In the spirit of improving your results, consider using only adults, only one race at a time (or at least tracking by race), and increasing the sample size per z level. Only 8 random creatures of varying constitution and age will yield reaults that look rather random. If you can do, say, 20 goblin adults per level, you may start seeing more consistency. Also, are you stripping them of all gear, or letting them just drop as they come?
All in all, this is good dwarfy research. Please continue on in the name of !!SCIENCE!!
I was just thinking, he needs bigger sample sizes and better control groups. There is so much more in-depth ‼science‼ to be done! There are so many questions that could want answers!
-Does armor or clothing protect against fall damage? If so, how much?
-Does age affect survivability?
-Toughness?
-What about the type of ground they land on? (smooth stone vs. unsmoothed stone vs. grass-covered soil vs. constructed metal floors (vs. bridges?))
-What if they subject is carrying something heavy, does that make them hit the ground harder and die more?
-Does lateral momentum change the amount of damage or the blood spatter? (i.e. if someone is shot out of a fast minecart and doesn't land on tracks)
-if there's water on the ground, does it reduce damage? Does more water reduce damage more?
To be clear, I heartily endorse your pursuit of ‼science‼, and I think you're doing a good job with the resources you've got. I merely offer suggestions for how to go above and beyond the calling, and I hope your endeavor to procure more test subjects succeeds.