Information about developments at snowballing problem

[Dabos37]

By the way because we solved snowballing problem we can now make imprisonment times longer. With 1.5.9 imprisonment times will be about 1.5-2x. Daily running away chances of hero prisoners will reduce to 5% from 7.5%. This will make gameplay better even its just 2 number change.

Also currently kingdom economies are good compared to previous versions so kingdoms usually do not go bankrupt (not because of any new cheat just they are managing economy better). In future we can easily remove AI cheat of not needing horses to upgrade cavalry troops to make things harder for kingdoms. Recent tests shows 1.5.8 snowballing score average is about 20-30. We can remove some AI cheats so snowballing score can rise to 30-45 in future versions. All snowballing scores < 45 is good. No need to make perfectly balanced.

Maybe I am the only one who thinks in this way but removing the AI cheat for getting cavalry units won’t bring much to improve the game IMO. I would prefer to reduce even more the scape chance or something else. The AI just has access to pretty small cheats while the player has a huge advantage over the AI, so I am ok with keeping these small cheats and do not sacrifice snowballing score for “fixing” something which won’t be much appreciable.

Thanks again for the great work!

 

[Blood Gryphon]

Maybe I am the only one who thinks in this way but removing the AI cheat for getting cavalry units won’t bring much to improve the game IMO. I would prefer to reduce even more the scape chance or something else. The AI just has access to pretty small cheats while the player has a huge advantage over the AI, so I am ok with keeping these small cheats and do not sacrifice snowballing score for “fixing” something which won’t be much appreciable.

Thanks again for the great work!

I agree with this, if taking away the cheat messes up the cav ratio of parties then I don't care for it. It is honestly unnoticeable when playing the game, but a lack of cav would be noticeable and take away from gameplay.

 

[sabatonsky]

Well... it may cause the factions like Khuzaits to have... "different" approach to things, so to say... we will never see, if we dont test

 

[Kut]

Maybe I am the only one who thinks in this way but removing the AI cheat for getting cavalry units won’t bring much to improve the game IMO. I would prefer to reduce even more the scape chance or something else. The AI just has access to pretty small cheats while the player has a huge advantage over the AI, so I am ok with keeping these small cheats and do not sacrifice snowballing score for “fixing” something which won’t be much appreciable.

Thanks again for the great work!

+1

I agree with this, if taking away the cheat messes up the cav ratio of parties then I don't care for it. It is honestly unnoticeable when playing the game, but a lack of cav would be noticeable and take away from gameplay.

Also, +1

 

[Marbles20]

I have a feeling that AI armies will either have to carry more horses with them so they can upgrade as they go or they will have to make more stops into town in order to upgrade. Neither seems like a clear cut winner.

The side effect of having them bring more horses around with them would be that AI lord's armies will move slightly faster if they do gain the infantry on horse bonus. I don't think that would be game breaking, but it could become a lot harder to catch individual lord's parties. One option would just be to deny them that bonus. Also it could impact the player's ability to get warhorses through town trade if they are buying up a lot of horses to replenish their ranks.

The side effect of having lord's stop into town explicitly for horses is that it could create a lot of irregular movement and disrupt the flow of campaigns. Even with a decently high threshold like 'If above 20 soldiers ready to upgrade to cavalry, then stop into nearest friendly town' it seems like it would disrupt a lot of lord's movements, especially late game. It too could disrupt the player's ability to buy warhorses.

 

[Kut]

Without extra precautions, we will probably see major horse shortages in Calradia. Plus, more foot-held armies. That, is a no go situation, imo..

It's not like "devs chaining their sky-net level AI with those cheats" anyways. It's the opposite. For catching up the player's satisfaction, they need those small cheats. Without those extra musters, they can't match players even in sterilised lab environment...

 

[mexxico]

Actually npc parties already buy horses and warhorses if they have less amount of horses and they usually carry 5-10-20 with them. Rich leaders buy more. You just do not see them at loot inventory because nearly 80-90% is assumed runned away during battles. So there will not be big changes if that AI cheat is removed. But I accept it is not big step in gameplay or anything. Just a new money spend area for them. Currently AI kingdoms do not have financial issues so nearly no defections happen. I assume it is now harder for player to find clans which want less than 100K to recruit their kingdom. Removing this cheat will make their economical inputs and outputs same with player except they have no caravans or shops and player do not get help from kingdom wallet. If this cheat is removed mostly Khuzait will be effected economically because their cav ratio is 40% while others change between 8-18%. However currently Khuzait is richest faction by far because they have higher cav ratio and they are fast and they mostly win battles because they join ones which they can win.

