Interestingly, the objection to my approach was not that it might be easier to fake track segments but that it would foul up the route length check performed by the validation tool which was the primary method for validating the track(!). This logic seemed a bit suspect as the track would be incomplete in both cases.
Are you sure?
From what I've read from the DIY organisers that do it, the tracklog isn't used to confirm any kind of distance. It's just used to confirm that you went through (or close enough to) the points you'd nominated as your control points.
The distance check should have been done by the DIY organiser on the list of controls you supplied when you entered the DIY.
There are two slightly confusing uses of 'DIY by GPS' around at the moment (or recently).
1 is the above, specifying a set of control locations for a DIY including some 'GPS' controls where there's nothing to get traditional proof-of-passage, and submitting a tracklog (or snippet) as proof that you went through those points.
2 was a trial based on submitting a GPX file of a proposed route (from bikely or similar), which can be checked for distance, then riding it and submitting the tracklog as proof-of-passage. The two can be compared to ensure that you followed the intended route and any deviations from it are either explained (closed roads/etc) and/or don't affect the distance ridden.
With #2, if the tracklog wasn't recording for a significant period of time it could mean that your ride isn't validated. That's just a risk you have to take (much like the possibility of losing receipts that can invalidate a ride).
I'm not sure what happened to the concept of method #2. Is it formally available or has it just become using method #1 with sufficient GPS controls to bring the route up to distance?