When you have something which saturates measurements, it's probably not 1/273, it's probably 1/n where n > 273.
In other words, we expect the next few meteors we find will also be made of weaker materials, since every single one has been thus far.
[edit, originally said 'slower' not 'weaker materials']
This is equally puzzling to me. It sounds like you're saying outliers are more likely to be extreme outliers than less extreme outliers, which is... not true.
The real number could plausibly be 1/250, after 10,000 samples. More than that would really call for an explanation: if it was 1/100, why didn't we find another one in the first 272?
Also, the real number could plausibly be 1/100,000.0, we've never seen a counterexample. There's nothing bounding the lower end of the range.
The distribution of these probabilities is substantially lower than 273. You can work it out precisely with Bayes' Theorem, if you'd like, this is a verbal explanation of how that sort of statistical reasoning proceeds.
There's no evidence for this, you can't just point to "there's a lot of numbers and we only saw 273 of them" to justify it. The distribution seems exactly what we're trying to determine; to me it sounds like a good example of when Bayesianism breaks down.
Zero.
Can someone explain why it's rarer than the product of the observations? Assuming hardness and speed are totally uncorrelated that would seem to be the maximum rareness. Only if they're anti-correlated would it be "rarer", but I don't see why that would be. (I would even naively expect them to be somewhat correlated, fragile fast things would be more likely to break apart.)