This article from the San Diego PBS station about the meteor breakup near Chelyabinsk quotes a planetarium producer saying that fireballs happen "about every day," but most happen in remote areas or over open ocean, and so do not reach the headlines.
It's true that fireballs (bolides, as in glowing chunks) happen almost daily, but not explosions in the kiloton-plus range.
Here's a Wiki list of significant meteor blasts, some leading to breakups and the touchdown of small meteorites. The list of powerful meteor explosions is considerably bigger than this, but if they happen in places so remote that no one is around to watch them, how do we know?
The answer is that someone is watching all the time, from the other direction: via infrared detectors in space. Perhaps the best known to Pentagon-watchers is an array of geosynchronous satellites under the highly generic name of the Defense Support Program, as in DSP East and DSP West.
DSP satellites have played rather important roles in the American early-warning system since 1970, such as the ability to detect boost plumes of other countries' intercontinental ballistic missiles (ICBMs) or submarine-launched ballistic missiles (SLBMs). Given how fast such missiles travel (about 30 minutes from their missile fields to our targets), and given the limitations of ground-based radar, DSP information is extremely important to military people who fear a surprise attack (Image, USAF):
The Russians have their own infrared satellites, Oko.
Now the Space Based Infrared System is replacing the sturdy old DSP sats. Here's a very interesting article from the Astronomical Society of the Pacific on what unclassified DSP and other missile-watching data told scientists about meteor explosions, covering roughly from 1972 through 1997. Here's a map of the events:
The article explains that this is by no means a complete list because the DOD didn't do a lot of data-preservation in the early years. A meteor explosion over the Marshall Islands in February 1994 (approximately 50-70 kilotons, in two explosions) was the biggest in the publicly released data set.
(Even so, researchers now say the blast on Friday was significantly more powerful than the one in 1994, as in hundreds of kilotons).
So that's what the data suggested as of 1998: on average, every week a rock at least three feet in diameter was exploding in the upper atmosphere. Pretty amazing information, in my mind.
Unfortunately, the DSP data-spigot has turned off.
Geoff Brumfiel noted in 2009 in Nature that meteor researchers suddenly lost access to American space-based infrared data (apparently because the Pentagon didn't want the abilities of the system to be in the public literature, and because SBIRS was soon to go on line).
While USAF sources suggested afterward they might open up the files soon, Geoff's meteor-blast article yesterday in Nature notes that the promised release of new meteor data hasn't happened yet.
But this week's Siberia blast was so remarkable and unsettling that I predict some DSP and perhaps SBIRS data will come out.