I appreciate the detailed response. First, the question that I answered was "why would they not be doping?", not "do some athletes dope?" or "does doping exist/occur". I do not doubt that doping exists and that some athletes use banned substances. My doubts are about whether such doping can produce *unnaturally* fast marathon performances, compared to legal methods like training at altitude, and where the evidence for that can be found. Establishing otherwise would make allegations that performances like 2:11:53 are unnatural more credible.
I claimed there is no scientific evidence that a PED exists, for the marathon, and you responded with "they can still improve performance" (speculation) and "increase red blood cell production and oxygen-carrying capacity" (not evidence of performance). This does not rise to the level of scientific evidence that a PED exists for the marathon -- an sub-VO2max event which it is not obvious how increasing oxygen carrying capacity would result in faster performances. It looks more like personal speculation about what can be, but might also not be. But more fundamentally, to my knowledge, doping performance in the marathon is completely unstudied in all of the scientific doping literature. Similar with the 1500m/mile. All of the time trial studies I've seen range from 3000m-10000m -- events which are run much faster than the marathon, making VO2 more important, and too short to have to worry about factors like energy management after 30km and heat dissipation.
I claimed that there was no strong anecdotal evidence that PEDs exist for the marathon. It is not enough to say one elite athlete was busted suggesting other athletes might also be involved. In addition to doping use, the evidence of a PED has to include the data that shows the "performance enhancement" -- that's what the "PE" is. Simply providing examples like Jeptoo (and Shobukhova) rises to the level of a "proof by example" fallacy, and not "strong anecdotal evidence". As an indication of stronger anecdotal evidence, compare it side by side with what happened in cycling, where many athletes came forward, and many lists were produced showing figures like 7 out of the top-10 were busted or otherwise linked to a dirty team, and the yellow jersey might rightfully belong to the guy who finished 8th. Better still would be enough examples to be representative of the whole group that can be used to establish a correlation between doping and faster performance (which still might be spurious).
I claimed no evidence (forgot to put "scientific") of other mechanisms like recovery. Compared to studies looking at red blood cells and oxygen and lactate, the connection to recovery and improved performance is relatively unstudied at all, let alone for the marathon. Again you respond with "can indirectly lead to faster marathon performances" -- this is speculation and not evidence.
I pointed to the lack of world-class fast marathon performances from non-Africans, and you responded that I cannot imply PEDs are not used. I am not implying that banned substances are not used, but that if they are used, they are not producing unnaturally fast marathon performances for non-Africans. When I looked at all-time performances in 2018, over the span of 33 years from 1985 to 2018, only 9 non-Africans worldwide ran faster than Steve Jones (2:07:13) and Carlos Lopes (2:07:12), the fastest being Ronaldo da Costa (2:06:05). None of them were implicated in doping, with the exception of Spaniard Julio Rey (2:06:52), who ran slower when busted for doping (2:07:37). That covers the entire EPO era up to 2018. Then we enter the era of new shoes.
I claimed that East Africans have been dominating endurance running since 1981 -- with a demonstrated depth of top quality. Think of this like cross-country scoring, where you need a team of 5. Sure the best non-African might beat the 5th best runner every once in a while. You respond again asking if PEDs are being used -- I don't question whether some athletes used banned substances. You also respond say "some non-African athletes have achieved world-class marathon performances" -- which is far short of "demonstrated depth of top quality". Recently, at the top of the non-African list are a few Japanese athletes -- a country not associated with doping. The fastest are just sub-2:05, not a big improvement from Ronaldo da Costa (1998) or Steve Jones (1985). I'm asking how powerful and how widespread can PEDs be if the best non-Africans have only improved 2 minutes in almost 40 years now (including all the other benefits from non-doping factors), and the best of those athletes (Japanese) are most likely clean?
I claimed researchers looking at 12 years of blood data showed the marathon to be cleaner than other events. You repeat again that doping can still occur in the marathon. Sure doping exists and occurs. With respect to blood doping and EPO, these scientists were only (indirectly) suspicious about 1 in 9 of the Olympic and World Championship medals won, meaning 8 out of 9 did not raise any red or yellow flags for these scientists.
We should consider "PEDs don't exist for the marathon" not as something I "categorically dismiss" but as the "null hypothesis" that has yet to be disproved. Establishing the existence of a PED not only requires a representive set of data that includes doping data and performance data strongly correlated, but also requires disproving the null hypothesis that all of the data collected thus far may have no statistical significance.