Damage Modifiers for Aliens

[Bomb Bloke]

Your patience is why I wrote that logger. The thought of you actually sitting there, reloading however many thousands of time manually... Ugh! Even I would settle for lower sample ranges then that. :lol:

[Zombie]

Most people probably couldn't handle it, but I can. I never complained about logging those numbers manually. (Well, maybe a little). One thing is for sure though: automatic logging is tons faster. It's kinda hard to stay interested in a project like this if the results take a week (or longer) to complete. Your logger definately is a huge help, both in time, and in effort. I thank you for your thoughtfullness!

 

----------

 

It took a while, but I finally completed the last Incendiary trial. If you remember, I first tested out Incendiary ammunition on the Quarterpod in the air, assuming that IN ammo is an area effect weapon like high explosives are. From that first trial, it sure appeared that IN ammo wasn't a true area weapon like HE is. However, just to make sure, I ran another 1000 trials where the Quarterpod was standing on the ground. Yep, sure enough the results are the same. Incendiary rounds deal the same damage, regardless of the ground. Here are my results, including a total for the two:

 

 Quarterpod vs. Incendiary ammunition
 

[Bomb Bloke]

My guess for the gap in the stats is that 5 is the minimum damage - and 0 is simpley a 'critical miss'.

 

Just thinking, correct me if I'm getting this wrong, but units take more HE damage in the air then on the ground, right? If that's true, then I reckon it might have something to do with fire. The idea is, if you're not on the ground, you won't be standing in fire, and thus won't take fire damage.

 

How much fire damage does a unit take per turn? I gather it takes one turn of standing in fire for a unit to be set alight, how much damage does that do? Can a unit take caught on fire/standing in fire damage at the same time?

[Zombie]

Bomb Bloke, you just might be right! Your comment about a unit on fire and a unit standing in fire spurred my memory. I remembered thet the OSG (Official Strategy Guide, by David Ellis) had some damage numbers for your question. So I looked it up (bottom of pg. 231 for those of you with this guide). What it says is this:

 

"Units that catch fire receive from 5 to 10 damage points per turn until the fire is out. Units standing in a fire receive from 1 to 12 damage points per turn."

 

There is that range of 5 to 10 damage points! The value of 0 should therefore mean that the unit did not catch fire. Makes perfect sense now. That explains the apparent discontinuity in the range. That range is hard-wired into the game, it doesn't happen from splitting the Sectopod into quarters like I thought at one time.

 

So to answer your second question, yes, units can take damage from both caught on fire/standing in fire at the same time.

 

It's stupid why the game would actually have different numbers for the various Incendiary rounds. I mean, you would think that a higher Incendiary strength would do more damage, not the same 5 to 10 points as a lower strength ammo. I guess Incendiary strength just refers to the area of the ground that gets affected by the blast. That's possible.

 

Thanks for your questions Bomb Bloke! They forced me to look something up, which in turn solved the problem. Sometimes all it takes is someone with a fresh point of view!

[Zombie]

In this post I am going to present my results for the Quarterpod vs the Stun Rod. Now, before I get ahead of myself, I want to explain how I gathered the numbers for my previous trials.

 

NKF had modified all of my test scenarios so that every ammo type dealt a 50 for damage. Those values do not occur naturally in the game, and the reason they were all set to the same damage was because of a couple of factors:

1) Too high of a damage strength might accidentally kill the alien instead of hurting it.

2) If the damage is set to a nice round number, the results are easier to compare against each other.

 

Those damage numbers are stored in a file called OBDATA.DAT which is inside a game folder called GEODATA. GEODATA is a global folder, meaning all parts of the game access this for information, from GEOSCAPE to BATTLESCAPE. If the OBDATA.DAT file were found in a mission folder (GAME_1 - GAME_10), those damages would be local to the battlescape mission you were playing at the time.

