Malkoff’s are double thermally potted and done very well, a copper shelf, potting, driver with brass battery contact/sink, potting, spring. Heavy duty to say the least.
I view it like that:
1. Battery stop
Initially the spring compresses, transferring some impact energy to MCPCB. Then there’s stop which quickly transfers the rest of impact energy directly to the body.
2. Stiff battery
All the energy is transferred to MCPCB but unless the spring bottoms out - the transfer is much smoother and peak forces much lower.
BlueSwordM, did you try to bottom out with protected cells? That should be easier to do.
Big force on the driver will make it crack. A little smaller - will make it bend. Repeated bending will make it crack eventually.
Intuition tells me that both long stiff battery and a stop should work well enough. Note at the tail cap there’s a similar issue. Though I think it’s easier to support the tail PCB with the tailcap wall, at least in lights that don’t have tail switches.
But that’s not all of it.
F.e. this light died due to a sideways impact. Managing battery impact wouldn’t help.
This may be related to
But there’s also another thing making me think this is not a full picture. DB Custom’s EE X6 experiments. A spring collapsed there so it was not very stiff and there was nothing that would prevent it bottoming out. Yet the driver (as well as the tailcap) survived despite being directly hit with a cell multiple times.
I did a bit of learning on solder joint failure in drop shock. It happens to be a rather well studied topic and there’s even a standard way of measuring resistance (JEDEC JESD22-B111). In the era of everyone carrying one and often more electronic devices with them, drop resistance is a common requirement.
It is often said that SAC305 is not great for shock resistance. Though in most studies that I’ve seen it’s actually a great performer. There’s a number of specialized alloys that target high shock resistance, typically proprietary with undisclosed composition. I’ve seen indication that these are low-silver products.
F.e. Indium’s SACM is Sn0.5Ag1.0Cu0.03Mn
PCB finish matters.
Overall…the disconnect between common recommendations and study results that I found makes me feel that I don’t have any idea of what’s going on.
Interesting read:
http://www.techni-tool.com/site/ARTICLE_LIBRARY/Indium%20-%20Choosing%20a%20Low-Cost%20Alternative%20to%20Sac%20Alloys%20for%20PCB%20Assembly.pdf
https://alphaassembly.com/-/media/Files/CooksonElectronics/Drop-Shock-Reliability-of-Lead-Free-Alloys—-Effect-of-Micro-Additives—-ECTC-2007.pdf
Well, I probably forgot what actually happened in the X6 road test… the cell compressed the cell to it’s max and then dented in the top of the lap pull cell. This made it too short to reach the spring, so between some spring collapse and the dented cell the light wouldn’t work. I still have that cell and that light, both still work. But at the scene, under an extreme duress, the light did actually fail.
My thought to do the test was remembering those times something got left on top of the car, or on the toolbox on the truck, even on the bumper of the truck after looking underneath. So the 60 mph test was an extreme real world possibility. The light survived, took very little to resume working condition. Again though, had it been potted it might not have been repairable.
The Type III ano actually survived sliding on asphalt quite well, the major damage came when a pebble in the surface started the light flipping and it rebounded into the air then impacted very hard at landing. Comparatively minor damage even then, but the sliding action that could be heard in the video did virtually nothing to the light.
I found the two video’s relevant to this conversation, uploading them to YouTube… it was September 06, 2014. 
Spring collapse may be fixable - get stronger springs or go springless. 
Performance of the rest of the light is remarkable and it would be good if future BLF lights could match it. Or exceed it.
[ADDED]
I mean: it doesn’t feature Elzetta-like wall thickness. It’s not made from high-end alloy. It doesn’t use potting. It doesn’t use any exotic construction. It’s not expensive. There seems nothing special about it but nevertheless it performed incredibly well. Maybe they just did something right…and I’d really like to know what it is.
Though seeing that was a single test of a single specimen makes me wonder if the result was not a pure luck. I guess we won’t be able to do tests at scale though…
[/ADDED]
I’m looking at X6 disassembly pics now.
Looks like there’s a generously sized ring for the driver to sit on. The driver doesn’t seem to be supported from the centre of its back which surprises me slightly. I wonder what’s the driver PCB thickness…
At the tail it looks like the spring presses on PCB which presses on the switch which presses on alu ring which presses on tailcap. So it’s complex but there is some support directly behind the spring.
BTW it just occurred to me that Zebralight or similar light engines are inherently safe against the batery damaging the driver - because it’s easy to make positive contact supported directly by the shelf.
Combine this with DQG style springless construction you have fat contacts at each end (which means relatively small chance of denting the cell) and the result seems very robust.
nice videos
Pretty cool indeed.
I have built a couple of lights for people that I know are hard on their tools, I potted the driver in JB Weld just to ensure they left it alone and didn’t beat it to death tossing the light in the back of the truck and such. Seems like the few times I’ve potted I’ve regretted it later.
Oh, my wife watched in the rearview and thought the light almost hit me in the head, refuses to do THAT again! 
If the light is well constructed potting is always going to be better, and if you are a modder, you may never agree. 30-40 USD I don’t think will get you a light that is good enough that potting might be a good idea.
I disagree. A well made light doesn’t need it. Really that simple. But if someone wants to pay extra for some epoxy, rock on….
what was the regret then, the light still dying, or couldn’t get to inside it?.
thanks.
I guess they weren’t well built lights DB…
does potting have a thermal impact though?
seems like if you just potted an existing light, you might screw up some of the assumptions the designers made about thermal paths, resistance etc…
wle
Potting, especially with a thermal epoxy, will improve the thermal handling. Generally speaking air is an insulator so unless the epoxy is designed for insulative properties it should be an improvement.
The issues were usually people. lol
Like, build a light with 4 modes and last mode memory, pot it, test it, ready to ship out… they say Oh no! I wanted 3 modes and no memory! UGH! Of course it is almost always in writing, the original requests and all correspondence, but you know how it is…
Here it sounds like you know it makes the light tougher.
Here is seems like it doesn’t.
I think most would agree that two lights that are identical will always be tougher when potted.