Your first paragraph contains 2 distinct ideas: first, exceptional seasons; and second, exceptional PR's.
When you look at world-beating seasons, you have to try to find out why, or how, that season was produced, because in addition to the performances of the particular athlete, that measure is also affected by the performances of all the other athletes.
If everyone at the top remains relatively the same for quite some time, and one guy is just better by a bit, then he is, well, just better. If he's consistently better by a lot, he must be considered more carefully.
If, on the other hand, a particular athlete all of a sudden has an exceptional season compared to the rest of their seasons, you have to consider things like how well-trained they were, how long they had been competing, injury status, modifications in technique, etc., because there has to be a mechanism of improvement over their previous years. After the mechanism is identified, the search for a cause or causes of that mechanism can be made. This is the significant individual improvement that can be enabled by luck, fitness, technique, drugs, etc.
Then, there is the situation where a particular athlete has a season that is not only exceptional compared to the rest of their seasons, but to everybody else's season as well. This situation immediately strains credibility, just for statistical reasons. There are so many guys competing, and trying so many things, and sandwiched so close to each other in terms of performance, that even if somebody were to have a really good season, the standout performances of other athletes would be among the top performances of the year. That is not the case with the 3 seasons by Robles, Oliver, and now Merritt, who were absolutely dominant in their years.
And last, but not least, there is the situation where a particular athlete has a season that is not only exceptional compared to the rest of their seasons and to everybody else's season as well, but also compared to everybody else's season, ever. Unadjusted for wind, the averages of the top 10 performances in Robles 2008, Oliver 2010, and Merritt 2012, are respectively 12.97, 12.99, and 12.94 That is a significant betterment of any other dominant season in recent history, if not ever.
When these things are combined--when Merritt is not only so much better than he ever was, but so much better than everybody else has ever been in terms of his single season performance--it is a red flag. It's not conclusive, of course, it must be compared to the baseline they all established. As I've indicated earlier, Merritt's improvement remains a massive red flag.
There are PR's, then there are PR's. Merritt's PR, however, is huge compared to all other PR's. Remember that there are 9 guys crammed into the next .07 increment, and only Merritt in the first .07 increment. You cannot ignore that fact. Is Merritt intrinsically THAT much better than anybody else has ever been? Has nobody ever gotten lucky before in history, with the perfect race? Of course they have. Has nobody ever 7-stepped to the perfect race before? Of course they have. And they are all crammed into the same amount of space that Merritt is out ahead of them all.
At least with someone like Bolt, people tried to find reasons for his inherent superiority. Although those reasons have been disproven fairly robustly, the point is that they were initially sufficiently credible to at least merit analysis, and in the minds of some, are still compelling in the argument.
Fair enough, for Bolt--but what about Merritt? What is intrinsically superior about Merritt over all the other top hurdlers? Absolutely nothing.
The only argument that can successfully be made at this point in Merritt's favor is that previous great hurdlers have just not performed to their potential, as suggested by some disallowed performances. There is no proof of this--people run what they run. Most of these guys have had very long, stellar careers, and the results are visible for all to see. There is even the category of top competitor in the 110mH, as I showed before, in which category the athletes had 1 huge PR, 1 even as big as Merritt's. It is not reasonable to suggest that only those athletes intrinsically inferior to Merritt have had "the race of their lives", for obvious reasons.
Merritt didn't run 0.07 faster than he has consistently run, he ran an incredible .16 faster than he has consistently run in this historically exceptional year alone as judged by the average of his top 10 performances excluding that best performance. Not only that, but the difference between the 12.80 and what he has "consistently run" over any period longer than the immediate 2012 season is ABSOLUTELY GIGANTIC when you consider that, of his 12 best performances listed on the IAAF all-time list, 11 OF THEM WERE RECORDED IN 2012. I'm not even going to show how much of an improvement the 12.80 is over, say, his past 5 or 10 years, somebody can do that if it matters to them. If somebody does it, they should also show how much his 2012 season average improved over his average of the past 5 or 10 years, excluding the 2012 season, of course.
Both Merritt's 2012 season, and the 12.80 individual performance, considered in the light of not only his personal history but of the recent history of the 110mH over about a decade, is most certainly a sign of drug use, IMO.