With Experiments, you have to explain what you are observing.
The least fantastical explanation is usually correct. In any case, you must be able to mathematically describe your theory to demonstrate it is at least probable. Math doesn’t lie, if your theory is correct, it fits like a jig saw puzzle into place.
With LENR type experiments, you have a reaction without a theory. LENR was the only theory attempted although discarded buy others. People are coming out with new theories all the time that can be tested against the observations.
Nice, but by no means necessary. All that is really needed is that the experiment be reproducible at will. And that has been the hard part with LANR/CF. There were just enough positive results to encourage further investigation, but not enough to meet the absolute standards of "good science".
But LANR/CF has turned out to meet Edison's motto "1% inspiration, 99% perspiration". After a very long and arduous effort, the ability to repeat is finally coming to fruition, and not just from Rossi. On one sense, Pons and Fleischman misled science, in that they made it look simple through their use of a simple apparatus. But as evidence has built up, people have found that the electrolysis of deuterium on/in palladium is VERY complex, and most of the postive results these days don't use electrolysis, but some other approach to get "high loadings" of D2 into Pd.