Except, high-drain 18650 batteries are NOT designed for flashlight use, they're designed to be used in power tools. They consider up to 90 seconds as "pulse" time, anything longer than that is considered continuous.
Absolutely true, but one thing is power tools and their known industry leaders, and another one the flashlight merchandising, branding and warehousing, which too often mix the sources like bad martinis....
So personally, I don't sweat it with regulated devices, since all of the branded 20 amp+ continuous batteries (LG, Sony, Samsung) have a 30 amp+ pulse duration that's usually far longer than our regulated mod firing times.
Besides that, on the IPV D2 in particular, with non-temp control builds, since it can't fire below 0.2 ohms, you simply can't exceed the 20a continuous rating.
Beware those misguided comments, please. First, any regulated device which drains off the battery more than it can provide, due to the high power requested, would likely fail to provide that power, or it should show a 'Battery Weak/Battery Low' message. In any case, you won't get what you asked for.
But your second phrase clearly points out to a common misconception. A regulated device, i.e., with a DC-DC transformer/converter between battery and power sink (attys in our case), provides some amperage to the sink, exchanged from another totally different current consumption from the battery. And, within certain boundaries which depend on the electronics and its hardware/software limiters, the battery current drain IS NOT dependant on the atomizer resistance.
In a mech, that resistance is everything, actually, it takes the role of "current valve" for the battery. In a regulated mod, it's the electronics the one and only who decides how much amps are needed from the battery, to provide the atomizer with a certain voltage (and amperage related by ohm's law). They are not the same, they are just related by energy/power considerations in each side, plus the yield/efficiency factor.
I do not encourage to read my blog posts, I know, they are TL;DR!.....

, but the math is there, if you want to guess correct numbers, do the actual math, please!
With temp controlled builds, you're limited to 50J (not 75W) and you're increasing resistance as you fire (lowering amp drain), so worst case during the first part of the firing cycle, a typical 0.1 ohm Ni200 build will only be pulling 22 amps for a brief moment, and drop that figure to under 10 amps rapidly as resistance increases.
TL;DR: Battery amp rating is important, as is buying quality batteries. But any of the quality 20 amp+ continuous units on the market are more than safe enough for regulated single 18650 mods. Mech mod paranoia is still warranted, but most of the battery safety "facts" people are spouting these days dates back a couple years and an entirely different category of devices and use cases.
Temperature controlled devices are forced to run with even lower resistances, and that strikes the electronic boundaries. That's why many mods pull out less effective power in TC mode than in normal constant VW mode, as they run with different resistances (due to the materials). An example:
P4Y IPV D2 (running away the O.T.!). It fires up to 75 W in VW mode, but restrained to 7 V max, so it really achieves 75W with 0,653 Ω or less. As it cannot fire down 3,6 V (it lacks of real DC-DC step down) with that same resistance it cannot go below roughly 20 W. In that mode the electronic accepts between 0,2 and 3,0 Ω.
Now in TC mode, nickel wires are so damned little resistive that any sizeable build ends up down 0,5 Ω or less (if you're lucky and really skilled with enough room for that coil), and the designers of almost all TC-nickel chips have put the resistance boundaries between 0,05 and 0,30 Ω, like the D2 (YiHi SX130H) does. But in that range of resistance the electronics cannot go beyond 50 W (it will fire up something!), so the maximum power is 50 J, that is, watts as YiHi christened them. Both James (Watt and Joule) are deeply concerned at their afterlives....
As titanium coils go a little up and the resistance boundaries are more relaxed, towards usual sub-ohm culprits, usually you can get more power..... But in the D2 the max. power remains the same. In the SX330-whatever (I'm talking about the IPV4S) they can reach 50J in nickel, 100J in titanium, and 120W in normal VW mode. It's the electronic boundaries.
And finally, now it's the turn for the lack of stepping down in VW but some simulation of it in TC mode for the D2. Well, this is an easy one.
Remember the SX130 in the IPV mini? It could be user set to DC-DC (no step down 3,6 V minimum output) and PWM (now it goes down, but using the PWM output). Now you change minor details in hardware and power train (those MOSFET and impedances/inductances) and apply some new logic to TC control......et voilá! ...SX130H!
But it remains without real step-down, unless it simulates it under a PWM output, which is absolutely needed to get to low outputs and avoid burning things (in the coil/cotton/whatever assembly).
Now, the marvellous thing. They had a good baseline to fix costs and control them, and they did all lobotomising the former SX130. Good for YiHiecigar boys! (and our pockets, not too deep lately....

).