It is very unlikely that you are bleaching the silver - it would take many hours in the fixing bath. I'd bet that blacks are intended to look like that. Conventional wisdom is you can (barely) read newspaper print through the blacks in the frame, though maybe not through the leader!
So I developed the first roll... 16 minutes, the more diluted developer, compensating agitation, 74 degrees.
What dilution did you use? I understand that you used 15 ml concentrate in your 325 ml tank, is that right?
That concentration would need about 19 minutes (ballpark figure) at 20 C (68 F), which would be about equivalent to 16 min at 74 F, which is what you used. In order to have an effect in the leader, the shortage in development time would have to be horrendous, like half the correct time or less. This does not seem to be the case.
The telltale sign of too short a development time is too low contrast. Look at a frame of a subject with high contrast and a full range of tones. You should see the whole range of tones in the negative, from black (almost as in the leader) to almost as clear as in unexposed areas. You should also see good detail in both black and clear areas of the negative. Detail in the clear areas (shadows in the subject) is put there by exposure. If you have bad detail in the shadows, you need more exposure. If you have bad detail in the highlights (black areas in the neg) you need more development.
More developing time will give you more contrast, but the increase in average density will not be very high after a certain point. If you want to try for effect of more time, increase time significantly - one or two minutes would be fine tuning. For instance, do 24 minutes instead of 16.
Another factor may be agitation. I'd do a run with standard agitation instead of compensating, to compare. Standard agitation is more reproducible; your technique may be different than what the manufacturer had in mind when they recommended times for compensating agitation.
OTOH thin negatives scan well, and can be enlarged easily (contrast can be controlled in wet printing) so maybe you could try your results before correcting.
I'm not even considering patological factors - bad chemicals in the developer, for instance. That could also be a cause of problems.