I don't think we can make any conclusion.

The premise is sound however, provided that this is not a scientific experiment. The reason I say this is there are proven examples of both of the techniques you mentioned being useful for improvement.

The big x-factor in this is that the human body (and how it specifically relates to cycling) is an inherently complicated system, and cannot be distilled down to an essential procedure (i.e. I do this exercise, and hence start winning races). If that were the case, then we wouldn't have the enormous variety of studies on the topic.

Allen/Coggan have their adherents, and the Ride Lots school have theirs. Both of these groups have successful examples. It is logical to assume that both systems work well when applied properly. That doesn't imply that either one of them is superior or (more importantly) that there is not some third system that may not yet be developed which outperforms them. Or a fourth. Or a fifth. etc.

Not to mention the fact that cycle racing is such a complicated sport that there is no one particular kind of development which will guarantee victory. You might spend years developing the ability to produce threshold power, only to be defeated by someone who has a better sprint, or vice-versa. This is just one example of many.