Other than the afore mentioned external adjustment, there are numerous changes to the Mk25 piston specifcally addressing cold water performance. This includes changes in material and design for two parts in the bushing system, several changes to the piston base cap, and mirror polishing the piston stem. I haven't heard whether or not these changes really help, but the piston upgrade is available for the Mk20.
I just looked at this and one thing is blindingly clear - there is a stunning lack of understanding of some of the issues here, while there is a good understanding of other parts of it.
Basically, my personal opinion is that this is a band-aid attempt and an abortion on top of it. A big part of the problem appears to have been
caused by SP going to the "bushing" system in the first place.
Basic thermodynamics explain how all this happens. Gas cools when it expands. The cooling happens at the site of expansion. In the case of a piston reg, this is at the edge of the piston and seat.
This is a dry area (inside the HP portion of the reg) and as a consequence the air is extremely cold (the higher the tank pressure the colder!) and the heat removal so generated is transmitted to the piston at the knife-edge. This does no harm because there is no water in that portion of the reg.
The problem comes about in that the piston is made of a highly-conductive substance (stainless steel) and as such it cools along its length by conduction and by heat transferrance to the gas along its length. As soon as the HP O-ring is encountered, the other side of the piston is immersed in water (the balance chamber.) It is here that the piston is at its coldest, and where ice will form first. This is also a bad place for ice to form, as if it sticks to the piston shaft it can bind the piston in the O-ring bore, preventing the seat from closing. That happening produces the classic freeze-up freeflow event.
SP's mirror-polishing of the piston shaft is a real attempt to address this, by reducing the ability of ice to "stick" to the shaft, and therefore reducing the risk of the shaft binding in the bore. Good - so far.
Unfortunately the bushings do the opposite thing; the ribs give the ice a place to "stick", which tends to counteract the good! Plastic has crappy thermal conductivity, so the huge thermal mass of the case (which is kept above freezing by the water around the case) cannot help melt the ice that may form in the immediate area of the O-ring. Argh! So we have one good thing and one bad..... let's keep going.
Further down the piston shaft towards the head of the piston, the "big end", there are two things conspiring AGAINST ice formation:
1. The water will tend to transfer heat to the metal of the piston incrementally from the point of the O-ring back.
2. The thermal mass of the big end of the piston is very large compared to that of the shaft. It therefore is unlikely to be a source of freezing until extreme amounts of ice are already in play at the HP O-ring site. As a conseqence all this claptrap of creating a "sealed room" is likely for naught, and in fact may be COUNTERPRODUCTIVE, as that bushing provides thermal INSULATION of the big end of the piston - exactly the opposite of what you want! You WANT the water in the ambient chamber to (1) circulate, and (2) add heat to the piston by conduction. It cannot do so if you insulate the big end of the piston!
I wish I had an easy way to bring a significant quantity (like a hottub worth) of water to near-freezing (say, 40F or so) and run some tests. Specifically, it would be interesting to compare performance of the Mk10, the Mk25 with the "cap" and "sleeve" REMOVED, and the Mk25 with all their "fun stuff" in there for comparative icing and attempt to provoke intentional icing incidents.
This winter, when we have outside temps near freezing, I may be able to arrange the water conditions necessary in my pool for some controlled tests, and can use my long hose to keep the second stage out of the water and breathe through it.
