DC's House of Whacky Variable Testing

[Mithel]

Soapy thanks, I've grabbed that and rewrote it slightly and documented it on my website as additional clarification of how setting the modifiers works.

As for why your French broke first and the Germans suffered very little org loss: I think that is because of the way the combat engine randomly selects targets. You have 24 German targets but only 4 French targets. So for any one hour of combat all four French divisions are on average going to be hit by six attacks. But a German division is only going to be hit (on average) once every six hours. Add in a bit of randomness and you'll find that some German divisions won't be hit at all for days.

Summary: the combat engine is slick in how it pairs off divisions against each other. It does not just add up both sides and make one big combat calculation. Each hour, each division will select one enemy division to be it's target.

- Mithel

 

[unmerged(19545)]

Way to go.

Mithel, you seem to have a real handle on collating and posting the effects of these variables. How comprehensive have you been? You seem to be on line a great deal, do you check everything? I have occasionally run across very interesting stuff buried in a thread on some otherwise banal or boring issue. (Boring to me that is. The fellows posting to those threads always seem interested)

And do you have the bandwidth to have everybody referring to your site for every question? Or should we also go ahead and make a new site, or post a collated list in the FAQ forum? Your thoughts please.

 

[Soapy Frog]

You have 24 German targets but only 4 French targets. So for any one hour of combat all four French divisions are on average going to be hit by six attacks. But a German division is only going to be hit (on average) once every six hours.

Of course! I don't know why I was confused in the end... the defenders attack/defence points might be 1/6th the defenders, but they still have 6 times the org and raw manpower.

Makes perfect sense...

 

[unmerged(19841)]

DC, yes, I will continue to document findings on my website. I own my own domain (actually I own two domains) and have maintained it for nearly ten years now, so it's not going to suddenly disappear. If you feel my website doesn't have something documented that we've tested and confirmed please e-mail me full info and I'll update my site. - Mithel

do you have a link to your web sight or a web address??

 

[unmerged(11775)]

Mithel's site

Mithel's Mod (StarfireMod)

He also has a lot of interesting stuff on his site as well.

Link Directory

 

[Mithel]

Anyone interested can also go to:
http://www.mnstarfire.com/ww2/ww2.html
Which is the base of my WWII website. This includes links to my HoI material, IL-2 Sturmovik: Forgotten Battles and also a general WWII history site where I'm beginning to organize a variety of information about WWII history.

DC, I don't see any harm in mirroring what I've put together on how HoI works. Go ahead. The domain that hosts my WWII material has a rather large bandwidth so unless the traffic becomes absolutely enormous it shouldn't be a problem. If I need to I can shift some of those pages to my other domain that has unlimited access (but a slower bandwidth connection).

I generally only read what's posted here in the mod forum. Occasionally I'll visit the AI forum or the General Discussion but basically if the information isn't posted here then I'll only pick it up if someone specifically brings it to my attention. I usually review (and I'm very active on) this forum roughly three days per week.

- Mithel

 

[O'Donnel Aboo!]

For purposes of completeness

I have discovered while modding that every province should belong to some country. The game will start up fine with unassigned provinces. They are the naked off-white color that they are in the editor. however, ordering tropps into the location causes an immediate CTD.

I thought i'd mention this since i had not seen it mentioned anyplace else. Also, this eliminates any Mods where colinization of terra incognita is the goal. It also puts the clamps down on my "Fantasia" Mod where everyone starts with one province and expands.

I had thought of another option, though. Setting one of the User Tags for vast lands of unconquered territory. Only trouble is, you'd have to annex the entire spot to be able to put your units down. So this is all probably a crappy idea.

The moral is: every province should be owned in a scenario.

 

[Gwalcmai]

To address the annex issue, you could have some kind of "colony" event that makes the uber-country cede the province. This would, of course, require one event /available province /playable country (so that each province would fall to the country that is actually controlling it).

 

[O'Donnel Aboo!]

yes, Gwalc, that's what I came up with too...so I aborted that idea all together!

I was pretty upset to discover you couldn't change province resources via event. THere's no way to "strike oil" as it were. Has anyone considered a way to tap the Alaskan oil reserves? Maybe puttig the oil on one of the Aleutians? Lame, I know, but for a modern scenario, the US needs to have access to them. I wonder how MDS is handling that.

