Slashdot Mirror
A Puffed-Up Extrasolar Planet
Maggie McKee writes, "New Scientist Space reports astronomers have found a planet less dense than a wine cork and 38% larger than Jupiter. It circles a star about 450 light years from Earth. A similarly bloated planet has been found before (HD 209458b), so these puffed-up planets may be quite common. But no one knows how they got so swollen. One possibility is 'that some poorly understood mechanism has separated hydrogen and helium in each planet.'"
In what other science does discovering two instances of the same thing make something potentially 'common'?
When you consider that they've only observed an infinitesimally small portion of the universe, seeing two of the same thing suggests that there's a good chance there are more of them.
This guy's the limit!
So, a twice-observed occurence makes a possibly common universal feature, and to explain it, we have a poorly-understood mechanism that somehow does something we don't understand with an effect we can't mimic. Ah, the joys of physics. :-)
Touting MyEclipse AJAX Tools
I'm curious to see your reasoning for this. If you know that your sampling is not representative of the population, or you have a reason to suspect a bias which makes it more likely that you are finding instances of the "something" than if you had a lot more samples available, sure, I'd agree with your reasoning.
"The human race's favorite method for being in control of the facts is to ignore them." -Celia Green
Okay, assuming you have a representative sample, no biases.
Let's suppose you have hat, with numbers in it. There are 10 tens and one each of 0-9. Each number 0-9 is relatively rare, compared to the 10s.
So let's draw a sample from the hat. Our probability of drawing a 10 is 10/20 or 50%. Our probability of drawing a NOT 10 is also 10/20. Suppose we draw two samples, one is a ten, one is a 3. Not knowing anything about the numbers in the hat (how many there are, how many of each kind there are or even WHAT numbers are there) we might decide that there are likely similar numbers of 10s and 3s. If we draw a second sample and it comes up a 10 we might think there are half as many 3s as 10s. From this population our simple estimate of the population is only accurate (judging 3 to be 1/20) if we actually draw every single number!
In a population that is very diverse (there are lots of individually rare items) your chances of drawing any PARTICULAR rare item are very low, but your chances of drawing A rare item are quite high. And we only ask about the rare items when we do find them.
So with a sample of one puffy planet out of ten samples we have three possibilities. 1. Puffy planets might be very rare and we happened to find one. That's unlikely but possible. 2. Puffy planets represent about 1 in 10 planets. That's possible, and fairly likely. BUT, equally likely is 3. Puffy planets are very rare but the population of planets is very diverse.
3. is even more likely since our observations do show us that there are lots of different kinds of planets out there.
Now, if you take an eleventh sample and it's another puffy planet, 3. becomes much less likely (if puffy planets are rare it's unlikely we'd draw two of them in such a small sample) and 2. more likely.
Excellent reply; thanks. Of course, I believe the TFA said that astronomers have seen two "puffy" planets out of ten samples being drawn, not just one; but your point about the difference between a PARTICULAR rare item versus a SPECIFIC rare item is still relevant until we get enough samples to have a better feel for the variation out there.
We've observed around 180 exoplanets via Doppler and have ten which perform transits; how many do we have to observe before we start getting a feel for the more common variants versus the rare exceptional categories...?
"The human race's favorite method for being in control of the facts is to ignore them." -Celia Green