The WAR fallacy

[SrMeowMeow]

As others have said, there's nothing wrong with WAR as a stat. I think Frobby's real problem is that it can be, and often is, misused to draw conclusions. Two examples of how I have seen it misused on here:

1) Cherrypicking all the negative WAR guys from last year's team, and saying if we got rid of those guys we are automatically going to pick up the absolute value of their negative WARs just by putting replacement players all year. The reason this doesn't work is because just about every team has some negative WAR guys every year, and predicting who they are is difficult. (If it weren't difficult, those teams would cut those guys and sign a journeyman to replace them before the season!) I don't think too many people on thise board would have picked Matusz to be a negative WAR guy in 2011.

2) The other argument I hear is, signing [hypothetical 350 pound free agent] will only add 5 or 6 WAR at most, and we are at least 20 WAR away from contending, so there's no way he could make us a contender, so we shouldn't sign him. That argument is making the assumption that for all the guys we retain, their WARs will stay the same. As discussed in this thread, there is significant variability from year to year with WAR. And the fact is, teams make 20 game jumps in the standings fairly frequently (I think there's more than one 20 game jump every two years), and just about every year there's at least one team that makes a 10 to 15 game jump. And most of those jumps usually have significant internal component -- players that were ON THE TEAM the year before improve their WAR. So an argument that uses last year's WAR as a baseline and tries to say that a 5 WAR improvement is essentially useless, is also flawed IMO. Significant changes in the WAR of retained players DO occur for almost every team.

1) That argument is truly idiotic. Agreed.

2) Our 2011 record should not be the baseline for projecting our 2012 record. Agreed.

Victory!

[CSB Jack]

First of all this is not accurate. The Orioles had more negative replacement value than just about every team with Pie (-2.1 fWAR) and Matusz (-1.0 fWAR) being the worst.

.

.

.

You replace Matusz with Chen who will most likely be around a 2.0 fWAR pitcher, and then you are looking at a 3 game improvement from Matusz.

Matusz is a good illustration of the difficulty of using the previous year for projecting the next year. What is his true talent level? In 2009 he produced 0.9 fWAR in 8 starts. In 2010 he was a 2.7 fWAR over 32 starts. Based on these numbers and anticipating normal (not exceptional) growth for a player at that stage of his career it would have been reasonable to project he should have been somewhere in the 2.5-3.5 WAR range; certainly nobody would have predicted a negative WAR value in 2011 based on the two prior years.

With the changes that the Orioles have made this offseason, it seems a safe bet that Matusz will not be given the opportunity to be a negative WAR player in 2012. It's also possible that he by himself could account for that 3.0 WAR swing from his position on the roster.

You can't get rid of sub-replacement performance, but you can minimize it by acquiring quality depth.

This is huge and seems to be a focus of Duquette's moves this winter. Again back to Matusz, the Orioles appear to have more and better options in the organization now to deal with subpar performance in the rotation. If Matusz shows up in 2012 and performs at the level he left off in 2010, great. If not, the Orioles aren't locked into running him out every five days until it becomes unbearable.