 

[Crusher Bob]

If there is no country limit you could also make each 'non-player' province it's own country, then you just invade and annex...

Planning on having national claims over the whole world, or will that prince of terror really power you up?

 

[O'Donnel Aboo!]

Unfortunately, there is a country limit. There is a limited number of tags. I beleive something like 20 undefined tags and maybe 30 non-implemented countries (in Vanilla HoI anyway) THat is a lot of leeway, but not enough for every province, unfortunately.

In most of the MODs i've seen, a good deal of the Tags are used to break up Africa south of the Sahara, which can be tricky. Lumping countries together is always bound to pi$$ someone off!

 

[Soapy Frog]

Here's an interesting tidbit that some of you may or may not know:

Anti-Air efficiency is not affected by any weather/terrain/night modifiers.

Anti-Air always has 100% efficiency regardless... this can be quite bad for the aircraft at night and in other adverse conditions...

 

[unmerged(14683)]

Here's an interesting tidbit that some of you may or may not know:

Anti-Air efficiency is not affected by any weather/terrain/night modifiers.

Anti-Air always has 100% efficiency regardless... this can be quite bad for the aircraft at night and in other adverse conditions...

Ok, so to balance it, air units should have seriously rised night_defense modifier? Or is it ignored in AA batteries fire equations as well?

 

[Soapy Frog]

Ok, so to balance it, air units should have seriously rised night_defense modifier?

night_defense is applied to both attack and defence efficiency when the unit in question is being attacked (i.e. another unit enters it's space and launches an attack).

So it will do nothing since the air unit is 99% of the time the attacking unit (not even sure how it could be the defending unit where AA is involved).

Solution #1: Eliminate night penalties for aircraft. Since units defending at aircraft are always defending at 100% (plus leader modifiers), this will not break anything EXCEPT it will have the highly undesirable effect of night-time air missions being just as effective as daytime air missions.

Solution #2: Err... don't use your aircraft at night... truly a micro-management hell.

So, no ideal solutions here, unless someon can think of something else (find a way to abstract air power completely??)

 

[unmerged(13914)]

I saw this problem in my Winter War AAR and tried to estimate how much difference it made. Russian air losses at night were double their daytime rate, maybe more. However, tac bombers and single-seat fighters really couldn't operate effectively at night and did have higher operational loss rates when they did that. It wasn't historical, because the Red Air Force didn't really spend half its time bombing at night. But it wasn't disastrously wrong.

I also tried making selective adjustments to night_attack for air units that historically operated at night -- mostly strategic bombers. In effect I boosted them ABOVE 100 % for night, not just negating the penalty but giving them an advantage. That works pretty well. Air-to-air is more correct (the fighters have a harder time scoring at night) and the AA losses are more consistent with history too.

However, it's all kind of fruitless because the AI makes almost no distinction between night and day. With tac bombers it will apparently try to ensure at least PART of the run is in daylight. (I'd get two runs per day from the Russians, and in the Arctic winter there wasn't enough daylight for that.) But strat bombers just cycle round the clock, so whether you rate them for day or night operations, half the time they're doing the wrong thing.

Getting rid of the night penalty for air is going backwards. It would be better to help the AI understand what it's doing. But offhand I don't have any clever ideas for how to do that.

 

[Soapy Frog]

Well speaking form a perspective of playability, managing your air units as far as making sure they only bomb during the day is an absolute nightmare. In the end, as far as it being a GAME goes something has got to give...

Either the bug gets fixed in 1.06, or I am going to allow aircraft to operate equally well in day and night, and halve their offensive combat stats to compensate.

I can't tell you how many times I have wanted to simply remove aircraft from the game and model their benefits through tech bonuses to land units, however poor a simulation that would be, it would improve the game immensely!

My other idea is to make a generic "air power" LAND unit, which moves quickly. I would probably delete Militia form the game use that slot to make the "air power" units and also artillery divisions and the like. Lots of problems there too, but I think in the end it would actually be much more satisfying.

 

[unmerged(14683)]

I was wondering...

