Honestly, statistics do lie. Statistics can give us an
idea what kinda of problem we are facing, but it doesn't tell us
the source of the problem.
For example, the AIM-7's abysmmal combat record had to be
analysed to understand why there was a flaw in the idea of a
Semi-Active Radar Homing missile. It had to be guided all the
way in order to achieve a kill. Although it provided a lot of
reach, it limited the pilot tactically, and was limited by the
constraint of the illuminating radar versus target's RCS and/or
aspect. These are factors that statistics don't say.
The AIM-120 has gone through several revisions ever the first
one was fired in the Gulf War. It is a missile that is still
undergoing revisions, and previous flaws being rectified and
capabilities improved. Statistics do not tell us that as
well.
Hence, it is more empirical to understand the dynamics of the
missile's kill probability than just to examine the statistics.
Its not that the statistics isn't important, but its only a
very surface level indication of problems.
The weapon's engagement zone for the AMRAAM is dynamic, but it
can achieve very high kill probabilities if the target is close
enough or the energy state disparity great enough in favor of the
shooter. It can guide itself to its target once active, or it
can even home in on jamming emmission until it can burn through.
With newer rockets, it's end-game profile has also become
more deadly as it now has more energy to defeat manuevering
targets. These things, statistics do not tell.
As far as it's class of missiles are concerned, in my opinion,
its leading the pack in terms of operational use and
revisions.