Anyway I accept no big deal so I skip this for now. Later I will make some calculations and if it will not change cav ratios much I can remove this cheat.

 

[Blood Gryphon]

Actually npc parties already buy horses and warhorses if they have less amount of horses and they usually carry 5-10-20 with them. Rich leaders buy more. You just do not see them at loot inventory because nearly 80-90% is assumed runned away during battles. So there will not be big changes if that AI cheat is removed. But I accept it is not big step in gameplay or anything. Just a new money spend area for them. Currently AI kingdoms do not have financial issues so nearly no defections happen. I assume it is now harder for player to find clans which want less than 100K to recruit their kingdom. Removing this cheat will make their economical inputs and outputs same with player except they have no caravans or shops and player do not get help from kingdom wallet. If this cheat is removed mostly Khuzait will be effected economically because their cav ratio is 40% while others change between 8-18%. However currently Khuzait is richest faction by far because they have higher cav ratio and they are fast and they mostly win battles because they join ones which they can win.

Anyway I accept no big deal so I skip this for now. Later I will make some calculations and if it will not change cav ratios much I can remove this cheat.

Sounds good, I do appreciate the general idea that the player and AI should play on even grounds. I was waiting for this hotfix to start a new kingdom playthrough, so I'll let you know how recruiting turns out. Cost was rarely an issue for me in the past due to how easy it is to collect millions by the time I start a kingdom. My main issue here still remains chasing around lords to recruit, need messenger mechanic to help with that. Sorry for bringing it up again, I know you planned on bringing it up later.

Just curious, is player clan party control still in your hands or did that get moved? This is the next thing I'm interested in seeing even with the watered down version that was approved.

 

[Kut]

Actually npc parties already buy horses and warhorses if they have less amount of horses and they usually carry 5-10-20 with them. Rich leaders buy more. You just do not see them at loot inventory because nearly 80-90% is assumed runned away during battles. So there will not be big changes if that AI cheat is removed. But I accept it is not big step in gameplay or anything. Just a new money spend area for them. Currently AI kingdoms do not have financial issues so nearly no defections happen. I assume it is now harder for player to find clans which want less than 100K to recruit their kingdom. Removing this cheat will make their economical inputs and outputs same with player except they have no caravans or shops and player do not get help from kingdom wallet. If this cheat is removed mostly Khuzait will be effected economically because their cav ratio is 40% while others change between 8-18%. However currently Khuzait is richest faction by far because they have higher cav ratio and they are fast and they mostly win battles because they join ones which they can win.

Anyway I accept no big deal so I skip this for now. Later I will make some calculations and if it will not change cav ratios much I can remove this cheat.

Understood.

Having buying is one thing, having buying and systematically expending (troop ugrade) is another thing. But you guys can add extra horse villages or adding few horse producement in every village, maybe this will cut it..

But then, horse prices will go down too. Esp. Aserai ones. Traders love those horses.. I don't know how it can effect the general economy/horse trading. But you're more knowlodgeable & experienced than me in this area, of course...

 

[AJAJP_Juan]

Cost was rarely an issue for me in the past due to how easy it is to collect millions by the time I start a kingdom.

Off-topic, but millions? Beating up all lords I can concentrating on maximum profits like selling prisoners I barely make the 50k-100k a week.

 

[Blood Gryphon]

Off-topic, but millions? Beating up all lords I can concentrating on maximum profits like selling prisoners I barely make the 50k-100k a week.

Millions was gratuitous by me, but 1 to 2 mil by clan tier 5 (when I typically start a kingdom). I basically have all companions run a caravan and set up as many workshops as I can in the first few years and remain a vassal for quite some time (typically vlandian so I can take over sturgian land). Ensured I received towns instead of castles. Also always keeping my party costs reasonable with lower tiers instead of high tiers (Tier 4 max in most cases). Lots of fighting like you said to sell prisoners, loot and keep my companion parties in the green. I don't smith cuz of obvious cheese.

To be fair though this was pre-snowball fixes so I'm not sure how this will impact things.

 

[LuciusDomitiusAurelianus]

By the way because we solved snowballing problem we can now make imprisonment times longer. With 1.5.9 imprisonment times will be about 1.5-2x. Daily running away chances of hero prisoners will reduce to 5% from 7.5%. This will make gameplay better even its just 2 number change.

