Yet more thoughts on this...
I don't know this for a fact but I would hazard a guess that friction between the rings and the bore is highest on the compression down stroke when the peak cylinder pressure gets behind the top ring (and to a lesser extent, the second ring) and forces it out to maximum splay and surface contact pressure.
In theory, this fuel delivered sacrificial layer would be laid down on the fully swept length of the bore on the intake down stroke. If this is so, the sacrificial layer would be 'sacrificed' on the compression up stroke when friction isn't really that high and would be absent on the compression down stroke when it would be most needed. That doesn't sound right to me.
The other thing is that of the three rings that you usually find on a piston, it's the oil control ring that generates most friction over the four strokes of a cycle. I think it's like this because the top two rings are designed to 'retreat' back into their grooves when not under pressure. The oil control ring never gets exposed to high combustion pressure and as such has very little inherent 'flex'. It must be more permanently tensioned towards the bore over all four cycles. Now because the oil control ring is the lowest ring on the piston, it's the ring that gets least exposure to this so called sacrificial layer. Furthermore, because it has two rings sitting above it, the layer is more likely to have been already 'sacrificed' before the oil ring hits the upper portions of the bore. Again, this doesn't sound right to me.
The only way I can see this stuff working is if the friction modifier partially burns to sticky black goo, which ultimately sticks the oil ring flush in its groove (thus totally eliminating OCR friction) which then allows engine oil to enter the combustion chamber and be burnt (it's got a calorific valve so it's fuel right?). Instant fuel economy benefit!! This I can understand.