Reload time to damage on gun ratio?

[Queso]

Out of curiosity, what are your testing procedures? There might be some damage you aren't accounting for. It is odd that the graph isn't linear. Nothing in code looks like anything your graph is showing.

[Extirminator]

Out of curiosity, what are your testing procedures? There might be some damage you aren't accounting for. It is odd that the graph isn't linear. Nothing in code looks like anything your graph is showing.

Going on an assumption that light guns have 200 HP and heavy guns have 425 HP (correct me if I am wrong), I use chemical spray on a gun, and hit F6 until it gets damaged. the damage done to the gun is 12.5 on each hit (again, correct me if I am wrong.) . that way I can measure with 200 HP , 187.5 , 175 , 162.5 , 150 ... 12.5 or in heavy gun case - 425 , 412.5 , 400 , 387.5 , 375 , 362.5 , 350 ... 12.5 .

All that reloading in different HP values gets recorded and then put in a editing program to count the frames it took for the gun to finish reloading. number_of_frames * 0.03 ( recording in 30 FPS ) = total reload time in seconds. I then pair them all up and plot them in matlab.

Attaching a file comparing both expected values and measured values for comparison.
EDIT:which for some reason uploaded 2 files? idk whats going on..

[Extirminator]

Actually the only thing that I could be thinking of right now is F6 not doing the same damage of 12.5 on a gun as I previously thought. Which could be quite logical now that I think of it. But I will wait for your confirmation on that - if that's the problem here, I derped up real good.

[Extirminator]

Did some measurements of Mercury fire rate - the results might not be as exact though since the error bars are +-0.06 sec and the mercury suggesting its following any linear pattern is supposed to have a jump of 0.133 seconds for every 12.5 hp down for it's fire rate.

^^So keep in mind these will not be too exact.

Looking at the graph it looks quite linear with several fluctuations that can be easily explained by the low accuracy of the measurements.
it looks like the fluctuations go either way of the graph, above and below with a nice avg of points being very close to the regression line, not like the reload time. So I don't see anything fishy about fire rate behavior.

[Queso]

One last confirmation. Are you going by the ammo count in the bottom right or by some sort of animated asset?

[Extirminator]

One last confirmation. Are you going by the ammo count in the bottom right or by some sort of animated asset?

Concerning the reload time, I was going for the ammo count in the bottom right. I don't know how would that affect the results since if it does increase linearly, the time that it takes for the ammo to reach back to full should still increase linearly too, right? can't you shoot right away after your ammo indicator is back to full? or am I missing another 'invisible' step to count towards the reload?

and talking about rate of fire, I went from ammo decreasing from 2 to 1 ( talking about mercury ) , to the little rotating bullet thing in the middle of your screen to stop rotating.

[Queso]

Yeah that's about the most accurate way to test for both of those things. Hrrmmm. I'm running my own tests now locally to see if any lag could be at play. Very curious indeed.

[Queso]

My last suspicion is that perhaps something odd is happening regarding chem spray. Does that other ship damage button have the same effect? It should be 50 shatter with a modifier of 2 for 100 damage on a gun. This means that if we see a result at about 75% speed then something odd is going on with fire or chem. If we see lower, then I'm pretty much completely lost for how it's operating.

[Watchmaker]

Okay, so Queso and I just shared a minor math epiphany that explains what's going on here.

The trick is the slightly nonintuitive relationship between reload *speed* and reload *time*.  The percentage modifiers from damage affect reload speed, and your graphs show reload time.

A gun at half health has a -25% reload speed penalty; equivalently, its reload speed is 75% of normal.

Think of the reload meter as a progress bar divided into sections, each of which is 1 second long; to reload the hwacha, we need to fill 14 of these.  This takes (14/1)=14 seconds.  But at 75% speed, it takes (14/0.75)=18.66 seconds to fill.

This inverse relationship between speed and time is why your graphs appear nonlinear; the linear relationship with gun health is applied to speed, not time.  If I'm thinking this through correctly: if you invert (y=1/y) your reload time values, so you're measuring "reloads per second", you should see a linear relationship.

[Extirminator]

Aha! Finally things make sense.

I am glad I finally got some linear relationship here, I guess I could say I am done digging up all of this, although I got out more confused than I started pretty much.

Thanks for your time here!

[Queso]

Thanks for the research. Helped get some old reload animation issues out of the code.

[Extirminator]

Thanks for the research. Helped get some old reload animation issues out of the code.

Ah, no problem there I guess. Glad to be of some help to you