Actually it's a pretty good stat (1 can tell, because normally HW doesn't bother hyping legit %-type stats, they only care about total $). But not when it runs in all of 10 theaters.
Its an OK stat when comparing two different movies, released at different times, when you can take the average theater count each played in (theaters/weeks run) and use both to weigh the results for a proper comparison.
Now if you FUDGE the number intentionally by releasing in only 2 theaters in the most LIBERAL cities in the US, and then running around quoting its own 'per theater' amount... thats entirely different, and ANYTHING but a 'good' stat.