Attack numbers are not evenly distributed

from GalCiv4 Forums

For a while I have thought there was something off about attack values in the battle log (and I'm not talking about the raising 1 above the defense issue). This recent battle to me is a good showing.

In this battle, all of my ships have a 2 missile attack. This means I would expect a roughly 50% distribution of 1 damage and 2 damage attacks over a long battle. Against this starbase, I had such a battle. 

When I look at the battlelog, here are my results:

1 damage attacks: 167
2 damage attacks: 8

While some high variance battles are expected, I hope we can agree this is way outside normal distributions. I've seen several others battles where I felt that the attack numbers were just too low, but this is the first I have noted with such a strong distribution.

Version: 1.0
Reproducibility: Click on the electra fleet vs space monster note in the upper left to bring up the battle log for this fight.
Files: https://drive.google.com/file/d/1TJoJuihPk9WPKT1hhPyJhfuIMR_BJCVh/view?usp=sharing

galciv iv bug reports
9,336 views 6 replies
Reply #1 from GalCiv4 Forums Top

Btw, in case people are curious how likely this is to happen "without an error", the odds are:

1 in 2,582,063,386,850,963,586,074,117,355,896,422,046,251

So unless weapons are not designed to be equally distributed, the chance this occurred based on pure random chance and not some issue is exceptionally unlikely.

Reply #2 from GalCiv4 Forums Top

The damage roll is normalized. It isn't a straight die roll. When the range is so small (1 or 2) its normalizing to something between and then losing the decimal and ending at 1 almost all of the time.

The normalization is good to keep battles from being to swingy, but it does look weird at this low of a range.

Siege of Centauri
Reply #3 from GalCiv4 Forums Top

Quoting DerekPaxton, reply 2

The damage roll is normalized. It isn't a straight die roll. When the range is so small (1 or 2) its normalizing to something between and then losing the decimal and ending at 1 almost all of the time.

The normalization is good to keep battles from being to swingy, but it does look weird at this low of a range.

There is also no info about this at all in the wiki, or anywhere else I can find. So what does the curve generally look like? (the formula that generates the result would be wonderful to have).

Does defense have the same formula, or use its own formulas for rolls?

Reply #4 from GalCiv4 Forums Top

I'm glad it's not just me that finds the opacity of the combat mechanics very unenjoyable.

Reply #5 from GalCiv4 Forums Top

well it looks like it's about 20x more likely to be 1 than 2... so my guess is it's generating some random number with 1 decimal place, such as 0.1, 0.2, 1.1, 1.2, etc. and then it's rounding everything to 1, except for when u get an exact 2. so 0.1-1.9 (19 times out of 20) it rounds to 1, and on the 1/20 chance u get an exact 2 you'll actually do 2 dmg. hopefully this is not working as intended and they'll fix it to a more even distribution

+1 Loading…
Reply #6 from GalCiv4 Forums Top

Devs, any chance we can get the formula for how attack and defense values are actually "rolled"? It would help a lot in various analysis I would like to do, and those results may help you with some of your balancing.

So....please please please? :)