Alain Sepeda <[email protected]> wrote: 2015-01-10 15:01 GMT+01:00 Stefan Israelsson Tampe <[email protected]> > : > >> That is the question we need to answer. Typically to validate or disprove >> cold fusion you make sure to draw a representative sample of the old >> results and do a serious examination to evaluate the evidences > > > cold fusion controversy is not statistic, like clinical tests. > > It is "does it exists or never". > > the good techniques is to take the best paper, the best experiment and > check if some theory compatible with old paradigm can explain the result. >
I agree! I have often said that the statistical approach of looking at a large number of experiments would be like this: Imagine in 1904 you want to know if airplanes are possible. At that time, many distinguished experts said they were not, and no machine larger than a model would ever fly. Suppose you were to look at all of the tests of airplanes from 1850 to then. You would find dozens of failed attempts -- possibly hundreds. Plus three successful flights by the Wright brothers, in December 1903. Would you look at all of these attempts to fly, and conclude that flight is impossible? That would be wrong. Or that on average it is impossible? That makes no sense. Along the same lines, one experiment by Storms, McKubre, Fleischmann or Miles is convincing. It makes no difference how many others failed, or how many others were poorly done. Storms cannot be held responsible for a badly done experiment in another lab. It is true that in experimental science, unlike aviation, one test does not prove the point. An experiment generally has to be replicated before we can believe it. So you do need more than one report to be sure. In the case of cold fusion you can take the top 5 reports, or the top 50. Anyone who would reject a claim with 50 replications does not understand the nature of experiments. It makes no difference if there are a thousand failed attempts. Something like an airplane flight or an atomic bomb test is more a demonstration of technology than science, so one test is often sufficient. Replication is not needed. That does not mean the test is obvious or self explanatory. A person who was not an expert in aviation might have had difficulty distinguishing the Wright brother's successful powered flight from one of the failed tests from before that, such as Maxim's in 1894, which did leave the ground. A person who does not understand calorimetry will have difficulty seeing the difference between a success and a failure. For example, that person might not realize that a temperature difference caused by a heat source in the reactor is not the same as a temperature difference caused by the surrounding ambient temperature falling. - Jed
