Fair enough, perhaps we can call that evidence. But in addition to the Hesch story been so heavily injury-related (ie. he clearly didn't increase his training load from 100 % of normal to 110 % of normal with rHuEPO), I also find the following issues with his account:
- If I read Hesch's account correctly, he puts all the benefit to the increased training load and none to the increase in his Hct from 44 % to 51.7 % (+17.5 %) because he claims that he didn't use rHuEPO in connection with races but took care that he ended the treatment a few weeks before races. This is hard to believe and sounds more like whitewashing and at least one your pals found this also hard to believe and didn't even believe half of his story:
Even Larry Eder from the other link wrote that "in reading the NYT piece, Christian Hesch will say that he did not use EPO in his racing EVER. Perhaps he is naive, or perhaps he thinks NYT readers are". If one thinks that Hesch is lying, it is a very slippery slope to the conclusion that the "training load" mechanism is just a madeup story (even when he most likely ran faster even when training with those extra RBC).
- Even if his account is 100 % accurate and he might've improved by hugely after his rHuEPO regimen, in larger context he improved very little even with the extra training load (and with some extra residual RBCs remaining after his treatment). He had clearly natural talent having 3:40.68 PB at 1500m and his half marathon PB improved by roughly a minute after his treatment (1:04:56 -> 1:03.53).
But here is some extra anecdotal evidence I found, not too convincing either. By going through the USADA-report and the CIRC-report, there is nothing about the recovery aspect of rHuEPO, but it is actually possibly mentioned in the Report on Doping in Danish Cycling 1998-2015 focusing in large part on Michael Rasmussen and Bjarne Riis: