Notice how I said that players with more snaps would be
more likely to progress faster. Not they be guaranteed to. And everyone still progresses. Let me put up a scenario for you.
Player A is a starter, and he has played 80% of the snaps at his position. We'll say he's a running back. Player B is his backup and played 20% of the snaps. (I know 3+ player rotations at RB are fairly common, I'm just making this easy).
Now the computer would calculate "base progression" which would basically be a toned down version of what they have now. We'll say that the computer determines that Player A gains 2 points overall and Player B gains 3 points overall.
Now the playing time points come into play. Here, the computer "rolls the dice" to determine if they get more points. Let's say the computer calculates that 7 more points will be given out to this position. For each point, we'll say that there is a 70% chance that the point will go to a player that got snaps. Within that 70% chance, there is an 80% chance that Player A will get it, and a 20% chance that Player B will get it (according to their snap count percentages). This way, it is not guaranteed, but Player A is simply more likely to get more points.
I know I'm over-simplifying this using OVR, but because I don't know exactly how it's calculated for any postion, I'm dumbing it down slightly. I'm confident I can be applied with more complex equations though by the programmers that know exactly how progression works now. There would also have to be some kind of plateauing or "diminishing returns" so that higher rated players progress less than lower rated players. There'd have to be some nice magical equation that goes through lots of testing. I'm just giving a base framework for how it could work.