AnalyticBridge

A Data Science Central Community

Debunking the story about the Russian meteor event

This Friday, about 15 hours apart, two rare events occured:

1. A 15 meters meteor exploded above Russia
2. A 50 meters meteorite missed Earth by less than 20,000 miles

Astronomers claim that both events are unrelated, because the meteorites in question were flying in opposite directions (South to North vs. North to South).

Here I claim that the 2 events are related, and that indeed, we are dealing with two fragments of the same space rock (the Russian explosion is NOT a test nuke from some rogue entity).

1. The probabilities do not make sense

The first event is said to happen every 100 years. The second every 40 years. Thus, if unrelated, the chance that (1) and (2) occur on a same year is 0.025%. The chance that they occur on a same day is 0.000068%. The chance that in your lifetime, you see such a coincidence is about 75 times higher. Now the chance that any kind of extremely rare event occur during your lifetime is actually very high, but that's another story.

One might suspect that some of the above numbers are wrong: maybe 15, 50, 40 or 100 is wrong, thus making this coincidence far more likely than it seems at first glance. However, I believe that these numbers are accurate enough to make the probability incredibly small, and I persist with my belief that indeed, the two events are thus related. Read below for my explanation.

2. Look at the graph below

The fact that both rocks were flying in total opposite directions actually makes perfect sense! Below A and B represent respectively the smaller and larger space rocks. Metorites are known to break due to tidal (gravitional) forces when approaching large bodies (Sun or Jupiter - or it might have happened long ago).

The graph speaks for itself.

Related articles:

Views: 6117

Comment

Join AnalyticBridge

Comment by Robert D. Brown III on April 3, 2013 at 11:21pm

For what it's worth, this is what NASA is reporting as of March 1, 2013.

http://neo.jpl.nasa.gov/news/fireball_130301.html

Comment by John Morrow on February 25, 2013 at 3:58am

As much as I generally enjoy analytic discussions about probability, I think it's a bit much to label this "debunking" when it contains little actual data.  Topics that can be fully explored through actual, observable data and the laws of physics (ala @David 's point) aren't really well served by cartoon graphs and speculation.

Perhaps a better title would have been "Speculating About the Russian Meteor Event"?  Probability is a wonderful thing, but not that useful when looking at a discrete event about which actual data is available.

Comment by David J Corliss on February 20, 2013 at 9:20pm

Vincent writes: "Also, when you hear "every 40 years on average", is there any regularity in this, as the sentence seems to imply? Or is it "nothing for 10,000 years" then 250 hits in just one year?"

As usual, you are right on target pointing in a statistical problem. 40 years on average with *zero* periodicity - at least for small stuff like this. There is some good evidence for asteroid catastrophies that wipe out nearly all life on earth that repeat every 65 million years or so. However, for small impacts like this one, there is just as much chance of getting hit by one in the next 24 hours as any other day in history. The *long term average* is perhaps forty years but the news media often leaves out the long term part. As far as we can tell, small meteor hits like this are as random as the lottery. It might be centuries before you get a hit - it might be tomorrow.

Comment by Harlan A Nelson on February 20, 2013 at 4:56pm

Objects not captured by the gravity of another body travel in a hyperbolic path around that object.  Objects captured by the gravity take an elliptical path. The Russian object was captured by earth's gravity if it hit the earth so in the end the orbit had to be elliptical.  Also the fact it "streaked" across the sky shows it must have done some bending relative to the earth. So I think the NASA diagram showing a straight path has to be wrong because it does not incorporate the "streaking" aspect.  (It was not a bulls eye hit.) The diagram given by Vincent seems far more plausible to me.  The object most likely circled many times before the orbit degraded.  That is what I think.

Comment by Vincent Granville on February 19, 2013 at 12:46pm

Some claim that the Russian meteorite is 2 meters long, some says it is 17 meters long. I guess 2-meter are a hundred or a thousand time more abundant than 17-meter. With 2-meter, the "coincidence" status changes from being categorized as "nearly impossible" to "very rare", or maybe "rare", or maybe "not that rare" if indeed there are a few such collisions every couple year.

I would think that 2-meter doesn't create such a destructive explosion. But maybe the number of casualties have been greatly exaggerated by local government and the press.

In short, it is clear that there is nonsense and/or errors in the information provided to the public. The purpose of my article was to provide another example where data science is abused and the laws of statistics violated without people noticing.

Comment by Nate Granatir on February 19, 2013 at 9:01am

I find it odd (and disturbing) that this supposed rational, scientific analysis, identifies a (possible!) problem, and then goes on (mostly in the comments below) to invoke a grand conspiracy theory to explain it away.  If you're going to come from a stance that says that probabilities matter, shouldn't Occam's Razor be your first filter?  Is it more likely a scenario that some of the numbers you're using are wrong (as is hinted at in the most recent comments) or that "it is also possible that newspapers and astronomers might be wrong, maybe because the phenomenon is not fully understood despite claims to the contrary, or maybe because they lie"?

Comment by John T on February 18, 2013 at 12:11pm

Somehow, I had the same feeling that the two asteroids were from the same origin. Thanks Vincent.

Comment by Vincent Granville on February 18, 2013 at 9:42am

@David: Ah, only 2 meters for A. That invalidates my discussion, based on the fact that it was claimed over and over to be 17 meters long. Because if 17 meters occurs every 40 years, maybe 2 meters occurs every year (most of the time in the ocean), turning "incredibly improbable" into "very rare".

Also, when you hear "every 40 years on average", is there any regularity in this, as the sentence seems to imply? Or is it "nothing for 10,000 years" then 250 hits in just one year?

Comment by David J Corliss on February 18, 2013 at 9:31am

Amy asks: 'How was the smaller rock not detected at least 5 minutes before impact, while the big rock was detected long ago?"

Good question!! Actually, the smaller rock was very small - only about 2 meters across. What's more important here is cross-sectional area, because that determines how much sunlight gets reflected towards Earth, enabling telescopes to see it.

The smaller rock was 1/25 the diameter of the larger rock, which means it reflected only about 1/625 times as much light for telescopes to see. It was also moving at 6.8 miles per second (as do all incoming meteors). This is what makes it so hard to detect.

Comment by Robert Vanderbei on February 18, 2013 at 9:16am

Here's a JAVA applet simulator (actually, simplectic numerical integrator) that I put together this morning using data from JPL's Horizons System...  http://www.princeton.edu/~rvdb/JAVA/astro/galaxy/Asteroid2012DA14.html

You can't get a full wrap-around as you pointed out.   But, you can get pretty close to it.   Add a little bit of drag from the atmosphere and you can get even further around.   Anyway, Granville's diagram was meant to be a cartoon, not a precise depiction.   My simulator, on the other hand, is meant to be precise.  I added to hypothetical asteroids---one wrapping around each way.