A.L. Wildcard Standings

FIP needs to be incorporated with other stats to be really meaningful, but it gives some insight into how a pitcher performs over-time and can be compared easily with ERA and tRA to see generally how much the pitcher has benefitted from defense and from luck.

Game Scores give you the info on how a pitcher did in any given game, but the average Game Score over a series of games does not give as much information toward future performance as each score is based on a relatively small sample size (<= 9 IP).

What this gives you is a more variable statistic that does not converge well with additional time, giving you relatively-weaker predictive ability going forward. On the other hand, a pitcher’s FIP or tRA for any given game is not particularly informative (see: Feliz’s first game) from a realistic standpoint.

His FIP was negative because FIP is normalized to ERA based on generalized values of K’s (reduces FIP), BB’s (increases FIP), and HR’s (increases FIP).
Having a perfect 100% K rate is an extreme outlier that “breaks” FIP, because that would result in a 0.00 ERA, therefore the FIP should theoretically be 0.00, but the “fit” does not predict that due to the extremeness of the situation and the resultant linear weights being applied.

Game Score is designed for a retrospective look at how well a pitcher performed in any given game, so it is more friendly toward this small-sample size quandry. It was set up as an expansion on the “Quality Start” statistic to be more informative.

For any given game, pulling the Game Score will give you a general indicator for how that pitcher pitched, but it fails to take into detailed account how much of the pitcher’s success was luck or defense.

As such, pulling the average Game Scores for a season with IP totals is simply not as informational to future production as FIP or tRA + IP totals.