Interesting calculation

Folks - thought I'd share this with you. Roger Sinnott of Sky and Telescope posted this to the MPML boards. The NASA website is also very interesting.

Dave

All,

I just made a somewhat interesting calculation (to me, at least).

I went to the NASA NEO risk page ( neo.jpl.nasa.gov/risk/ ) and extracted the cumulative impact probabilities for all the listed objects during the next 100 years. Now, suppose we want to know what the chance is that at least one of these objects will impact the Earth.

The way to do this, I think, is first to calculate the joint probability that *none* of them will hit. If we call the cum. impact prob. for the first object P1, for the second P2, for the third P3, etc., then the joint probability of no impact at all is (1-P1)*(1-P2)*(1-P3)....

The result? For the list as it now stands, I get 0.98918 for no impacts at all. This means that there is a roughly 1 in 92 chance that at least one impact *will* occur in the next 100 years!

Of course, this result is almost entirely dominated by the two objects that currently have the highest chance of an impact, 2000 SG344 and 2006 JY26.

-- Roger

What if you are now 90 years old Don't think it would bother you now

Seriously though, that a high chance for a 100yr timeframe

Keith..

The result? For the list as it now stands, I get 0.98918 for no impacts at all. This means that there is a roughly 1 in 92 chance that at least one impact *will* occur in the next 100 years!

So when did we last get hit? Tunguska June 1908?

So if there is a 1 to 92 chace we get hit in a hundred years and we havent got hit in 98 years then that must mean that we are almost 100 times more likley to get hit in the next year. 100ish times a 1 to 92 chance means there will almost certainly be an impact in the next year and half or do!

Statistics are great when abused

Cheers,
~Al

The result? For the list as it now stands, I get 0.98918 for no impacts at all. This means that there is a roughly 1 in 92 chance that at least one impact *will* occur in the next 100 years!

So when did we last get hit? Tunguska June 1908?

So if there is a 1 to 92 chace we get hit in a hundred years and we havent got hit in 98 years then that must mean that we are almost 100 times more likley to get hit in the next year. 100ish times a 1 to 92 chance means there will almost certainly be an impact in the next year and half or do!

Statistics are great when abused

Cheers,
~Al

all the end of worlders will be out in force then next year .

all the end of worlders will be out in force then next year .

As will those calling for Benitez's/Moyes's heads

Al - the Tunguska impact did not result in a catastrophic event wiping out life from the planet. I'm not a statistitian, but surely, if there was a 1 in 92 chance of being hit in 100 years, that does not equate to 'we will be hit once in 100 years and because Tunguska happened, we WILL be hit in the next year".

I think I remember the probability of heads showing in a coin toss as 50% (or 1 in two). The probablity of a sixth toss remains the same even if you got heads the previous five times.

Its not statistics that have been abused - I think the mathematics used are correct. Its the inference from the data that is not correct.

I could be wrong of course, not being a statistician

I'm not a statistitian, but surely, if there was a 1 in 92 chance of being hit in 100 years, that does not equate to 'we will be hit once in 100 years and because Tunguska happened, we WILL be hit in the next year".

Absoutly true. But wheres the fun in that logic

Statistics are more fun when abused