- #36
Vanadium 50
Staff Emeritus
Science Advisor
Education Advisor
2023 Award
- 33,294
- 19,814
Jonathan Scott said:The paper looks into the statistics quite thoroughly.
True. But I don't find the statistics convincing: a common flaw of a posteriori significance calculations. What does the paper say? It says there were no events passing their requirement. They they changed this, and changed that, until they got a signal compatible with LIGO (and two dozen other signals). Now I don't want to say this procedure is generating a fake signal, but I am saying that once you go down this path it becomes pretty much impossible to calculate a p-value for whatever you find.