How reticent are police in using "justifiable" force to kill criminals?

In 2011, there were 399 reported justifiable homicides by police.  There are real problems with the justifiable homicide numbers as few police departments report these numbers.  But they can still serve as a proxy for measuring how frequently police resort to deadly force when they are threatened.  In 2011, the FBI has data on the number of police officers who are assaulted (54,774) or assaulted with injuries (14,578).  Of course, police may use justifiable force to kill criminals without being assaulted, such as when a criminal is threatening to shot them, and we don't know how many of these cases there are.  But take these numbers as being correct.  These numbers imply that only about 0.72 percent of assaults and 2.737 percent of assaults that injured police end in justifiable homicides by police.

To the extent that the FBI's count of justifiable homicides by police underestimates the true number, the true percentages are higher.  Suppose that justifiable homicides by police is three times higher than reported, that would imply that about 2.16 percent of assaults ended in justifiable homicides by police. For assaults where police were injured, about 8.2 percent of those cases would end in these justifiable homicides.  People can adjust these numbers as they think best.

The problem with FBI justifiable homicides by police is no where near as bad as the numbers for justifiable homicides by civilians, but we don't really even know the size of the errors here.

To the extent that this measure of assaults doesn't include all cases where police might feel threatened, these percentages are too high.