results we have since 1.5.4 :

eJfMz.png

According to these results we have 117 snowball score at 1.5.4, 104 at 1.5.5, 61 at 1.5.6 and I think 1.5.7 will be around 45 with new developments mentioned in last posts :

1-Reducing cavalry speed bonus to 40% from 60%. I did not get approve for this yet because I fear it to be rejected but I will apply it. Because it seems this is a must. I do not want to risk this.
2-There will be a bit higher horseman troop ratio in 7 non-Khuzait factions. Parties will tend to upgrade their troops not 50% - 50% randomly if they have 2 options they will upgrade according to troop tree to reach horseman troops more if their cavalry ratio is low.
3-Rebel clans will join weaker factions & same culture ones more instead of joining powerfull kingdoms if they succeed surviving first 30 days.
4-Reducing effects of OP 2 perks giving too much passive xp and being a side effect in povetry. Also needed xp for tier-3, tier-4 and tier-5 are a bit increased. So it will be a bit hard to reach high tiers with only passive xp.
5-30% longer preliminary phase at sieges.
6-Heart values of Khuzait villages will be a bit reduced.
7-Khuzait default policy grazing rights policy will be +0.5 loyality -0.25 hearth instead of +1 loyality +1 militia.
8-Weak kingdoms will try to face with less enemies and pay more tributes if needed. Agressiveness of kingdoms which lost lots of territory will be reduced. This is partly 1.5.1 feature but its effect will be a bit increased.

And this change will probably also help a bit with snowballing.
 
Last edited:

froggyluv

Grandmaster Knight
NW
This is all tiresome.
Does anyone supervise the code in TW or its like a pot where anyone can throw anything?
What is the point of even doing balancing in current state when most components are probably don't even work properly?
I appreciate Mexxico fixing stuff but it look like he is only one trying to put down fire using bucket when rest of brigade slack off.

Yeah on one hand its nice to see a community come to the aid of its game development but in another its pretty worrisome that Taleworlds seems dependent on it or else we are stuck with what we were given. What if the 3-5 guys from this thread just decided to stop or had other life pressing needs? Is the current game we have in hand the one Armagon envisioned and is satisfied with?
 

sniparsexe

Knight at Arms
Yeah on one hand its nice to see a community come to the aid of its game development but in another its pretty worrisome that Taleworlds seems dependent on it or else we are stuck with what we were given. What if the 3-5 guys from this thread just decided to stop or had other life pressing needs? Is the current game we have in hand the one Armagon envisioned and is satisfied with?
I can only think of 2 reasons for this stuff (mexxico needs help from players to test, which I am and probably others testers are completely fine with, so no offense against mexxico) is, TWs QA is incapable (or unwilling) to test things. Or simply, TW leads are like "let's finish this mass ASAP and move on to other projects, so don't test things too much and just throw them at players", or again simply, both.
 

Dabos37

Sergeant Knight at Arms
Developers getting feedback from players is something pretty usual and relevant for improving games balancing. If the 3-5 guys in this thread decide to stop, then other 3-5 guys will come and help devs giving feedback. I do not think that the community is coming to aid anything, the community is just complaining about issues and giving feedback for helping developers to fix these issues faster. This is usual for most of games where developers care about community and want to improve/fix existent issues. Same happens for Total War games, they even hired a guy from the community to improve balancing in Warhammer Total War.

Anyway, sorry for the off topic and let’s keep focusing on snowballing issue.
 
Last edited:

froggyluv

Grandmaster Knight
NW
Developers getting feedback from players is something pretty usual and relevant for improving games balancing. If the 3-5 guys in this thread decide to stop, then other 3-5 guys will come and help devs giving feedback. I do not think that the community is coming to aid anything, the community is just complaining about issues and giving feedback for helping developers to fix these issues faster. This is usual for most of games where developers care about community and want to improve/fix existent issues. Same happens for Total War games, they even hired a guy from the community to improve balancing in Warhammer Total War.

Anyway, sorry for the off topic and let’s keep focusing on snowballing issue.

It should be as an auxiliary role - not for primary design which without would be an incomplete system. Taleworlds needs to build a cohesive complete game first.
 

Sarissofoi

Veteran
M&BWBWF&S
irMlklK.png

10 tests done with changed(asymmetric troop trees and modified party templates)
most factions have only one branch with 5th tier common troops(HA for Aserai, Cavalry for Vladians, Skirmishers(foot and mounted for Battanians) Khuzait and Empire have none)
after 4 first tests it was adjusted to give back Battania some of its cavalry(in templates), taken 2nd tier of Khuzaits cavalry(turned it into foot archers), and give larger ratio to Vladians for cavalry(which was mistake probably and need to be adjusted(by lowering it in their templates)
 
Top Bottom