The diminishing return part is correct! Currently, the DEF uses the classic "survivability" method of determining efficiency, with 140 as target value. What this means is that for every 140 DEF, your effective HP is increased with 100 %. So if you have 500 HP, at 140 DEF you take damage as if you have 1000 HP, and at 280 DEF you would have 1500 HP.
Sorry for bringing this up again, but having gotten stuck back into my spreadsheet I've noticed a few things.
I thought I'd actually set up a formula for this, such that for every 140 DEF you have, the amount of damage done by the Yeti's slam is halved. It should operate like so:
Min DMG = Base damage for attack * 0.5^(DEF / 140)
For the Yeti's slam:
Min DMG = 90 * 0.5^(DEF / 140)
because the Yeti's base damage for the slam is 90 (when you have 0 DEF, on Hard)
If we have 140 DEF, then the (DEF / 140) becomes 1, and 0.5^1 is just 0.5.
0.5 x 90 = 45
Which is how it should work.
Let's now put a trendline for that formula on this graph:
Look at that small dotted line there. At a DEF of 140 we go through 45 on the graph. And it completely misses the plotted points.
I'll forgive, but I still can't find what the actual formula is. I've tried using DEF / 78 instead, which fits it a lot better:
But then we still have that kink after 90 DEF.
Also, the problem I'm having with this line is that I can't get it to fit the other sets of results I've got. I'll go do some more fuckery with trendlines and try find it by trial and error. If
@Teddy doesn't get back before, that is