The Endowment Effect Strikes Matt Garza | The Process Report

The Endowment Effect Strikes Matt Garza

A common perception about people is how they value what is theirs higher than they do items that are not. In economics, this is called the endowment effect. Essentially: a person can buy an item at one price, but will only sell at a higher price. The baseball equivalency is as simple as executives placing higher value on players from their own teams than those on other teams with equal production and ability. This phenomenon is prevalent in fan bases as well and tends to emerge during rosterbation sessions.

The Matt Garza trade rumors are legitimate and therefore common. I am not privileged to what each team thinks about Garza, only my opinion and whatever public perception I can manage to detect. I believe the public likes Garza more than I do, but of that I cannot be sure. Take the fan-submitted projections at FanGraphs. Only 18 ballots have been cast for Garza, but the results suggest Garza will record two fewer outs in 2011 than he did in 2010 while posting a slightly lower ERA (a five point decrease) and increasing his strikeout-to-walk ratio so the best part mark of his career (2.57; his current high is 2.39). I find this a modest projection and one I have little problem with.

It takes a special kind of person to visit FanGraphs and vote in their projections. It takes another special kind of person to visit The Heater blog and vote in their polls. A few days ago, Jason Collette passed along a link to a poll on The Heater blog. The poll revolved around – what else – trading Garza. I winced after reading because of its format. The question of whether to trade Garza or not is a complicated one, yet here it’s asked in a loaded manner. There are only two responses which are slightly more detailed than the generic “yes” and “no” options.

You can vote “Yes” if you hold this reasoning: “The Rays have a surplus of starting pitchers, and Garza will bring back more than, say, James Shields.”

You can vote “No” if you hold this reasoning: “He’s been one of that AL’s most consistent starters the past three seasons.”

The phrasing on the “No” response made me curious. When I think about Garza, I think about an occasionally ill-tempered pitcher with highly perceived upside and stuff whose results are above average. The word consistent does not enter my mind. In fact, I don’t know of any consistency mythos that Garza fits. This was only my perception, though, so I chose to go to the numbers and test how valid it was or was not.

Garza’s workload has increased in each season since joining the Rays. At some point, the amount of innings pitched will either cap or trail off (similar to James Shields’ innings pitched totals) but for now that is the trend. Garza has two-straight 200-plus inning seasons and while that speaks to his durability, it doesn’t do much to suggest he’s a super-consistent starter. Garza’s earned run averages are pretty similar, but his ERA+ has dropped by nine in each of the past two seasons. Meanwhile his home runs allowed per nine have increased in each season as his walk and strikeout ratios play popcorn – bouncing up and down year-by-year.

If Joe Morgan is reading this then I doubt his consistency meter is causing convulsions. As it stands, the most consistent aspects about Garza are actually reasons to trade him (an increasing allowance of home runs and decreasing ERA+). Since his consistency is supposed to be a reason against trading him, I’ll allow that perhaps “consistent” is not meant to be taken literally, but is rather an euphemism for a more direct and precise term. Like, say, “good” or the uber harsh “talented”. For whatever reasons, writers and fans tend to think being consistent is the most desirable attribute despite common knowledge that one can be consistently bad, but not good bad – unless, of course, one means bad meaning good and not bad meaning bad.

Pretend that the response is implying that Garza is simply one of the best starters in the American League and the Major Leagues by extension. Is that true? Before getting into the research phase of this post I asked two writer pals of mine about Garza. More specifically, I asked for a round number of current starting pitchers that Garza would fit in. Ben Lindbergh of Baseball Prospectus gave me 30 and Jack Moore of FanGraphs said 50. That means, in so many words, that if every starting pitcher in baseball (and for the sake of simplicity, let’s pretend this is not a keeper league and far away prospects don’t count) were placed into an allocation draft, whereabouts would Garza go?

Because I trust both of their opinions I assumed the research would find that Garza was somewhere between the 30th and 50th best starting pitcher. The only thing left was to do the work in order to find out where he ranked quantitatively. To get the subject pool, I took each pitcher with at least 300 innings pitched since 2008 (100 innings per) which resulted in 127 pitchers. From there, I found Garza’s rank in each of the worthwhile categories. For the sake of readability I broke it down by category.

Traditional counting and rate stats:
Innings Pitched: 28th
Strikeouts Per Nine: 57th
Walks Per Nine: 67th
Home Runs Per Nine: 82nd

Traditional run average metrics:
Earned Run Average: 42nd
Adjusted Earned Run Average (ERA+): 44th

Advanced run average metrics (Note: For some reason, FanGraphs’ innings pitched filter limited the pool to only 127; so the FIP and xFIP rankings are out of 117, not 127. Another note: The SIERA rank is an average of his finishes, as Baseball Prospectus does not offer multiple year leaderboards at this time) :
FIP: 67th
xFIP: 71st
SIERA: 67th

Win metrics:
Baseball-Reference’s WAR: 29th
FanGraphs’ WAR: 47th

Because different folks prefer different strokes, naming an exact range is a highly subjective art. A stubborn old timer might look at this data and say Garza is a top 45 pitcher in the league while a stubborn new timer might look at this data and say Garza is no better than the 29th best pitcher in baseball. My guess is that Garza is indeed a top 50 pitcher, but that he’s not quite top 30. Meaning the original guess of splitting the difference doesn’t change too much here.

