When you say "you would trust" -- I interpret this as your personal "hypothesis". Personally, I would not trust the rationale of coaches and athletes when they are never expressed. There would be too much temptation to introduce personal self-serving bias.
I ask for data, evidence, and controlled observations, and you seem to concede we don't yet have these things, and we may never really know the reasons behind the rationale for athletes doping or coaches doping athletes. This is really my whole point -- there are many claims about the universal power of doping which are not backed up by real world data, whether in the lab, or on the track or the roads or the fields. What is the right word if not faith? Without data, it is not knowledge or evidence or proof. If you think faith is the wrong word, we can more diplomatically say these are all hypotheses still waiting to be confirmed.
You mentioned before: "trial and error" -- this means trying things without knowing the result, and making mistakes -- trying many things that don't work and some things that seem to work. And generally speaking, we the public, you and me, do not know the results of these many trials, nor its success rates, nor the appetite for risk when the risk of getting caught is perceived to be low. The "rational conviction" is part of a risk analysis based on a perceived risk and a perceived reward. Athletes and coaches are not good scientists taking the time to do controlled studies on single variables and identifying and eliminating confounders, so they may draw premature and wrong conclusions, and then tell others, who try things on the strength of word of mouth, or even unspoken perceptions, that that is the thing that worked for someone else. There is a rationale and a logic, but it is not necessarily accurate nor grounded in reliable data.
You accused me of not looking at the reality -- all time performances are the ultimate measure of reality. Recall we are in a thread where the premise is that the fast times of Kiplimo and Chepengetich are strong indications of doping -- so it is relevant to look for prior evidence of comparable fast times everywhere, especially in nations known to dope, to see how reliable these indicators were in history. I did not select countries or athletes at random, or cherry pick athletes, but looked at the fastest times of all athletes worldwide, attempting to find comparative confirmation of popular claims of drugs producing the fastest times, yet not finding it among the fastest times of athletes from 5-continents for nearly three decades. It doesn't make much sense to look among slower times, if the whole premise is that fast times are only possible with drugs. We may find local exceptions for one country or another, for some short time frame, but I looked worldwide to help confirm the power of EPO during the entirety of the EPO-era. I find it inconsistent that when the fastest times are undisputedly comparatively slow, I am reminded to consider "other factors" that explain why such massive world records never happened in rich nations who dope (nations which successfully doped in cycling), but when the times are fast, this spirit of considering "other factors" is readily dismissed or forgotten, and any suggestions to consider them are met with ridicule or hostility.