Game engine has problems with calculus

[jpd]

After 6 months of intense Hoi 2 beta testing, I started a fresh CK game today. It goes reasonably well, until I got an event to hire mercenaries. Seeing that I had the cash for it, I decided to the bold move of hiring them.

Well, they are parked on my capital of Nassau for about a year, when my ally, France, declares war on one of the minors on the channel coast, and pulls me into a war with the King of England.

Well, a couple of months into the game, I get a visit from a couple of english armies. I use the mercenary armyand sure enough, the first opponent is eradicated. The second army is more troublesome. My mercenary army (1600 men, of which 400 are knights) are defeated by a peasant army half the size. What the f*ck, I think, and mobilize my regular army. Those beat the english into submission (as it should ), and I demob afterwards, keeping only the mercenary army, now reduced to 1468 in strength.

Or so it seemed. The game nicely reports 1468 men in the army, but when I click on the army to view it's regiment, all components (knights, archers, peasants, cavalry, etc) are zero. Yup, you read that correct. 1468 men had instantly turned into ghosts.

So I save the game to use wordpad to figure out what's going on. I find the mercenary army as a unit with one division (I like the HoI keywords in a medieval game ), and sure enough, the numbers are there in the save game, and add up nicely to 1468.

So, info is correct in the save game but not in the running game itself. Well, I surrender to reload from this save game. Sure enough, the mercenary unit is now back to the strength it should have, but now the capital is in another place. I go bonkers here. Back to wordpad to checkup on it. And sure enough, the save game still lists province 121 (Nassau) as capital. So why does the game loader interpret 121 as 253, which is where my capital appears after reloading is complete?

I thought computers were good with numbers until now, but this game sure as hell makes me doubt the accuracy of both my processor and my brain.

Jan Peter

PS, playing with the april 8 beta.

 

[unmerged(21937)]

For the ghosts, I have no idea.

For the capital relocation, game checks where your capital should be always when you load up a save, so as Nassau is not your richest province, it will relocate, no matter what it says in the save.

 

[Grosshaus]

I´ve never seen that kind of behaviour with the armies, might it be something that happens only to mercenaries? Would be nice to get more feedback.

The capital thing is indeed WAD.

 

[jpd]

Hmm. when that capital thingy is considered WAD, why then does it only move to a different location upon save game reload.

In other words, if I had never found a reason to force a save/reload cycle, then my capital would have stayed in Nassau (#121) indefinately, irrespective of whether ot not it's the richest province (however that may be defined).

Since the tech research is impacted significantly upon where the capital is (it moved to the province with the least tech spread, grrrrrrr), I found myself researching the same stuff all over again which I already had acquired in my official capital.

In other words, if it should be WAD that the capital moves to the 'richest' province, why then does it not do so during the game itself?

Jan Peter

 

[unmerged(2037)]

jpd: I remmember the 'ghost' behavior of armies in early versions, not sure about current. But the cause was IMO that the component numbers show the number of each troop type able to fight (BUG or WAD?). So if the regiment morale is zero or bellow the component numbers are zero and the regiment will break as soon as it enters combat.

 

[unmerged(21937)]

Hmm. when that capital thingy is considered WAD, why then does it only move to a different location upon save game reload.

This observation is actually wrong. I've seen capital change in following situations:

A) Loading up the game
B) When doing peace
C) When dying and being inherited
D) When giving titles to people
E) When current capital is lost

So, four out of five happen during gameplay.

 

[jpd]

Lol. I was referring to this in context of the reason which was stated why it moved upon reload. And precisely that reason is not a trigger to move the capital in-game.

So, why should it be WAD when it moves to the richest province (by whatever measurement that is derived) upon reload, while at the same time it's not a reason to move it in-game whenever the conditions change that define the richest province.

My definition of WAD is that it should
1) Either move both upon reload and in-game when the conditions determining the richest province change.
2) Or don't move it when conditions change.

As it works now, only players stupid enough (or being forced to) to save and reload get confronted with a sudden shift of the capital province, and thus upsetting the list of techs 'known in our realm'.

As it works now, it's like the old HoI 1 'research cost calculus bug', when upon reload the cost of research programs were recalculated without regard to decimal precision (thus always rounding down), while changes to research programs in-game were calculated with decimal precision. Thus, after a reload, a player could suddenly do more research programs for the same amount of IC invested.

A WAD save/reload cycle, IMHO, is only WAD when after the reload the saved condition is restored. I don't consider a reload WAD when reloading produces a different situation then existed on the moment of the save. And moving the capital definately qualifies for this.

Jan Peter

 

[unmerged(6777)]

I think the limited checks on criteria to move a capital are a concession to the processor demands that more prequent checking would entail. IIRC it should move:

• Upon game load
• January 1st of each year (but maybe this has a bug?)
• Upon loss of demesne capital to a non-coreligionist (ie you actually don't have a demesne capital any more)
• Upon agreeing to a peace (which is the other thing that would normally have a large impact on change of realm richest province) or upon making any other action that changes the physical make-up of the character's demesne (granting titles, etc.)
• Upon realm-ruler dying (maybe the son is tired of daddy's run-down ol' palace?) when inheritance laws will cause a check of all states

Checking it more frequently would be nice but I think the issue is that it would cause a noticable bump/delay when it checks thus the more often it checks it, the more annoying it would be.

I'll ask Johan to look into the ghost army issue.

 

[jpd]

Well, regarding processor capacity.

I just took a shot of my own medicine (http://forum.paradoxplaza.com/forum/showthread.php?t=147392) and got rid of the tiles.bmp file. This makes the game much, much more responsive.

Just a shame that CK does not come with a default province color set for terrain type (like EU2, Vicky, HoI, HoI2) and all provinces are now white-ish. But hey, I rather have them all a shade of white than a sluggish mouse pointer

Jan Peter

 

[Caranorn]

Just on the army I have seen the same phenomenon. I think it's indeed the 0 or worse morale. I expect you let your mercenaries fight out that battle (no wnder they lost with an anonymous leader) and they routed from the field. When you next checked them their morale must still have been 0 or less (lucky they didn't decide to run home alltogether as tends to happen in enemy provinces).

And yes, capitals occasionally move during game play as intended, but I don't seem to recall a precise january 1 move (so that check might indeed be missing).

 

[jpd]

Just on the army I have seen the same phenomenon. I think it's indeed the 0 or worse morale. I expect you let your mercenaries fight out that battle (no wnder they lost with an anonymous leader) and they routed from the field. When you next checked them their morale must still have been 0 or less (lucky they didn't decide to run home alltogether as tends to happen in enemy provinces).

Hmm. What you really are saying is that I've uncovered a way to cheat. Because, if what you say is true, then the morale of the troops doesn't get saved in the save game, and thus is reset to max upon a reload.

Jan Peter

 

[Caranorn]

Hmm. What you really are saying is that I've uncovered a way to cheat. Because, if what you say is true, then the morale of the troops doesn't get saved in the save game, and thus is reset to max upon a reload.

Jan Peter

Not necessarily, some armies I have now found seem to remain routed eternally even after their morale has recovered. I mean by that that during any battle after their first route they will appear with the routed arrow and effectively don't participate in the battle regardless of their actual morale. The only way I have found to work around this is to reload the game (or disband the army). Maybe these two phenomenon are related.

The very fact that the problem is solved by reloading from save also means there is no way I can think of to replicate this for the folk at Paradox to test.

 

[unmerged(28030)]

I've noticed that when I hire mercenaries, I can keep them around as long as I want, but if I've had them a few years, they just disappear when I send them into battle.