If it was possible to implement Math Guy suggestions to the naval battles in HoI 1.06 (and Vicky), maybe it would be also possible to do the same with air-to-air, air-to-land rules?

Of course, we should earlier have clear vision HOW exactly air battles should work.

 

[unmerged(19545)]

Cripes! He's back.

Hey, I just had a brainwave. Something to report in the Whacky House of Variables.

What if you set the manpower requirement of a brigade to a negative value? Would it reduce the MP requirements of a division? Could we thus model the later war Triangular Divisions vs the early war Square Divisions? Reduce other values as well? Supply, and the attack and defence values? Have actual cavalry regiments, or armored brigades. All sorts of things, if you can make a negative value work in brigades. Heck, you can just call it the undersized division brigade, and let it do completely different things for each kind of division, in each tech where it might have an effect. If attached to Inf div, for example, it makes (for example) a Soviet emergency rifle division with very little artillery. that way, at each arty level that improves a normally equipped inf div, this one will have a negative modifier, to take that plus back out. I see a plethora of possibilities, but nothing compared to what my much more intelligent and creative neighbours might see. (Yes, I mean you guys in the other mod teams - you know who you are)

And here’s another. We have already tested and proved that you can’t have construction costs with decimal values, because the game doesn’t read anything after the decimal point in the build queue. Does the reinforcement part of the game also read only before the decimal? Or does a unit that costs 5 to build only cost 90% as much to reinforce as one that costs 5.6? Defective minds want to know.

I’ll do some testing for both questions when I get a chance, but maybe one of you will get the chance first (hint hint).

 

[Soapy Frog]

Hey, I just had a brainwave. Something to report in the Whacky House of Variables.

Yay this thread is still alive!!!

What if you set the manpower requirement of a brigade to a negative value? Would it reduce the MP requirements of a division? Could we thus model the later war Triangular Divisions vs the early war Square Divisions? Reduce other values as well? Supply, and the attack and defence values?. Have actual cavalry regiments, or armored brigades. All sorts of things, if you can make a negative value work in brigades.

Interesting idea. You could accomplish the same thing by simply having different models though, e.g.

Infantry Model 0: 'square' division
Infantry Model 1: 'triangular' division
Infantry Model 2: German reduced division

etc. and they can all have completely different stats (although obviously they pick up the same techs, so if your units get 95% of their stats from tech as in vanilla HoI this won't work so well).

Heck, you can just call it the undersized division brigade, and let it do completely different things for each kind of division, in each tech where it might have an effect.

Alas this would not work since tech effects are applied to the brigade type regardless of the parent division type... e.g. tech modfies brigade stats (and then just adds them to the division type), not division type + brigade stats.

And here’s another. We have already tested and proved that you can’t have construction costs with decimal values, because the game doesn’t read anything after the decimal point in the build queue. Does the reinforcement part of the game also read only before the decimal? Or does a unit that costs 5 to build only cost 90% as much to reinforce as one that costs 5.6? Defective minds want to know.

It has been proven conclusively by Mithel and others that reinforcement calculations DO indeed use the decimal values.

So yes a unit that cost 5.6 would cost the same to build as one that cost 5, but indeed would be more expensive to reinforce.

 

[unmerged(19545)]

Yay this thread is still alive!!!

Yay! Someone cares.

Thanks for the word on Mithel's testing. I hope the information becomes more widely known, now that it's here too.


"Alas this would not work since tech effects are applied to the brigade type regardless of the parent division type... e.g. tech modfies brigade stats (and then just adds them to the division type), not division type + brigade stats."

Well, actually, I wasn't thinking in that mode. The starting values placed in the brigade.txt file in DB wouldn't need to have any values other than the negative MP one. Any other effects you want become tech activated, where a given tech WILL do completely different things for the different kinds of units. Anything you would like to be a starting change, rather than something happening later due to tech, is just added to the description of a tech that everyone has, like level 1 infantry doctrine.

The only question then is whether a negative MP requirement in the brigade.txt file will be read as negative, or the game reads things as absolute value. I just talked to Math Guy a little while ago, and he relates that some things are read as written, and some things are taken as absolute value.