3 ms·
> Based on the recent observation, LIGO scientists estimate that these heavy pairings should be almost as common as the lighter binary star systems that astrono
by rwallace 7y ago
> Based on the recent observation, LIGO scientists estimate that these heavy pairings should be almost as common as the lighter binary star systems that astronomers have been studying for decades. Big neutron star pairs should be all over the universe, including our own Milky Way. Why, then, have they never been spotted before?
I don't understand this part. Observations: lots of light neutron star pairs, one heavy pair. This implies heavy pairs are as common as light pairs? How? I would've thought it implied heavy pairs exist but are much less common. What am I missing?
- sfink 7y agoThe observation of lots of light pairs was from directionally focused visible light telescopes, which is a very heavily biased estimator. LIGO has flipped a weighted coin twice. It came up heads once, tails once. That suggests that the weighting is probably not 999999 to 1 (or whatever 5 sigma works out to) in favor of heads, as the current models (based on visible light observations) predict.
- amerine 7y agoYour second paragraph is an excellent summary of why this data is super impactful. Curious about the “but the sample size” crowd.
- perl4ever 7y agoI was thinking the other day along these lines - statistically, I am most likely to be born when the world's population is at its peak. But apparently it is still growing and has been since I was born. Is this surprising? Is it more surprising if it continues?
- simonh 7y agoPrior probability will tell you a lot about populations, but very little or nothing about individuals. For example the prior probability of someone being born with your specific genetics is so close to zero it’s not worth mentioning. That’s true of any specific combination, yet humans exist and so it’s guaranteed that some combinations get to ‘win the lottery’. It’s just not knowable beforehand which ones. In the context of a specific outcome though, and only that context, prior probability just doesn’t matter. I think the problem is our brains can’t help ‘polluting’ the context of an outcome with unrelated contingent information. This makes it very hard for us to reason about these questions.
- perl4ever 7y ago"For example the prior probability of someone being born with your specific genetics is so close to zero it’s not worth mentioning." I don't understand what this has to do with distinguishing between individuals and populations. You could just as well say that the probability of a whole population being exactly as it is, is even lower. You reference "polluting the context of an outcome with unrelated contingent information", but isn't that what you just did? If you think that confuses the issue, then what's the point?
- steve_taylor 7y agoTwo isn’t statistically significant. It may well be coincidence.
- maxnoe 7y agoOne is very statistically significant if your previous hypothesis was zero.