Historical Variability

[KalZakath]

Would it be feasible to have a slider to set the variability in the events that occur during the game, especially for the non-player countries? Would this be something that could be modded in?

In the WWW, even though they wanted to play 'historically' (cough, Anglo-Franco, cough) there seemed to be an awful lot of events firing that were very non-historical. Alf Landon in the US, Trotsky in the SU, even Poland giving up Danzig when Johan didn't guarantee beforehand. For the last one, they actually reverted to a saved game to make sure it fired.

I'm thinking of a slider where on one extreme would be a very very historical game, with little chance of these ahistorical events firing, and the other extreme would be a much higher percent that the ahistorical ones fire? Sort of like on one extreme, there's less than a 1% chance of Alf Landon winning, and the other extreme, it might be 25% (just a figure pulled from thin air)... That on one end, there's less than a 1% chance of the Trotsky event happening to the SU, and on the other it might be 25% (again, thin air), etc.

This would allow a very historical game for those who wanted one, but also allow for a lot of monkey wrenches to be set in place if the player(s) so wanted?

Not sure how many things are triggered by choices made by the players (i.e. Johan's choice of waiting to guarantee Poland), and how many of them are just plain random, but might be a way to satisfy those who want a historical gameplay and those who want a non-scripted game as well.

 

[hkrommel]

In the WWW, even though they wanted to play 'historically' (cough, Anglo-Franco, cough) there seemed to be an awful lot of events firing that were very non-historical. Alf Landon in the US, Trotsky in the SU, even Poland giving up Danzig when Johan didn't guarantee beforehand. For the last one, they actually reverted to a saved game to make sure it fired.

Alpha. This will almost certainly be fixed by release.

 

[Bernard Black]

The historical option is supposed to force the Non-Player-Countries along the historical focus tree, nothing more. The game can not go historical, if the players go ahistorical. For instance if a player germany doesn't honour the MR-pact the SU should not stand idly by until they get attacked or in the WWW case if GB doesn't guarantee Poland its perfectly sensible for them to give up Danzig. If you force the AI down a historical path you need to limit the player to a historical one as well.

But the moment you unpause at the start of the game you go 'ahistorical' as you don't make the same decisions as in history. The whole game is set out to simulate the WW2 setting and letting the player experiment and play with different approaches. In a fully historical game Germany should always loose.

That beeing said, you should be more clear what you actually want, when you talk about a more historical gameplay. What is this slider supposed to do? Coding a specific sequence of National Focuses is pretty easy everything else after that gets difficult.

Regards

 

[KalZakath]

That beeing said, you should be more clear what you actually want, when you talk about a more historical gameplay. What is this slider supposed to do? Coding a specific sequence of National Focuses is pretty easy everything else after that gets difficult.

Regards

Thinking more about the events that fire during the game, rather than the players actually playing. i.e. Single player playing Germany - What is the chance during the game that Alf Landon wins. What is the chance that the Trotsky event happens as it did in the WWW.

Looking for whether that variability - the chance that these events have to fire that are ahistorical - can be set before the game starts by a player. For example, as a single player playing Germany:

A very historical setting might have these events happen less than 1% of the time (i.e. Landon beats FDR <1% of the time)
A somewhat historical setting might have these events happen 3% of the time. (i.e Landon beats FDR 3% of the time)
A quasi-historical (just adding another level) might have these happen 5% of the time (i.e. Landon beats FDR 5% of the time)
A non-historical might have these events happen 10% of the time (i.e. Landon beats FDR 10% of the time)
A total non-historical might have these events happen 20% of the time (i.e. Landon beats FDR 20% of the time)

This isn't for the events that I as Germany control - it's for the seemingly random stuff that happens around the world that what I do as Germany has very little control over. Yes, if I want to go historical, I can build up to and trigger Danzig or war on Aug 31, 1939 and away I go. I know I have the power to control that as Germany, but I wouldn't be able to, as Germany, trigger the Landon Victory in 36?

