I've long felt there were players who could be described in a similar way, though I've never taken the time to try to quantify it. Some players do things that bring wins your way. However, these players may also do things that bring losses your way, and thus on average they are, well, let's say average. Basically, if you're a cellar-dwelling team, such a player may be interesting to you, because a few of them may be a quick route out of the 100-loss wilderness and let you reach .500. However, to go above .500, you eventually need to move on, to the scarcer player who brings the positives but not the negatives. I guess the same holds true for reaching the post-season versus excelling in it. Win Probability Added is a rough cut at this form of analysis, though not for fielding - it offers a positive and a negative component to plate appearances, which when combined gives the aggregate rating. WAR and its relatives, by contrast, just gives an aggregate. My thinking along these lines has been more about starting pitchers - some guys bring Cy Young stuff to some of their starts, and then in other games they can't find the plate, and though they tantalize you with potential they never do much better than .500 ball. But there's no reason a position player can't fit that profile too - Eddie will occasionally save your bacon with a laser throw to home plate that nails a Bosox runner carrying the tying run, but at too many other times will endure wasted at-bats where he eats out of the pitcher's hand. I've loved rooting for Eddie because he seems like a genuine good guy. But if the team moves on from him, I think I understand their thinking.