Now you are being silly. You don't even know if an experiment that meets your requirements is already in the library because you won't look. And if it was you would probably say 'that's just one experiment, it doesn't prove anything because there are 4499 which are no good'.'.
Alan - now you are being silly!
To work out whether an experiment is decent takes me, an outsider, a lot of work. 4500 X that. It is a no go. However those familiar with the LENR literature will surely be better placed to identify good candidates; ideally the principal investigators for "good candidates" would get together and argue it out.
Also - I can guarantee - if I pick something the LENR community will tell me I have done it wrong.