Again - not sure of the mechanics behind a lot of the events, but it does really seem as though there's a RNG factor on many of them. What I'm looking for is how to tweak that RNG to trigger more ahistorical events if I should so want to, that I wouldn't be in any control over.

 

[Bernard Black]

Well in theory that should be possible but it seems to be quite a bit of work.

One would need to go through all historic events and disable all the random trigger conditions on the ones deemed ahistorical. More elegantly you could define a global variable 'historicity' and let the player choose a value between 0.01 and 1.99, that would basicly be you slider.
The code for the random trigger conditions would have to be adjusted to something like this (just the rough idea):

Begin
Historic Event.
Conditions: - ...
Event_Chance:= Event_Chance *historicity

Effects:....

End

You would still need to go through all the possible events with some Randomness involved and chang the code. The question is, if that would make sense in all cases. Its also not clear which events have random components. I also don't know how this could interfere with other stuff in the code and at which point the events are compiled. So in short: Its not too difficult in theory, implementation would probably quite tedious and would need a lot of playtesting and could mess up the other stuff.

It doesn't seem to be out of scope for a mod and seems one of the thinks the community should take care of and the devs should focus on bigger issues.

Regards

 

[KalZakath]

Thanks Bernard - was assuming a mod if it wasn't something quick and easy (like there wasn't already a variable in place that could just be tweaked).

 

[TheCrimsonMajor]

I think this is a nice idea. Similar to how AI aggressiveness worked in HoI2. If a slider is too difficult, I have a few ideas for different modes they could implement.

Normal Mode (Historical Outcome Mode)

• Game Balance: tweaked to favor historical outcome
• AI behavior: strong preference for historical outcome

This mode would match the standard HoI2 experience. In Normal Mode, the game would be balanced to ensure that the war turns out more or less according to history more often than not. This means that certain nations (such as France) would be weaker in game than they were in reality, and the AI may be hard-coded to make decisions that a rational player never would make (like Japan halting their advance in China) in order to ensure a historical outcome. Without player interference, all of the major events of World War II should happen more or less on time. Where there are deviations from history, they are usually not massive. The percentage chance of the AI choosing a non-historical route in national focuses, elections, strategic priorities, etc. would be low (maybe between 5-15%). However, the AI would also offer limited reactions to non-historical changes made by the player.

Historical Start Mode

• Game Balance: attempts to model each nation's capabilities accurately
• AI behavior: preference for historical outcome, but potential for branching is greater

The balance in this mode would try to accurately reflect the historical capabilities and potential of all the major powers in 1936. In addition, AI behavior and decisions would comport with each nation's historical priorities and ideology, but the AI would be much more reactive and potentially exploitative of non-historical opportunities. This means that the outcome of the Battle of France (if it happens in 1940) would not be a foregone conclusion, and you could see WWII starting earlier or later than September 1939 much more frequently.

Sandbox Mode

• Game Balance: attempts to model each nation's capabilities accurately
• AI behavior: no hard-coded preference

The AI will pursue national focus, diplomacy, and conquest goals in a more or less random manner. Any hard-coded blocks against certain outcomes will be disabled (like no fascist USA in 1936, or certain nationalities being predisposed to a particular ideology). Play this mode for a potentially wildly ahistorical experience. That being said, the AI should still behave rationally, and should still stand by allies it makes and behave in a somewhat plausible manner.


Obviously, multiple modes seems very unlikely at this point, but who knows? They had difficulty and aggressiveness sliders in the past...

 

[Bernard Black]

That is actually much more workintensive than the slider idea, where one would just scale the randomness of certain events.

The problem is that your description is very vague in respect of how to implement that.
In your first mode:
How would you nerf France specifcly?
What do you consider a non massive deviation from history?
How would you make it so that the major events in the game start at their historical date with their historical outcome.

