I know from working on a large number of regs by various companies and of various designs, especially those used in murky water or regs poorly rinsed by traveling divers, that a sealed diaphragm has a significant advantage as the sealed ambient chamber protects more of the internal parts.
Even with an unsealed diaphragm, the internal parts are still protected by the diaphragm and no moving parts are subject to dirt, salt, corrosion etc due to exposure to water in the ambient chamber - and that chamber is generally easier to rinse than on a piston reg.
In a piston reg like the Mk 25, the piston stem, piston stem o-ring, piston head o-ring and the bearing surface in the swivel cap for at least one of the piston head o-rings is exposed to potential dirt, salt and corrosion. And it is fairly hard to rinse even on later Mk 20/25s with larger ambient pressure holes.
In terms of long term reliability, the Mk 5 and 10 could go years between servicing - if using a properly maintained SPEC system (silicone filled environmental system) and if diving only in fresh water. But in salt water and or with voids in the silicone filler, corrosion could be a real issue making annual servicing essential. The Mk 20/25 in comparison can go fairly long periods of time without service provided it is properly maintained, but the plastic replaceable bushings are now the weak point and limiting factor.
Similarly, the Mk 16 had issues with seat reliability and longevity, but that has been largely addressed with the Mk 17 which uses an entirely different seat and orifice system - one that has been retrofitted back into the Mk 16. Still, I tend to see IP creep on my MK 17s at a slightly greater rate than I saw it on my previous Mk 20s and 25s and I think the Mk 25 can probably go a longer number of dive hours between services - if all other things are equal and the reg is kept clean, silt free and salt free. The reality is however that all things are seldom equal.
As a reg tech, and owning a LOT of regs for a two technical diver family, I tend to defer my own maintence until a reg fails a pre-trip flow check, often (due to SP's painfully slow replacement of warranty parts) I find myself using my own planned annual service parts to service a customer reg and nby default defer my own service until the next parts shipment arrives. With that in mind, I can't say I have noted any higher rate of failure on customer MK 17s than an MK 25s and would conclude that if serviced at the required interval, both are 100% reliable from a design/longevity standpoint.
Perhaps more importantly, I have yet to see a Mk 17 fail in a manner other than IP creep. I can't say the same for the Mk 20/25 as I have seen failed piston stem and piston head o-rings that resulted in fairly substantial leaks. To be fair, I have seen a lot more Mk 20s and Mk 25s with more years on them, but I think the issue relates directly to the exposed nature of those parts and sealing surfaces - something that is absent in the Mk 17 or other sealed diaphragm reg.
Again, where the Mk 17 has the distinct edge is when used in murky water or when not thoroughly rinsed. Mk 25s subject to those conditions come into the shop in worse condition with more issues, greater wear and with far more required cleaning.
In the end however, if properly maintained, all of the above are more than adequately reliable.
----
I have flooded both the Mk 25 and the MK 17. I am probably more comfortable flooding a MK 25, than a Mk 17, but the Mk 17 seems to survive it ok. That said, I try to avoid flooding anything and it's generally bad form. In any case the potential to flood a reg is not an issue in the sense of an unintentional flood as both will survive.
----
For deco reg use I prefer a piston reg as I still regard the diaphragm (in any diaphragm reg) as pretty much a great big piece of fuel. If you are inclined to configure your stage and deco regs in a manner where you can use them in either role, then the O2/diaphragm caution applies to your stage regs as well.
In that regard I think a flow by piston design like the Mk 2 is nearly ideal for O2 use (at least as good as it gets for scuba regs) as the moving part (piston) and dynamic o-rings (piston stem and piston head o-rings) are in the intermediate pressure areas of the reg with no HP gas downstream of the orifice. However the flow rate is not adequate at depth as a stage reg. Consequently, since I value commonality and reduced tools and parts in the save a dive kit, I use Mk 10s for deco and stage regs as they are readily available on the used market, are fairly inexpensive, have adequate performance in both roles and are fairly bullet proof.
But for my use-on-every-dive primary and back up reg, especially for sidemount diving in potentially silty conditions, I use Mk 17s. When you consider I have had them 210' deep while a couple thousand feet back in a cave with my less than perfect personal maintenance schedule, I think that speaks highly of them and of my confidence in them and in their likely failure modes.
For example, I can live with a failure in a Mk 17 when hat failure is just a little IP creep that will not even be a issue on the active reg and would be noticed only as a slight freeflow in the inactive reg - one that can also be turned off to eliminate any on-going gas loss. In contrast, a leaking o-ring on a Mk 25 will leak all the time when breathing off the reg and result in a need to feather it during the active portion of the dive to minimize gas loss - that is a major PITA in comparison.
Those are just my own thoughts and biases and your mileage may vary. But it is probably worth noting that prior to the Mk 17, I had the same basic "piston" reg bias and did not really consider a diaphragm reg for technical diving use.