 

Anyhow, some of you may already know this fact, but the Stun Rod does not have any damage numbers listed for the amount of stun it does. I always wondered how much damage the ol' prod could do, so I thought that this was an excellent opportunity to find out. Instead of keeping the modified damage of 50 for the Stun Rod, I copied my original, unmodified OBDATA.DAT file back into the GEODATA folder. Now I started to run my tests:

 

     Quarterpod vs. the Stun Rod

Minimum:              0 
Maximum:             130
Range:               131
Median:              68
Mode:                17
Ant. Ave:            65
Act. Ave:           66.988

Since the Official Strategy Guide (by David Ellis) lists that the Sectopod will take normal (100%) damage against stun weapons, that means there is no modification to the numbers. Look at my Anticipated Average in the table above. The Stun Rod deals (on average) 65 points of stun damage!

 

Remember my smash 'n grab alien base missions I always run? The base I have been attacking has beginner skill level Mutons there. That means they have a health of 125. Every once and a while I was able to stun a Muton with one jab from the cattle prod. I knew the stun damage of the Stun Rod must be at least 125 points, but I never knew if it was more. Now we all know for sure that the Stun Rod can pump out a maximum of 130 damage points!

 

I had a lot of fun collecting the data for this scenario, simply because there was more riding on this answer than any other trial. So everyone can now write down 65 for the damage a Stun Rod does. I hope some of you actually find this interesting.

 

I'll get those other unreported Quarterpod tests up as soon as I can. Just give me a little time to get the numbers pulled together!

 

- Zombie

[Danial]

So the Stun Rod is 65 and the Stun Bomb does 90 (although that has an area effect).

 

I've read most of this thread as you've been compiling it, Zombie, but unfortunately I get lost with so many number flying around. One thing I've been wondering: Have most of your numbers matched those of the OSG? If not, which ones have been vastly different?

[Zombie]

Yes, the Stun Rod does 65 and the Stun Bomb does 90 for stun damage. However, I wouldn't be too quick to judge the Stun Bomb as a "true" area effect weapon. What I consider a "true" area effect weapon is the high explosive shells. The damage inflicted with a HE shell depends on the ground. The Stun Bomb is completely different in nature as it has a strict range of 0-180 (with an unmodified damage).

 

I'm sorry if you are having a tough time following what I have been doing. Unfortunately, all those numbers are necessary if we are ever to compile a new table based off of experimental results. If I reduce the volume of numbers by just posting the Min, Max and Actual Average, something might get lost in the "translation".

 

So far, all of my tests have reflected what the OSG shows. Those damage modifier numbers the OSG gives in Table 8-7 are correct in terms of susceptibility to ammo types. However, some of my trials (HE ground and air, and also the Incendiary) the damage range could not be predicted from what the other trials showed. Just because the range didn't jive, it doesn't mean that those susceptibility numbers are wrong. For now, just accept those numbers as true.

[OmegaTwo]

I think I might have lost the post explaining it early in the topic, but what are those "Median" and "Mode" values?

[Danial]

In the complete list of values, the "Median" is the middle number, while the "Mode" is the most frequent number.

 

At least, I think that's what they are

[Zombie]

You are correct Danial.

 

Since my results show that the mode is random within the range, perhaps I could drop that stat from my future tables. The median is also a little redundant, especially since the ranges are kinda small. I dunno, what do you guys want? A condensed table listing the Min, Max, Ant. Ave, and Act. Ave -OR- just list it the way I have been doing it. It's up to you, as I can show the results any way you want.

[Danial]

Min, Max, and Av. are fine by me, but I'm not really too fussed

[OmegaTwo]

Oh, thanks for the answer Danial.

 

 

Now, about the results... I agree showing only Min, Max and Average damage, since the Median is pretty obvious and the Mode is random (so it can't figure in a final results table). Plus, I'm terrible with big tables (I can't follow the lines correctly), so the less numbers the better for me.

[Zombie]

Ok, that's fine with me (less numbers to type).

 

Anybody else have an opinion? NKF? Bomb Bloke? I'm open to suggestions!

[NKF]

Who, what, me? Opinion?

 

Hmm, I don't really use the mode myself. I'm more interested in the min and max, but the median cannot hurt either. Just don't start using standard deviations and variances or else I might poke my own eyes out with a pen.

 

- NKF

[Zombie]

So far, I have the list narrowed down to the Min, Max, Median and the two Averages (Anticipated and Actual). Sound about right?