I understand roughly what you mean, but thats not enough for actually coding these modes. You need to make a lot of decisions that can be pinned on the numbers and outcomes you want.
They can not be easily systemized in contrast to the slider idea.

This idea needs to have a lot more thought put into it on how to make it work.

Regards

 

[FOARP]

Thinking more about the events that fire during the game, rather than the players actually playing. i.e. Single player playing Germany - What is the chance during the game that Alf Landon wins. What is the chance that the Trotsky event happens as it did in the WWW.

Looking for whether that variability - the chance that these events have to fire that are ahistorical - can be set before the game starts by a player. For example, as a single player playing Germany:

A very historical setting might have these events happen less than 1% of the time (i.e. Landon beats FDR <1% of the time)
A somewhat historical setting might have these events happen 3% of the time. (i.e Landon beats FDR 3% of the time)
A quasi-historical (just adding another level) might have these happen 5% of the time (i.e. Landon beats FDR 5% of the time)
A non-historical might have these events happen 10% of the time (i.e. Landon beats FDR 10% of the time)
A total non-historical might have these events happen 20% of the time (i.e. Landon beats FDR 20% of the time)

This isn't for the events that I as Germany control - it's for the seemingly random stuff that happens around the world that what I do as Germany has very little control over. Yes, if I want to go historical, I can build up to and trigger Danzig or war on Aug 31, 1939 and away I go. I know I have the power to control that as Germany, but I wouldn't be able to, as Germany, trigger the Landon Victory in 36?

Again - not sure of the mechanics behind a lot of the events, but it does really seem as though there's a RNG factor on many of them. What I'm looking for is how to tweak that RNG to trigger more ahistorical events if I should so want to, that I wouldn't be in any control over.

You're assuming that these events stand in isolation, and are firing according to a monthly chance of firing and nothing else. We don't know how HOI4's events work, but in a well-design system they would not stand in isolation (since this would lead to events occurring that make no sense given the prevailing in-game circumstances) but instead fire according to prevailing in-game circumstances.

Take the Trotsky event, for example - it seems likely that this is going to be a result of not carrying out the purge, for example. And why wouldn't you carry out the purge? Because you fear an early war perhaps, something triggered by world tension?

Something like this, which it would be very good to have in-game, cannot be implemented as a slider because it does not rely on random chance but instead on prevailing global factors.

 

[TheCrimsonMajor]

That is actually much more workintensive than the slider idea, where one would just scale the randomness of certain events.

The problem is that your description is very vague in respect of how to implement that.
.....
I understand roughly what you mean, but thats not enough for actually coding these modes. You need to make a lot of decisions that can be pinned on the numbers and outcomes you want.
They can not be easily systemized in contrast to the slider idea.

This idea needs to have a lot more thought put into it on how to make it work.

Regards

Thanks for your thoughts Bernard. I don't pretend to have any sort of workable plan, and I agree that these ideas definitely need more iteration.

In your first mode:
How would you nerf France specifically?

What I meant by "nerf" was just what was done in HoI2 (and HoI3?) - lower France's IC (relative to its historical industrial output in comparison to the other major powers) and reduce its starting number of divisions to ensure that Germany rarely runs into problems conquering it. While it seems from the dev diary that the devs may be using the "Victors of WW1" National Spirit to accomplish some of these aims in HoI4, the point is that France must lose to Germany most of the time in order for the game to be fun for most people.

However, my understanding is that in reality, France's defeat was far from certain. Even with its investment in the Maginot Line, France had more tanks than Germany, and many of them were more technologically advanced than the models used by Germany at the time. Overall, France's forces were pretty formidable on paper. France's defeat in 1940 wasn't a near-certainty, and resulted from a number of French failures and German successes on the strategic and operational levels. Contrast this with the Pacific Theater of World War 2, where its clear in hindsight that Japan never could have beaten the US. Therefore, no nerf of Japan's naval and marine forces is necessary, as the United State's slow-moving but massive industrial and material might is pretty well simulated in game.