The story doesn’t end there. I had all of this data for all of these pitchers covering these three seasons sitting there. This is a dangerous scenario because a curious mind can lead to wandering eyes more often than not. Sure enough, the next thing I knew, I was looking for comparable pitchers using the statistics. I found Brett Myers (who sat only a few rows down from Garza) a close fit and after checking PECOTA’s comparable player list, I checked Randy Wolf to find another reasonable match. By the way, both of those players are in the top five of PECOTA’s list, alongside Bob Welch, Alex Fernandez, and Ron Darling. The next five are a mixed bag of delight and dementia: Ian Snell, Jered Weaver, Kerry Wood, Scott Sanderson, and Dan Haren. PECOTA uses those comparable player lists to form their projections (as does ZiPS, although it uses a different comparable player list). Take a gander at the numbers below and it’ll become apparent as to why these two are on that list:

Garza (08-10): 591.2 IP, 7.1 K/9, 3.04 BB/9, 1.1 HR/9; 3.86 ERA, 4.24 FIP, 4.39 xFIP
Myers (08-10): 477.1 IP, 7.33 K/9, 2.85 BB/9, 1.24 HR/9, 3.90 ERA, 4.30 FIP, 3.91 xFIP
Wolf (08-10): 620.1 IP, 6.73 K/9, 3.13 BB/9, 1.07 HR/9, 3.89 ERA, 4.33 FIP, 4.46 xFIP

The numbers are a go, but when finding comparables I like to match more than numbers. When possible, I want to match build, I want to match stuff, and I want to match whatever else I can; placement on prospect lists, age, speed, and so on. For those reasons, I was hesitant about the comparisons. Garza has youth, relative cheapness, and stuff on his side. At least, I thought.

It’s true that Garza’s average fastball is roughly four miles per hour faster than the average of Wolf and Myers, but his strikeout rates are the second best of the trio and his career whiff rate is actually inferior to both of the slow-throwing sides of the triangle. Velocity might be easier to place a number on than movement, control, command, and what have you, but it’s still only a part of what makes up stuff. I won’t claim to know why Garza’s strikeout and whiff rates are lower, although I would suspect that a byproduct of competition (Wolf and Myers have never pitched in the AL, Garza never in the NL) and perhaps of reliance upon fastballs (although Wolf rides his fastball nearly as hard as Garza). Those are untested and therefore unproven guesses.

As it turns out, the comparisons are actually worthwhile. Since Garza turned 27 in November, then his next season will go down as his 27-year-old season. In the three seasons preceding Myers’ 27-year-old season, he totaled 482 innings pitched, struck out a batter per inning, walked a batter every three innings, and held a 117 ERA+. He missed time in 2007 due to injury and found himself making 48 relief appearances with three starts sprinkled in. He moved back into the rotation the year after and those numbers are reflected in his 2008-2010 statistics above, which are reprinted below:

Garza (08-10): 591.2 IP, 7.1 K/9, 3.04 BB/9, 1.1 HR/9; 3.86 ERA, 4.24 FIP, 4.39 xFIP
Myers (08-10): 477.1 IP, 7.33 K/9, 2.85 BB/9, 1.24 HR/9, 3.90 ERA, 4.30 FIP, 3.91 xFIP

Wolf appears to be the weaker of the two comparables until one looks into his seasons by age. In the three seasons before Wolf’s 27-year-old season, he averaged 191 innings pitched (including back-to-back 200-plus inning seasons), struck out 7.9 per nine innings, walked three per nine innings, and held a 109 ERA+. For comparison, Garza has averaged 197 innings pitched while striking out 7.1 batters per nine innings, walking 3.1 batters per nine innings, and holding a 109 ERA+. Those are frighteningly similar.

Frighteningly only because Wolf got injured in the season thereafter and never truly regained form. He averaged more than three wins (FanGraphs version) in his four big league seasons before the injury. In the seven that followed the injury, he managed to top three wins only once (in 2009) and failed to top the two-win mark (otherwise known as the league average threshold) in five of seven opportunities. There’s no reason to think Garza will suddenly fall into the injury and ineffectiveness vortex anytime soon except that the inherent attrition rate among pitchers suggests it can happen to anyone at anytime. Just ask Kerry Wood, Mark Prior, Brandon Webb, Chien-Ming Wang, and countless others whose names you either forgot or never had the opportunity to learn.

Should the Rays trade Matt Garza? It’s a difficult and multistep question to answer that relies heavily on what you think and know about pitcher attrition, the baseball salary structure, aging patterns, and the team’s competitiveness level in the recent future, and numerous other variables. The case can be made either way in a variety of manners and the most appropriate response is probably that it simply depends on the offers. The folks responding in The Heater blog poll did not have that option selectable and evidently they didn’t need it. Sixty-four percent replied using the “No” option.

Given what we suspect about human behavior I do not believe it would be outlandish to propose that a similar number of fans would have voted no to trading Myers or Wolf at similar ages. Of course, the circumstances between the Philadelphia Phillies at those points and the Rays differ, but folks have the tendency to overvalue what is theirs. There’s a good possibility that more than 64% of Rays fans were not in favor of trading Delmon Young for Garza to begin with and yet look at where we are now. Such is human nature.