Event troops function based on pre-2.0 levy values

[unmerged(133356)]

Overall, I have to say that I absolutely love this thread.

 

[ChildeR]

It matters a tremendous deal to pay attention to the leger when it provides data accurate values for event spawn troops, while the military page does not, and both calculate their numbers differently

Sure, of course. But I'm still not seeing where you are seeing the new (large) 2.0 modifiers not being used for event troops. If they were missed, the event troops in large enough realms should be up to around double the size that match_mult implies, but in practice they are withing ~15%. Clearly (?) the 50% LIEGE_LEVY_SIZE_MULTIPLIER must have been taken into account, or else it alone would cause a larger discrepancy than I see.

I guess I'll have to go look at the ledger as you do, but it's such a pain, since the interesting cases (i.e. large realms with non de jure vassals) will have a long list of numbers to add up.

 

[Zireael]

Why do the ledger and the military tab differ? This very fact is an oversight and should be fixed ASAP.

 

[Bortron]

Why do the ledger and the military tab differ? This very fact is an oversight and should be fixed ASAP.

Add it to the list. This latest patch was not ready for release. Besides the problems Mann42 has pointed out it is also full of bugs.

 

[c_none]

Thank you so much Mann42 and everyone who contributed to this post.

...
This results in some fun side effects. For example, if everyone in your country hates you so much that they would never dream of giving you levies but your crown law is low, it actually reduces the number of event troops so significantly you could probably handle them with just your direct levies. Essentially, since the game factors opinion and laws into the event troop calculation, but ignores liege levy penalties, you are essentially making it harder to fend off event spawns by maintaining good opinion with your vassals or raising crown law - effectively punishing proper feudal play even more.
...

Thank you so much especially for this pointer. Yesterday my Ironman broke down the moment event troops spawned after raising the CA to High with my ruler who enjoyed wonderful relations with all vassals

Interestingly, this is what revolt spawn troops do with scaled_by_biggest_garrison, to an extent, by taking the Total Levy + Garrison Value of your largest holding in a province, multiplying it by 3, and then occasionally scaling it up by 33%.

While I'm not going to defend the concept of troops magically spawning from thin air, I do think there's valid reasons to scale a threat to the player's capability, but I do agree that if the match_mult doesn't take liege levy penalties into account, it shouldn't take opinion, law, and demesne bonuses into account either. The main problem I see is that, by factoring bonuses in while not factoring negatives, they're effectively encouraging you to play either sub-optimally or so optimally you're exploiting and missing the point of the game.

From a practical perspective, it makes it an absolute nightmare to balance without introducing those undesired side effects, and from a selfish perspective, it makes it an absolute nightmare to attempt to mod the values to something that makes sense.

I'm all for adding challenge, but if Paradox wants to increase challenge in revolts/event troop spawns, they should do so in a way that encourages people to play the game 'properly' as a feudal simulator. Right now, the calculations active work to negate all your advantages and penalize you twice for properly managing your vassals, once by reducing your levies, and then again by boosting event spawns without consideration of your levies.

I'd like them to fix their calculations such that it encourages people to play in the way the developers intended (fuedally), and I'd love if discussions of difficulty centered on how to properly scale and present these larger number of event troops instead of a stalwart defense of a bug or oversight on the side of the developer.

I want Crusader Kings to be an awesome game. I make games for a living, but even though Paradox isn't paying me a cent, I enjoy their game enough to want to see it become better.

I'm not doing this work to make the game easier to 'beat'. I'm working to fix a developer oversight that introduces tons of unintended negative side effects. I'm doing this work in the hopes of making it easier to balance in the future for both developers and modders, easier to discuss as a community when rationally brainstorming the difficulty curve, and easier to play the way the I believe the developers intended. The current systems and event troops calculation are actively working against all 3 of those purposes.

PI should print this out and paste it on the walls of each cubical in their offices!

I'm so glad to let you guys know that apparently PI is working on it: http://forum.paradoxplaza.com/forum/showthread.php?737108-quot-Large-bands-or-rebels-quot-event-with-latest-patch. While I would appreciate a quick-fix so I can go back to playing vanilla, I'm OK with it taking a while this time so that we don't encounter more unexpected behavior.

 

[redsimonDE]

I'm so glad to let you guys know that apparently PI is working on it: http://forum.paradoxplaza.com/forum/showthread.php?737108-quot-Large-bands-or-rebels-quot-event-with-latest-patch. While I would appreciate a quick-fix so I can go back to playing vanilla, I'm OK with it taking a while this time so that we don't encounter more unexpected behavior.

Haha, and the first guy answering in the linked thread was already one of these "It's not a bug, it's a feature that is supposed to make the game more challenging" people.

 

[ChildeR]

Out of curiosity I had a look at how the math worked prior to 2.0 and... it's not pretty either.

Again, this is three match_mult = 0.1 armies. Event armies are 4110 men, realm levies over 55k. I.e. they didn't break the math in 2.0, it was already broken. It simply swung from being someting like 20% lower to being something like 20% higher.