 

NKF, I did try using variances and standard deviations when I first started these tests. Because the individual variances were so small, the standard deviation was also small. I decided to leave them out for this fact. So don't worry, your eyes should be safe for the moment!

[Bomb Bloke]

What I'd like is the formulas used to generate the numbers you've recorded. But that's probably too much to ask, and I could do it easily enough myself.

 

All the info I can really use is all the things that come into effect (level of targets/shooters, fire damage, ammo type damage, etc), the minimum damage, the maximum damage, the average, and the chance of a 'critical miss' (if there ever are any).

 

For example, with the IN trails, the minimum damage was 5, the maximum was 10, and the chance of a critical miss was about 17.5%.

[Zombie]

Formulas you ask? Sure!

 

If the Minimum of any of my tests is 0 you can use the following equation:

 

Damage = INT( rand# * range ), where INT is the integer part of the number, rand# is a random number between 0 and 1 (not inclusive), and range is my test range.

 

For example, take my results for the Quarterpod vs. Laser beam.

The min was 0, max was 150, and the range was 151. The equation would simply be Damage = INT( rand# * 151 ). That would yield a random number between 0 and 150.

 

For my HE tests where the min was something other than 0, use the following equation:

Dam = Min + INT( rand# * range)

 

For the Quarterpod on the ground and the soldier on the ground the equation would be:

Qtrpod HE ground Damage = 16 + INT( rand# * 33 ).

If the Quarterpod is in the air the equation would be:

Qtrpod HE air Damage = 20 + INT( rand# * 41 ).

 

A generalized equation encompassing the damage of the weapon and the susceptibility of the unit to that ammo would be this (not for HE or IN):

Damage to unit = INT( rand# * (Dam of weap * Suscep to ammo + 1))

Ground HE would be:

Damage to unit = 16 + INT( rand# * (Dam of weap * Suscep to ammo + 1))

Air HE would be:

Damage to unit = 20 + INT( rand# * (Dam of weap * Suscep to ammo + 1))

 

The 16 and 20 for the HE trials are a constant, irregardless of the damage of the HE ammo.

 

I'll have to work on the Incendiary equation a bit because I do not know how to model a discontinuous range. Hope that helps.

[Bomb Bloke]

In code, it would go more or less like this.

 

if ( (rand# * 100)

{

damage = 0;

}

else

{

damage = 5 + (rand# * 6);

}

 

That is to say, first it rolls a dice of sorts, and if the value is less then 18 (out of 100), the damage is 0.

 

Otherwise, the damage is calculated as if the minimum damage is 5, and the max 10 (by memory).

[Zombie]

Amazing what a simple IF-THEN-ELSE statement can do (at least that's what we always said in our programming class). You are absolutely correct there, Bomb Bloke. However, that "18" you have for the probability that 0 damage will show up is the actual count. We should be using the Anticipated count for that percentage. 2000/7 = 285.7. 285.7/2000 = 14.3% - assuming a linear distributed count for IN ammo.

 

----------

 

Speaking of Incendiary ammunition. I started some trials with the full Sectopod to see how much damage Incendiary ammo dealt. According to my data, The range is still 5-10, with the chance of a "critical miss" of 0. Therefore, I will not have to do Incendiary trials for the Sectopod because they will be the same as my Quarterpod tests. No sense in beating a dead horse, as I've said before.

 

Since things were going way too easy for me today, I figured that something must go wrong. Yep, and this is a biggie. I was working on the Sectopod in the air vs. HE ammo doing 50 for damage and noticed some problems right off the bat. See, HE ammo not only inflicts damage, but also stun damage. Because the Sectopod is a bigger target, I was actually able to stun that alien a few times. Now, that isn't that serious of a problem because BB's program will still log the health of an alien, even if is unconscious.

 

What is a problem is the level of health left. After 400 trials, I've reached a health level of 6 points left out of 200. 6 more damage points and the Sectopod will die. That's exactly what we didn't want to happen: instant kills. I'm thinking that 50 is a bit too close for comfort.

