You assume many things. The basis for your posited reasonable inference is a cherry-picked set of assumptions.
What doesn't make sense is to take Tokyo or Monaco out and say the view is more realistic. The special thing about Monaco is where it sits in the Diamond League schedule for 1500m -- when athletes are trying to qualify for the final.
It doesn't make sense to remove Kiprop, since we were looking precisely for an indication of the number of dopers faster than Elliot, Snell, Cram, and Coe. Before 1997, the number of performers faster than Coe was five. After 1997, the number of fast performers increased, and after the ABP, increased yet again.
You say "unexplained" but this ignores several non-doping explanations for the destruction of world-records in the 1990s, despite providing some of them yourself. EPO is a particularly bad explanation for dominance from athletes originating from local regions, even after they change nationalities and move to countries with testing, given the worldwide nature of availability, the worldwide lack of testing until 2000, and the worldwide lack of OOC blood testing until the ABP in 2009.
Being on the AIU red flag list is about what the AIU doesn't know. It is not evidence of a high degree of doping, but evidence of a low degree of anti-doping -- the process by which the AIU gains knowledge.
All of the myths surrounding EPO and top performance are built on a mountain of what we don't know -- this lack of knowledge is essential to propagating the myths.
Speaking of what we don't know, while you and others take it for granted, we don't even know if El G doped. Not too long ago, 43% of knowledgeable letsrun readers believed his world records were clean.