What do you consider a non massive deviation from history?

This is obviously subjective, but what I meant was simply something similar to HoI2. For example, in HoI2, the Republicans might win the Spanish Civil war, but it changes little about the overall outcome of World War 2. Or you might see the Soviets and Allies defeating Germany in 1944 or 1946 instead of 1945. The point is that the major events still take place (assuming no playing interference):

1) Germany invades Poland, WW2 starts more or less on time
2) Germany invades and defeats France
3) No successful Sealion
4) UK isn't kicked out of Mediterranean
5) Germany invades USSR
6) US joins the Allies at some point, and goes to war with Japan
7) Allies invade mainland Europe
8) Allies and USSR defeat Germany and Japan

How would you make it so that the major events in the game start at their historical date with their historical outcome?

Just as the event system did in HoI2, have the event trigger on a certain day, or in a certain month, etc. I assume the devs have some kind of similar capability for the national focus system and AI presets for various countries.

 

[Shatterfury]

8) Allies and USSR defeat Germany and Japan

Even if the player is Germany or Japan ?

 

[FOARP]

I think this is a nice idea. Similar to how AI aggressiveness worked in HoI2.

Whoa! Stop right there! Johan once basically just admitted that aggressiveness in HOI2 didn't actually make the game any harder (can't find the post but I do remember reading it, it's referred to here so I wasn't the only one), and in the end playing with aggressiveness on normal was the recommended way to play since they couldn't afford to spend the time to balance gameplay for every setting. Which is the basic problem with all this stuff about different settings - it leads to different verisons of the game, all of which need to be balanced.

This means that certain nations (such as France) would be weaker in game than they were in reality, and the AI may be hard-coded to make decisions that a rational player never would make (like Japan halting their advance in China) in order to ensure a historical outcome.

But actually a rational player would make exactly the same decision (China was not worth occupying entirely and Japan did not have the manpower to do it anyway) if the game modelled the war in China realistically.

Without player interference, all of the major events of World War II should happen more or less on time.

Do you ever remember this happening in a hands-off game of HOI2? Because I don't. Germany might invade the USSR and then lose in 43 or 44 at the latest, but there would be no D-Day, and the Pacific war would be nothing like the historical one. Inr eality, for this to happen you would have to script every single move, since so many of the major events of WW2 relied on chance events.

The percentage chance of the AI choosing a non-historical route in national focuses, elections, strategic priorities, etc. would be low (maybe between 5-15%).

These decisions don't stand in isolation, so this would not actually work in the way you intend. If there's a 5% chance that Germany declares war on the Czech Republic, then all that requires is for that event to fire and then a bunch of other events that were supposed to be 95% likely to fire then never fire. In reality it doesn't take long for even unlikely events (assuming there enough of them) to throw the game completely off track and on to a historical path that is not scripted for.

 

[TheCrimsonMajor]

Even if the player is Germany or Japan ?

No, sorry, I meant in a hands off game, without player interference.

 

[bengast]

I thought Johan said somewhere (can't recall where) that the AI would go historical if you did, but if you didn't the AI would adapt I assume? Things like trotsky and landon are indeed random but they might happen if the AI doesn't goes the proper path (like soviets getting couped by trotsky because they didn't do the great purge, tho the coup was really early).

 

[Shatterfury]

No, sorry, I meant in a hands off game, without player interference.

Even in an off hands game I think that once in awhile Germany, if not Japan as well, should be able to hold the Allies at bay and get a negotiated peace out of them.

 

[TheCrimsonMajor]

Whoa! Stop right there! Johan once basically just admitted that aggressiveness in HOI2 didn't actually make the game any harder (can't find the post but I do remember reading it, it's referred to here so I wasn't the only one), and in the end playing with aggressiveness on normal was the recommended way to play since they couldn't afford to spend the time to balance gameplay for every setting. Which is the basic problem with all this stuff about different settings - it leads to different verisons of the game, all of which need to be balanced.

