Hacker News Reader

"This is why it’s been crucial to have two similar detectors separated by nearly 2,000 miles - one in Washington State, the other in Louisiana - and to seek events that show up in both detectors, thereby ruling out effects caused by local seismic events, passing trucks, and so forth."

Sounds like a good old survival bias ;) Seriously though, what is confidence interval in LIGO?

View on HN

The standard for physics reporting on things like this is "five sigma" or that the test statistic in their hypothesis test must be greater than 5. Since this is a one-tailed test, this corresponds to a p-value smaller than 0.00000029, so the chances of being wrong are about 1 in 3.5 million. This isn't a "confidence interval" like you were asking about, but I think that it answers what you meant to ask. Let me know if you wanted something else.

Thanks, that pretty much explained what I was after.

Reply on news.ycombinator.com