Also currently kingdom economies are good compared to previous versions so kingdoms usually do not go bankrupt (not because of any new cheat just they are managing economy better). In future we can easily remove AI cheat of not needing horses to upgrade cavalry troops to make things harder for kingdoms. Recent tests shows 1.5.8 snowballing score average is about 20-30. We can remove some AI cheats so snowballing score can rise to 30-45 in future versions. All snowballing scores < 45 is good. No need to make perfectly balanced.

If this is done it would be good to have benchmarks done to see what the overall cav mix of parties are and for horse production to increase, especially if lords will now buy more of them. It's pretty hard to keep a stock of war horses, especially in empire, you need them to get the noble lines upgrade and for bucellarii. If it won't drastically change composition of cav ratios then I would support the AI being put on par with the player, but it definitely seems like low-hanging fruit.

It would make sense if it was taken up in parallel with troop tree refreshes proposed by @Terco_Viejo - selfishly I would really like to see Equites moved into the main empire troop tree as dedicated non-noble light cavalry and start the noble line from a heavy horseman, it's annoying to need 2 horses across 2 types for the noble line. It would also help Emprie get the amount of cav to compete with Aserai and Khuzait - in all of the snowball benchmarking you can see S. Empire especially always ends up with the lowest amount of fiefs.

 

[Marbles20]

Through skills it should be lower. I think a 5% daily default rate is good.

The things to keep in mind though are:
- There is (or will be) the 'break lord out of prison' quests/opportunities and if the chance to escape on their own is too high there will be no point to implementing prison breaks. By the time you see the person was imprisoned, talk with one of their clan members, then find and approach the castle/town they are held in they will have already escaped.
- The skills are on the very high end of the skill trees (escape from party -50% in scouting is a 225 skill, hero escape -30% is 200 in rogue (a pain to level)). The scouting one isn't hard if you have a companion scout as that levels fast, but roguery atm feels like a waste and you wouldn't want to spend your own points seriously leveling it unless you were roleplaying a bandit lord.
- The 5% is also just if they are in a secure dungeon too. A 10% base minimum escape chance in the field, though much better than the 22.5% base minimum currently, is still really high.

That's why I was curious if this was an incremental adjustment toward something much smaller like 2%, which would come out to 4-10% chance to escape in the field based on the 2x to 5x.

 

[Blood Gryphon]

2nd test on the hot fix, looks really good again.

Test 7 1.5.8.hot	Test 7 10 yrs	Test 7 20 yrs	20 yr Strength	Original Clans	Recruited Clans	Rebel Clans	Mercenaries	Total Noble Clans	Total Clans w/ MERCs
Aserai	28​	23​	6492​	9​	0​	0​	1​	9​	10​
Battania	20​	19​	4573​	8​	0​	0​	0​	8​	8​
Khuzait	24​	17​	5121​	9​	0​	0​	1​	9​	10​
Northern	13​	13​	3783​	9​	0​	0​	2​	9​	11​
Southern	20​	28​	5550​	9​	0​	0​	0​	9​	9​
Western	23​	33​	7347​	9​	0​	0​	0​	9​	9​
Sturgia	20​	19​	3473​	9​	0​	0​	2​	9​	11​
Vlandia	25​	21​	4975​	11​	0​	0​	1​	11​	12​
Snowball score	7​	12​

隐藏: Test 7 10 yrs

隐藏: Test 7 20 yrs

 

[Nalgasucia507]

The things to keep in mind though are:
- There is (or will be) the 'break lord out of prison' quests/opportunities and if the chance to escape on their own is too high there will be no point to implementing prison breaks. By the time you see the person was imprisoned, talk with one of their clan members, then find and approach the castle/town they are held in they will have already escaped.
- The skills are on the very high end of the skill trees (escape from party -50% in scouting is a 225 skill, hero escape -30% is 200 in rogue (a pain to level)). The scouting one isn't hard if you have a companion scout as that levels fast, but roguery atm feels like a waste and you wouldn't want to spend your own points seriously leveling it unless you were roleplaying a bandit lord.
- The 5% is also just if they are in a secure dungeon too. A 10% base minimum escape chance in the field, though much better than the 22.5% base minimum currently, is still really high.

That's why I was curious if this was an incremental adjustment toward something much smaller like 2%, which would come out to 4-10% chance to escape in the field based on the 2x to 5x.

True that. I like your arguments.

