HOI Combat Calculator discussion

[unmerged(14603)]

Some first post!

Just the sort of thing I've been looking for in helping to quickly estimate the combat effects of changing attack and defense values for units. Up to now I've been using a simple formula that gives reasonable but far from perfect results.

I very much hope you continue improving the application as planned.

 

[unmerged(14603)]

I tried out the windows version. It will definitely help modding when you get all the modifiers working.

I was unable to use the calculator because the window was too long (couldn't see output file and below) and I couldn't resize the window. I use resolution 1024 x 768.

Also couldn't exit the program without a crash (used File->exit). Message said waiting to close window and I had to hard reboot.

From what I was able to see you will want to put in a skill = 0 option for leaders. Modders often use this to observe effects without any leader skill interference. The same is true for all modifiers.

I read the readme. For the most part you seem to be following the manual. We are not sure the manual is accurate but (in this case) it might be. We think that block probability (assuming effectivity = 100%) should equal 66.667% not the 10% you are using.

Also, we think only 1 defense value gets a chance to block 1 attack value no matter how high GD is. In other words, if an Inf SA = 5 is attacking an Inf GD = 9, 1 GD will try to block 1 SA (2/3 chance of success) for each of the 5 SA's using up 5 GD's (the other 4 GD's are idle). If GD is less than SA then the excess SA have a 100% chance of hitting/hour.

For Inf vs Inf and Arm vs Arm each hit/hour appears to give either .075, .150, or .225 strength damage (we think each outcome has a 1/3 chance so the average would be .150).

We are less sure how org damage is calculated. Whereas the savegames show strength damage to 3 decimal places, org damage is always a whole number.

There is also the complication of rounding. The Math Guy has suggested that leader bonuses are rounded down. We have not yet figured out the rounding issue.

I have not observed any critical hits in my testing. In spite of what the manual says, they may not exist.

As you can see we are still far from fully understanding how combat damage is calculated. The latest wrinkle is the ground_def_eff variable. We are trying to understand its effects on damage. Unfortunately it appears to be working perversely (higher ground_def_eff seems to reduce the opponent's damage not the unit's). This variable will eventually need to be added as another modifier.

If we can figure out the combat damage formula, your battle calculator would be a great addition.

 

[unmerged(13914)]

Grendel2, I tried out your calculator today. I endorse everything Mikel said about it. A very good beginning.

As I said before, sorry if my earlier posts made it seem as though the manual was right . . . I didn't do enough testing, and my approximations weren't actually as good as I hoped.

However, it seems like we've made a lot of progress on figuring out the real combat system, so in another week or two you might have enough for version 0.2.

Thanks again.

 

[unmerged(14603)]

Since my last post we have made some progress in determining how strength damage is calculated. We are reasonably sure about the following:

Each hour, attack values "line up" against defense values.

Each attack value that has a "matching" defense value has an 80% chance of being blocked (20% hits).

When defense values outnumber attack values, the excess defense has no impact.

When attack values outnumber defense values, the excess attack values have 100% probability of a hit.

Each hit does damage. The damage per hit is variable:

- damage done to all units by soft units (1/3 = .075, 1/3 = .150, 1/3 = .225). Average = .150

- damage done to hard units by hard units sames as above.

- damage done to soft units by hard units (1/2 = .075, 1/2 = .150). Average = .1125

Results are affected by effectivity and ground_def_eff modifiers. We are still working on those items.

Early results seem to indicate that the effectivity% is applied against the attack and defense values. We do not yet understand how rounding may affect this.

The ground_def_eff modifier appears to work perversely. It is added to the opponent's block %. So if a country had a ground_def_eff modifier of +5, this would allow its opponent to block 80 + 5 = 85% of shots. Maximum block % allowed appears to be 99%. A modifier of -5 would make blocks = 80 - 5 = 75%. We don't know if there is a lower limit. Again we do not yet understand how rounding may affect this.

All of the above applies to strength damage. We have not yet looked into organization damage.

Hope this helps.

 

[unmerged(14164)]

Thanks for the responses.

Thanks for the responses.

If you have errors (or more information about the combat system) post them here and I will attempt to get out a fix as soon as possible.

Mikel are you still having 'File->Exit' issues?

Thanks,
grendel

 

[unmerged(14603)]

I just tried out your latest version. Great progress!

I no longer get any CTD's. However, I can't read the battle report very well (it looks like the font size is bigger than the height allowed on each line).

Some ideas on the modifiers:

Both attacker and defender get dig_in bonuses. Sounds strange but that's the way HOI works.

Both attacker and defender have an effectivity% (I think you called it defense modifier or something like that). All terrain, leader, weather, night, dig_in etc modifiers affect effectivity% which in turn affects the atacker/defender points.

Ground_def_eff is a seperate modifier (just like you have it) which directly affects the blocking %.

 

[unmerged(14164)]

It was not very clear in the program but defense percent and dig-in affect both attacker and defender. The 'dig-in bonus' for the defender is a bonus 'head start' dig-in time in days. A number of '3' (the default) will give the defender a 3 point efficiency bonus, which is still, like the attackers bonus, capped at 10.

Effectivity for both the attacker and defender is calculated hourly and is based on day/night and leader traits. This is shown in the output as 'Eff: XXX'.

The output text should no longer overlap, it was not using font metrics for the offset, it is now.

Thanks,
grendel

 

[unmerged(14603)]

Originally posted by grendel2
It was not very clear in the program but defense percent and dig-in affect both attacker and defender. The 'dig-in bonus' for the defender is a bonus 'head start' dig-in time in days. A number of '3' (the default) will give the defender a 3 point efficiency bonus, which is still, like the attackers bonus, capped at 10.

Effectivity for both the attacker and defender is calculated in and is accurate based on day/night and leader traits. This is shown in the output as 'Eff: XXX'.

The output text should no longer overlap, it was not using font metrics for the offset, it is now.

Thanks,
grendel

Sounds good. What does the "Defense %" represent?

Also I forgot to say that we see no evidence of "critical hits" when we look at strength damage.

EDIT: Just downloaded your latest version. Answers all the above so ignore those comments.

 

[unmerged(13914)]

Hi Grendel,

I was thinking of you just today, when I realized there had been enough time for you to update your calculator with our latest results.

I'll download v0.02 and have a look.

By the way, in another few days I will have some very interesting results on low-infra attrition and combat in my AAR, "Winter War Diaries". I haven't had time to write everything up yet, but basically units lose as much as 10 % per day, once each day, if they are fighting in a province they don't own with less than 34 infra.

More on that later, in the meantime, great to see the calculator updated!

 

[unmerged(14603)]

Tried the new combat calculator and it worked great. My only comment regards the readme:

"First the attack does damage to the defender than the defender reciprocates, this gives a slight bonus to the attacker as the defender will be damaged when it first counter-attacks."

IIRC the tests showed simultaneous damage - no attacker advantage. The manual mentions that if the attacker "surprises" the defender then he gets to shoot first. I didn't try to create a surprise situation (set the variable to 0.000) so I don't know if surprise works at all.

Unfortunately RL precludes any more testing (or playing HOI) for a while so at the moment I don't have a better answer.

 

[unmerged(18738)]

I haven't tried it yet but wanted to commend you for your work!

-PK

 

[Aregorn]

Thanks a lot Grendel, I will check it out.