Blood doping is one of the hardest forms of doping to detect. Athletes take their own blood out and store it for a period of time, and then, right before their competition, they inject it back in, giving themselves a high level of red blood cells, and thus, oxygen-carrying capacity.
It's hard to detect because it's your own blood, with your own DNA, and who's to tell whether it's "new" or "old" blood? I've heard that because it's often stored in plastic bags, that microscopic traces of plastic in the bloodstream indicate that the method was used. But wouldn't it be easy to store the blood in a glass container? And when was the last time you ever heard of an athlete who was barred from his sport because he had too much of his own blood?
Ye Shiwen, the Chinese girl who won the 400 meter individual medley at the 2012 Games, is thought to have blood doped. Before the start of the 400 IM, she and a trainer later found to have syringes went into a bathroom, locked it, then reemerged a few minutes later. (It's a safe bet they didn't go into the bathroom for a quickie right before the start of the race.)
What is preventing athletes from taking their own blood at intervals before the big event, extracting the testosterone from that, and then periodically reinjecting just the testosterone back into their own bloodstreams occasionally, resulting in occasionally extremely high levels of the hormone, which would have a strong muscle-building effect. It would be untraceable, as it would be your own testosterone with your own DNA markers.
It would wreak havoc with your endocrine system, but regular steroids do the same thing, and athletes certainly seem willing to use those.
I don't know whether this method would be feasible; I may be completely out to lunch here, and the amount of testosterone that can be recovered from one's own blood may be too small to be of any use. But I can't help but wonder if the idea hasn't already been tried by some of the countries with state-sanctioned doping programs.