 

[mexxico]

The things to keep in mind though are:
- There is (or will be) the 'break lord out of prison' quests/opportunities and if the chance to escape on their own is too high there will be no point to implementing prison breaks. By the time you see the person was imprisoned, talk with one of their clan members, then find and approach the castle/town they are held in they will have already escaped.
- The skills are on the very high end of the skill trees (escape from party -50% in scouting is a 225 skill, hero escape -30% is 200 in rogue (a pain to level)). The scouting one isn't hard if you have a companion scout as that levels fast, but roguery atm feels like a waste and you wouldn't want to spend your own points seriously leveling it unless you were roleplaying a bandit lord.
- The 5% is also just if they are in a secure dungeon too. A 10% base minimum escape chance in the field, though much better than the 22.5% base minimum currently, is still really high.

That's why I was curious if this was an incremental adjustment toward something much smaller like 2%, which would come out to 4-10% chance to escape in the field based on the 2x to 5x.

Yes you are right. I will think about reducing more. Lets go step by step. First lets try 5% (settlement) / 10-25% (field) instead of 7.5% (settlement) / 22.5-45% (field). And I will collect your feedbacks. If no problem happens and if players are happy with this reduce (probably they will be) later I will lower it to 2.5% (settlement) / 5-12.5% (field). We have no snowballing problem anymore so I am more comfortable about reducing this probability.

Thanks for lots of tests @Blood Gryphon much appreciated! If possible can you also place map screenshot from your last test.

Thanks for ideas @LuciusDomitiusAurelianus, I also prefer 1 more cav troop in empire troop tree too but I do not think it will be accepted. I can suggest it. By the way enabling all castle village rural notables giving noble troops (even their power is lower than 200) can help Empire getting higher cav ratio and can make castles more important. Tomorrow I will give more information about this idea with all datas.

 

[Marbles20]

cool, sounds good.

 

[Blood Gryphon]

Thanks for lots of tests @Blood Gryphon much appreciated! If possible can you also place map screenshot from your last test.

Forgot to do it for that test ?, added it to the post. Plan to get 3 more over the next few days so we have 5 from me for both pre and post hotfix.

 

[LuciusDomitiusAurelianus]

Yes you are right. I will think about reducing more. Lets go step by step. First lets try 5% (settlement) / 10-25% (field) instead of 7.5% (settlement) / 22.5-45% (field). And I will collect your feedbacks. If no problem happens and if players are happy with this reduce (probably they will be) later I will lower it to 2.5% (settlement) / 5-12.5% (field). We have no snowballing problem anymore so I am more comfortable about reducing this probability.

Thanks for lots of tests @Blood Gryphon much appreciated! If possible can you also place map screenshot from your last test.

Thanks for ideas @LuciusDomitiusAurelianus, I also prefer 1 more cav troop in empire troop tree too but I do not think it will be accepted. I can suggest it. By the way enabling all castle village rural notables giving noble troops (even their power is lower than 200) can help Empire getting higher cav ratio and can make castles more important. Tomorrow I will give more information about this idea with all datas.

Thank you for the response and the suggestions @mexxico. I do agree with the other points regarding prisoner escape being lowered to 2.5% in settlement, this will help make sure that lords are likely to be captive in cities for Prison Break quests. Will the AI also participate in these?

Hopefully more cav in Empire will be considered, but I do like that Castles will be given more importance and more chance to have Cav. Would a suggestion to start the Noble Line at Equite (removing Vigla Recruit) be more likely to be considered? This will at least help Empire Lords not needing both a regular horse and a warhorse per Vigla Recruit. I think giving Empire more Cav will help with snowballing as well, even if only from Notables.

 

[Blood Gryphon]

Test 8

Test 8 1.5.8.hot	Test 8 10 yrs	Test 8 20 yrs	20 yr Strength	Original Clans	Recruited Clans	Rebel Clans	Mercenaries	Total Noble Clans	Total Clans w/ MERCs
Aserai	26​	24​	7945​	9​	0​	0​	0​	9​	9​
Battania	15​	6​	5225​	8​	0​	0​	3​	11​	11​
Khuzait	27​	22​	6104​	9​	0​	0​	0​	9​	9​
Northern	22​	14​	5505​	9​	0​	0​	1​	9​	10​
Southern	16​	15​	4502​	9​	0​	0​	2​	9​	11​
Western	18​	34​	8914​	9​	0​	0​	0​	9​	9​
Sturgia	21​	22​	6739​	9​	0​	0​	0​	9​	9​
Vlandia	28​	36​	8159​	11​	0​	0​	0​	11​	11​
Snowball score	11​	26​

隐藏: Test 8 10 yrs

隐藏: Test 8 20 yrs

I checked rebels again for this test and test 7 and found the same low count of rebels.

隐藏: Test 7 rebels

隐藏: Test 8 rebels