 

NKF: Here's where you come in. I need the HC-HE damage to be re-modified to something like 25 or so. That way, there aren't going to be any unanticipated instant kills going on. There's nothing worse then running through 1000 trials only to find out that all your numbers have to be trashed because of too much damage being inflicted, especially when the range is so large. So if you could fix that for me, I can restart the HE trials for the Sectopod. No rush here, but I'd like to start gathering some numbers because I have some time now. Thanks!

 

- Zombie

[NKF]

Right 'o. Same place, as usual:

 

https://www.geocities.com/nkfarma/fun/testgames.html

 

Just modified the HE shells to do 25.

 

- NKF

[Bomb Bloke]

I'm not sure if dead units will be logged as having health or not by my program. It's just possible that their health values will under flow...

 

You'd expect the health stat to be 0, but, programming wise, why bother to set it to that if the unit's out of play? It's a redundant step. You might like to try killing a few units, and checking what the log program puts them down as. If the value isn't 0, then you should make sure you note it well if a unit dies so you can discount that trial!

 

If the damage weapons can deal will prevent death of units, fine, but you might like to check that out anyways. Your call.

[Zombie]

Thanks for your quick reactions there NKF. Very much appreciated!

 

----------

 

Bomb Bloke:

 

I did some minor "testing" with your program as soon as it was finished, just to see if it worked, and did what it was supposed to do. When a unit is killed, your logger puts down a value of 0. No overflows, underflows or negative numbers were shown. Trust me, if I would have seen this, you would have heard about it already. I can't use a faulty logging program for gathering results, now can I? That's why it's always important to test out a new program before using it on something where the results count.

[Bomb Bloke]

The logging program would only report what was there. If the game had failed to update that values correctly, then I'd be most sad for you to lay that on my programming.

 

If it was the case that the values showed incorrectly, it would indeed be fixable. I'd just need to get the logger to check whether the unit was dead or not, and default to 0 health if it was. However, the stats are correct, so there's no issue there.

[Zombie]

For some reason, I couldn't find my results for the Quarterpod vs. a Stun Bomb modified to do 50 for damage. I looked high and low on my computer but came up empty-handed. Luckily, while rummaging through a couple of backup discs, I found that file today. (Do you see why it's always a good idea to make a backup of important data)? Per some of your requests, I have abridged the stat list down to the Min, Max, Median, Anticipated Average and Actual Average to make things easier to read:

 

Quarterpod vs. Stun Bomb

            Damage=50
Minimum:        0
Maximum:      100
Median:        49
Ant. Ave:      50
Act. Ave:    50.249

 

Things have been going rather slow for me lately while gathering numbers for the full Sectopod vs. High Explosive ammunition. Multiple problems, oversights, and just plain carelessness on my part, led me to stop for a while and regain my composure.

 

First problem: 50 was too high of damage for HE. Had NKF re-modify the damage to deal 25. Trials completed - 450. Reloaded the game and started over with less HE damage.

 

Second problem: While running the new trials, I noticed that the HE shell sometimes didn't do any damage. I finally realized that the soldiers firing the HE shells were actually missing their target. Darn it, I was using the old version of the game before NKF "upped" the soldier's stats. Trials completed - 1000. Reloaded the game and started over after realizing that oversight.

 

Third problem: I was just shooting at random "quarters" of the Sectopod to finish the HE trials quickly. Then I thought that perhaps orientation of the blast might have some bearing on the results. From my initial observations, it seems to. Trials completed - 550. Reloaded the game and had my guys shoot at the same "quarter" of the Sectopod.

 

Now that I have all the problems ironed out a bit, the testing should go smoother. (Hopefully). I'll try and finish the first set of trials off this week. However, I probably will not post my results anytime soon. This is because I need to run some trials for the Quarterpod vs. the newly modified damage of 25 to compare against the full Sectopod. I also am doing a concurrent test to see how much stun a HE shell inflicts on a target with a given damage number. In any case, I'll keep you updated on my progress.

 

- Zombie

[Bomb Bloke]

I tried one of NKF's files. I got a soldier, with stats upped to max, to miss twice. I admit I was using snapshots, however. Old habits die hard...