I think a couple of things may be happening.
First, there are three Paint Box jammers and two Badger jammers in the field of view of your F-15s' radars, and I think they may be causing some of the problem. So I made a modified version of your save, removing the jammers, as well as the nearby SAMs and all long-ranged missiles on the enemy aircraft. This should remove many of the variables from the test (no jamming, no dodging incoming shots, etc.).
Even with the jammers gone, I had Sparrows going blind on essentially all the shots. Sometimes the missiles would recover and start tracking again, but often they would not. Each time a missile went blind, or recovered from being blind, I compared their courses to see what the relative angle was. Here's an example of the results. Eight missiles fired had nine blindness events. Only one managed to engage its target.
Blind 94.8º - unblind 81.1º
Blind 84.2º - unblind 82.3º
Blind 71.5º - lost
Blind 96.0º - unblind 101.2º
Blind 84.7º - lost
Blind 108.8º - lost
Blind 71.0º - lost
Blind 71.0º - lost
Blind 75.7º - unblind 72º
If you plot the relative courses on a compass rose you get this:
So you can see that the blindness seems to be happening when the courses are roughly 90º +/- 20º. Presumably, this would correspond with the 'doppler notch'. (The cluster at 71º seems suggestive of a zone boundary.)
What puzzles me is that in many cases the missiles would come un-blind when they were still within the notch. And in some cases the missiles would fly within the notch for a long time without going blind. Perhaps it's not an automatic effect, and there's only a small chance of losing lock within the notch, which is checked repeatedly during flight?
I certainly agree that this is a lot more blindness than I recall seeing before. Perhaps that's because I like to micromanage my aircraft. If a closing target's bearing is moving to my right I tend to crank left, and vice versa, putting us on opposing courses, and relying on my radar's ability to look sideways to maintain lock. This not only reduces my rate of closure, and leaves me well positioned to extend and escape, but it would also reduce the likelihood of having a relative course within the notch.
On the other hand, maybe something has changed, and the blindness calculations deserve a look.
(I've attached the test scenario.)