That means if they actually fix the math, but don't touch the match_mult values, event troops will be pretty close to halfway as powerful as they were in 1.111 and 2.0. Everyone is happy?

 

[Bortron]

Out of curiosity I had a look at how the math worked prior to 2.0 and... it's not pretty either.

Again, this is three match_mult = 0.1 armies. Event armies are 4110 men, realm levies over 55k. I.e. they didn't break the math in 2.0, it was already broken. It simply swung from being someting like 20% lower to being something like 20% higher.

That means if they actually fix the math, but don't touch the match_mult values, event troops will be pretty close to halfway as powerful as they were in 1.111 and 2.0. Everyone is happy?

If you're right that might solve the problem.

 

[Mann42]

Out of curiosity I had a look at how the math worked prior to 2.0 and... it's not pretty either.

Again, this is three match_mult = 0.1 armies. Event armies are 4110 men, realm levies over 55k. I.e. they didn't break the math in 2.0, it was already broken. It simply swung from being someting like 20% lower to being something like 20% higher.

That means if they actually fix the math, but don't touch the match_mult values, event troops will be pretty close to halfway as powerful as they were in 1.111 and 2.0. Everyone is happy?

Oh wow. Great catch. Now I feel like I need to pop open my 1.11 branch and see what I see there.

I'm finding that my calculation for the military screen are wrong. I can consistently get close to the military page values using my formula, but it's not exact, and as I said, I'm not happy with close. I would have posted a follow up last night but the numbers were elluding me. Maybe looking back at 1.11 will help me find a common thread.

It is obvious that there's a massive discrepancy between the leger and the military screen. I guess the real question is, which screen is wrong. (And for my own self edification, how in the holy hell are levy values on the military screen actually derived?!?)

 

[ChildeR]

It is obvious that there's a massive discrepancy between the leger and the military screen. I guess the real question is, which screen is wrong. (And for my own self edification, how in the holy hell are levy values on the military screen actually derived?!?)

The military screen is "correct" in the sense that it shows how many troops will pop up if you press the button. That said, I wouldn't be surprised if the numbers there are not exactly what the tooltips and defines suggest, either.

 

[Mann42]

The military screen is "correct" in the sense that it shows how many troops will pop up if you press the button. That said, I wouldn't be surprised if the numbers there are not exactly what the tooltips and defines suggest, either.

That is absolutely true - what that screen shows is what you get when you raise levies.

My issue now is that none of the old calculations, or current assumed calculations for determining levies, actually work to calculate what you see on that screen and what you actually raise.

For a single count vassal, I have to assume that all their holdings and the holdings of their sub-vassals (mayors and bishops) are added together to get the military page total. However, , I've tried deriving from Leger values, from Total Levy values, I've factored in (or out) all known levy modifiers, I've tried using my count's vassal's opinion for sub-vassal liege levies, I've tried using the sub-vassal opinion of me for liege levies, I've tried factoring with and without the old 0.5 modifier, I've tried factoring with or without the liege levy penalties as seen in defines, I've tried factoring my count's levy laws when trying to derive sub-vassal levies... I can get within a reliable +/-5% margin of error, but I'm currently stumped on exactly which modifiers apply where and how to derive actual liege levies as derived on the military page and raised.

I'm going to keep trying to figure it out though.

In the meantime, what is confirmed is:

1. Event troop scaling from match_mult is based on leger values.
2. Leger values don't take liege levy penalties into account.
3. Leger values are significantly different from the military page values.
4. There were discrepancies between leger and military page values back in 1.11, except that it instead it caused event troop stacks to be proportionally smaller instead of larger like they are in 2.0.

 

[ChildeR]

Looks like the math wasn't "fixed" in 2.0.2. The match_mult numbers were simply scaled back. Between 20% and 33%, from what I see. That probably means event troops are generally more powerful than 1.111 for large realms, but significantly less powerful than 2.0 for all.

Edit: Prepared invasions were about halved, as were some starting troops.

 

[BadDecisionDino]

Maybe it wasn't fixed, but it's enough for me to relish the fact that people were wrong on the internet.

Still, it's a multiplier, so it'll take some testing to see how it ripples out. I imagine they'll say that the growing disparity for larger empires is Working As Designed, which I can live with, but it highlights just how bananas North Korea mode is, as the One True Way to rule large territories.

 

[Assalander]

So what about the beta patch which supposedly fixes this?

(sad that they didn't consider balancing retinues and mercenaries to the new levies as well... apparently they want everyone to play as a merchant republic or EU4...)

 

[ChildeR]

So what about the beta patch which supposedly fixes this?

If you look two posts above yours, you find that it does "fix" it, except the math still doesn't match.

 

[Assalander]

If you look two posts above yours, you find that it does "fix" it, except the math still doesn't match.

Oh yes sorry haven't slept in 2 days :/