Ah, thanks for the correction. Aggressiveness was a poor example - probably a placebo effect I was experiencing. I agree with you that implementing these different modes would be a headache for the devs, which is why we'll probably never see them do it. But we can dream, can't we?

But actually a rational player would make exactly the same decision (China was not worth occupying entirely and Japan did not have the manpower to do it anyway) if the game modelled the war in China realistically.

I disagree with you here. I found that annexing and then puppeting China was much better than partially occupying it. That or never invading in the first place and focusing on Southeast Asia instead.

Do you ever remember this happening in a hands-off game of HOI2? Because I don't. Germany might invade the USSR and then lose in 43 or 44 at the latest, but there would be no D-Day, and the Pacific war would be nothing like the historical one. Inr eality, for this to happen you would have to script every single move, since so many of the major events of WW2 relied on chance events.

You're right - I don't think the AI was very good at D-Day in HoI2, and the US and Japanese AIs in the Pacific needed a ton of work. I do recall a couple of games where the war ended in 1946, though. I guess what I'm trying to articulate is that the approach to HoI2 would be the basis for their ideal default mode for HoI4. A mixture of historical accuracy and balance which gives the player a sequence of events that still feels like World War II, and is a war they can participate in. After all, what fun is it for the Soviet or US player if France challenges Germany about the reoccupation of the Rhineland, and this confrontation leads to a war between France + UK and Germany in 1936? That would have been a pretty short war, don't you think?

These decisions don't stand in isolation, so this would not actually work in the way you intend. If there's a 5% chance that Germany declares war on the Czech Republic, then all that requires is for that event to fire and then a bunch of other events that were supposed to be 95% likely to fire then never fire. In reality it doesn't take long for even unlikely events (assuming there enough of them) to throw the game completely off track and on to a historical path that is not scripted for.

I've never written complex event chains before, so you're right that I don't know how this would work specifically, and I'm sure that there's a lot I have not thought of, due to lack of modding experience. However, I think the initial low probability of deviating from history is still workable. Taking your German-Czechoslovakian war example, the goal of the default mode would be to ensure that in most playthroughs, Czechoslovakia accepts annexation of the Sudetenland in 1938 (if the historical Munich agreement occurs), and accepts full annexation in 1939. That is how it worked in HoI2, unless my memory is faulty. As long as that is usually the outcome, the default mode would be working as intended.

Also, in the event that the 5% chance option to fight is selected, the goal wouldn't be to make sure everything else happens exactly the same as it did historically, but one change in the timeline also doesn't have to mean that everything changes. I think even in 1938, Germany would have been able to hold its own if the UK and France had declared war in response to a fighting invasion of Czechoslovakia. Plus, there's the whole "Sitzkrieg" - the Allies didn't exactly go on the offensive. It seems likely that Germany would still invade Poland and then later France after this event occurs.

I get that this would require significant forethought and testing to determine what the appropriate triggers for each National Focus Decision/Event are, but aren't the devs already doing that with the Focus Decision Trees? Don't all of those have prerequisites beyond their order in the particular tree? Many other Paradox games do feature such triggers/prerequisites, in any event. Finally, these triggers could be tailored carefully. For example, why would the triggers for Operation Barbarossa need to specifically reference the outcome of the Munich Agreement? All it would seem to need is a check for the date, the diplomatic relationship between Germany and the Soviet Union, the German government type, and whether or not Poland and the rest of continental Europe is under Axis control.


I guess what I'm trying to say is that with respect to historically variability, what HoI2 gave us was a pretty decent default, with a lot of room for improvement, and what were likely to get as a default mode in HoI4 is something similar to its approach. The other modes I suggested are unlikely, I know, but they would be nice to have.

 

[TheCrimsonMajor]

Even in an off hands game I think that once in awhile Germany, if not Japan as well, should be able to hold the Allies at bay and get a negotiated peace out of them.

For sure, as long as its not too frequent.