BIPA is recognized and accepted as a volatile hitting stat, where the major year-to-year variations are ASSUMED to be luck.

DER/UZR each specifically attempt to quantify the specific arena of balls-in-play.  The assumption when measuring these is that the results here are SKILL.

The problem here is that it doesn't seem imperically obvious that measuring something for offense should be classified as luck, while measuring it as defense should be classified as skill.  This, I believe, is one (of many) reasons that defensive analysis is so fraught with peril.

For the most part - the reality in all sports statistical analysis is an underlying assumption of facing an "average" pool of competition.  Over a full season, hitters will fact Halladay and Felix, and they'll also face French and Vargas.  The "default" belief is that most variables even out over the course of a season.  BUT ... when there is a clear reason why things should not even out, (playing 81 games in one park), it makes sense to adjust for that context.

The problem is that this may not ALWAYS be true.  Circumstance may skew PAs by hitter X, (because of his handedness, injuries during the season, managers quirks), so that he doesn't actually face an average set of competition.  (This is precisely what Matt's Matrix is attempting to capture and quantify, just fyi).

But, in addition to average competition, there is an additional assumption of "league appropriate" skill.  The standard production of players (and the formulae created to judge them), are based on large pools of players, playing large numbers of games.  The so-called AAAA player, who cannot deal with the talent level don't stick around long - so they have very few PAs - and very little impact on the aggregate formulas.

But, what happens when a player LOSES ability and drops below the normal range of performance?  Well, you get Richie Sexson.  INITIALLY, the .210 BIPA is viewed as a random luck-based result ... but as ability drops below the normal threshold of ability, the formulas don't matter.  "Hitters tend to have .300 BIP rates."  True for most MLB hitters.  Some guys, (Ichiro, Jeter), maintain higher numbers.  Others trend a bit lower, (typical .280s).  But, NL pitchers -- as a group -- have a .218 BIPA.  It's not always easy to see the divide between luck and ability.

The truth is - when dealing with an individual hitter, BIP skews wildly.  But, overall, the LEAGUE BIP is very steady.  But, TEAM DERs (the inverse of BIPA), never skew to the extent that individuals do.  You'll never see a TEAM post a .800 DER or a .600 DER --- though some players will produce .200 or .400 BIP (for a season). 

The important item in this is that one should not assume that just because there is a severe fluctuation that it is therefore, definitely, 'luck'.  It may be.  In many cases, it probably is.  But, rather than assuming "luck" for every wide skew in DER or BIPA - it is probably best to simply identify these outliers as "unexplained variance". 

Me?  I think the league BIP/DERs wouldn't be so steady if DER were "heavily" influenced by luck.  But, I also believe that "desire" may play a larger role in DER results than actual athletic ability.  In the end, I believe we know FAR less about defensive results and how they are achieved than we do about pitching and offense.  So, trusting "exact run" numbers for individual defenders is not something I am willing to do at this time.