I think Chris asks a really good question. I tend to be cynical like Sniper and say we will be able to conclude for sure the peloton is clean only when there is no major monetary advantage to winning. But of course that is not going to happen.
One of JV’s main arguments (echoed by Ryo) that the peloton is cleaner if not clean is that wattages are lower than they were in the late 90s. I remember his making a statement that the winning time in some recent MTF would not have been top 15 in 1997, or something like that. He might be right, but the obvious problem with comparisons like these is that there are many factors that can’t be controlled—weather, the difficulty of the ride preceding the finish, race strategy, etc. IOW, there is a lot of noise or variation in times that may be as great as the expected differences between a doped and clean performance.
I suggested a while back that a lot of these factors could be neutralized by having a climbing TT over the exact same route in every TDF. If the grade was fairly constant, it would be a good measure of power/weight. We couldn’t compare times on a route like this directly with those in the past, but over a few years we could get a very good idea of the maximum possible power outputs, and compare these with the maximum outputs performed under various conditions in the past. IOW, if we observed that, say, 6.2 watts/kg was the absolute best any current rider could put out for forty-five minutes, and if greater outputs were observed frequently in the past (particularly by more than one rider*), we could conclude with much greater certainty that current riders were at least less doped. We could also see how power outputs changed in the future.
I think we also need to take the notion of 24/7 observation seriously. I know it is wildly impractical, but if a few elite riders could prove beyond reasonable doubt that they had not performed any illicit operations on their bodies over the course of a season, we could take their performances seriously as clean. Some day not in the too-far future, there may even be technology that will be able to record everything that an individual has put in his body by any route, oral, IV or i.p. Together with the mountain test I suggested, this would give us a very good idea of how much power the best clean riders can put out.
Beyond this, better passports. Ones that look not just at blood values, but at all physiological parameters, and far more radical, functional parameters. Eventually, I think, it will be possible to know not only a rider’s natural physiology but his natural peak performances so well that almost any enhancement capable of making a significant difference in a race will be detectable. IOW, I would extend the passport beyond physiology, to create a baseline of performance, so that any performance significantly beyond this (taking into account a very detailed knowledge of how much clean training can raise performance) will immediately be suspect. We will probably need 24/7 information to establish natural performance maximums, but also contributing to this baseline will be a thorough genetic analysis of each athlete, using information about key genes, and the effect on their regulation by training, to estimate his maximum possible peak performance. I understand how controversial this sounds--it is basically the hardline Clinic position of "since he did so well, he must be doping"--but I really think this is where anti-doping eventually has to go--and can go. I have seen how careful scientists are to interpret blood passport values, and I really believe if this kind of care is extended to a performance passport, it can protect clean riders who have improved solely through legal methods. I raised this point with JV on one of his threads a while back, but he didn’t respond.
*This will make it less likely that a once-in-a-generation talent appears who is significantly better than anyone who follows him in the next several decades. However, even this possibility can be anticipated going forward. Science is approaching the point where we will know enough about the relationship of certain key genes to performance to be able to estimate the best possible performance by any conceivable genetic combination. IOW, we will be able to analyze the genetic makeup of athletes, and from this information, accurately estimate the best possible performance they are capable of.
Edit: This does not mean we are approaching a time where we will be able to predict the outcome of races just by a detailed knowledge of performance and physiological maxima. There are still factors like strategy, timing, motivation, etc. (though eventually I think these too will become more predictable). What I'm talking about here is just power, and a way of differentiating maximum power from a clean rider from that of the same rider doped.