Did Spring Training Matter for Free Agents?

Despite missing a portion of spring training, Jake Arrieta has actually beaten his projections thus far.
(Photo: Ian D’Andrea)

As was documented on a number of occasions at this site during the winter months, the 2017-18 offseason represented the slowest free-agent market on record. While the economic implications of the offseason remain unclear, that’s not the only way in which the game was disrupted this winter. Because a number of players signed late, a number of players also benefited from something less than a full complement of spring-training games. Logic dictates that could have an effect on performance. But does the data support that logic?

We could attempt to answer that question by examining performances from just the first couple weeks of season — when late signees would still, hypothetically, be getting reacquainted with the speed of the game. There would so much randomness involved in such a study, though, it would be impossible to reach any real conclusion based on a handful players. It might not be entirely responsible to try and draw conclusions from two months’ worth of performances, either. As it stands, though, we at least have a larger sample with which to work — and if we considered production too far removed from spring, we might end up not testing the effect of missed time, at all.

Of the Top 50 Free Agents, 10 signed in March or later. One of those players, Greg Holland, is a reliever. We could compare Holland’s performance to the 15 other relievers, but I’m not sure that would be a worthwhile endeavor. Holland has made 18 appearances, only pitched 13.1 innings, and recorded 15 walks, 10 strikeouts, a 5.97 FIP, and a 9.45 ERA. His -1.52 WPA is the worst in the National League. I didn’t run the numbers, but suffice it to say, there’s no way the rest of the free-agent relievers have been as bad as Holland. You could chalk that up to a lack of spring training, but it might be more worthwhile to look at position players and starting pitchers.

There were 22 position players in the Top 50, and 17 players signed before March 1. Here’s how the early signers were projected as well as how they’ve performed so far this season.

Early Free Agent Signing Performance
Name Sign Date Actual wRC+ Projected wRC+ Difference
Welington Castillo 12/1/17 112 93 19
Zack Cozart 12/15/17 88 101 -13
Carlos Santana 12/20/17 110 124 -14
Yonder Alonso 12/23/17 107 116 -9
Jay Bruce 1/16/18 74 106 -32
Howie Kendrick 1/18/18 115 99 16
Lorenzo Cain 1/26/18 125 100 25
Alex Avila 1/31/18 22 94 -72
Todd Frazier 2/7/18 117 104 13
Eduardo Nunez 2/18/18 68 98 -26
J.D. Martinez 2/19/18 173 138 35
Eric Hosmer 2/19/18 135 116 19
Jarrod Dyson 2/19/18 60 75 -15
Carlos Gomez 2/20/18 69 94 -25
Cameron Maybin 2/21/18 73 88 -15
Logan Morrison 2/25/18 81 112 -31
Lucas Duda 2/28/18 96 113 -17
AVERAGE 96 104 -8

We’ve seen very good performances from J.D. Martinez and Lorenzo Cain, with Welington Castillo performing well before his suspension, Howie Kendrick performing well before his injury, and Eric Hosmer turning in a solid year thus far. We’ve also got a bunch of players within 15 wrC+ of expectations and then some disappointments — mostly notably in Jay Bruce, Eduardo Nunez, and Logan Morrison, the last of whom has been a little better after a slow start. Then there’s Alex Avila, who is responsible for half of the overall difference between the projections and results so far. Overall, there are five players beating projections by at least 15 points, six players within 15 wRC+ of their projections, and six players 15 points worse than their expectations. This is pretty close to what we might expect to see.

We have just five position players for comparison. These players all signed in March.

Late Free Agent Signing Performance
Name Sign Date Actual wRC+ Projected wRC+ Difference
Jon Jay 3/6 99 85 14
Mike Moustakas 3/10 116 110 6
Neil Walker 3/12 59 113 -54
Jonathan Lucroy 3/12 92 113 -21
Carlos Gonzalez 3/12 84 95 -13
AVERAGE 90 103 -13

Jon Jay and Mike Moustakas have both exceeded expectations with the Royals. Jonathan Lucroy has performed much closer to last year’s disappointing season (82 wRC+) than his career numbers. The same is true for Carlos Gonzalez. Neil Walker has put up a wRC+ of at least 106 in every professional season, including last year’s 114 mark for the Mets and Brewers, but he has not gotten it together this season at all, losing playing time to Gleyber Torres. The group has disappointed overall, though Lucroy is actually doing better than last season and Gonzalez is about the same. There probably aren’t any broad conclusions to reach here, and we can’t say definitively that a lack of a spring has hurt these hitters.

Pitchers received a bit more scrutiny in the spring, as their arms need some time to get ramped up for the innings necessary to pitch. First, here are the pitchers who signed early.

Early Free Agent Signing Pitcher Performance
Name Sign Date Actual FIP Proj FIP Difference
Doug Fister 11/28/17 5.20 4.81 0.39
Tyler Chatwood 12/7/17 4.99 4.31 0.68
Jhoulys Chacin 12/21/17 3.98 4.66 -0.68
CC Sabathia 12/26/17 4.72 4.66 0.06
Yu Darvish 2/13/18 4.81 3.39 1.42
Jaime Garcia 2/15/18 4.94 4.51 0.43
Andrew Cashner 2/15/18 5.18 5.27 -0.09
Jason Vargas 2/16/18 5.18 4.63 0.55
AVERAGE 4.88 4.53 0.35

The one who sticks out among this group is Yu Darvish, currently on the disabled list. The Cubs were hoping for ace-type performance but haven’t gotten it yet. Fellow Cub Tyler Chatwood is also on the ugly end of the spectrum, with Jason Vargas also disappointing by FIP. Jhoulys Chacin is the only pitcher clearly outperforming his projections. Doug Fister, CC Sabathia, Jaime Garcia, and Andrew Cashner have been pretty close to expectations. As for those free agents who missed most of spring, the table below shows their performance compared to projections.

Late Free Agent Signing Pitcher Performance
Name Sign Date Actual FIP Proj FIP Difference
Jake Arrieta 3/12 3.51 4.01 -0.50
Lance Lynn 3/12 4.40 4.65 -0.25
Jeremy Hellickson 3/17 3.29 5.04 -1.75
Alex Cobb 3/21 5.14 4.39 0.75
AVERAGE 4.09 4.53 -0.44

Jake Arrieta is off to a very good start with Philadelphia, while Jeremey Hellickson, currently on the disabled list, got off to a great start with the Nationals after a disaster of a season with the Phillies and Orioles a year ago. Even removing Hellickson from the group, we find that Arietta, Lynn, and Cobb have combined to pitch to the projections as a group thus far. The two pitchers perhaps most associated with struggling after signing late deals are Lance Lynn and Alex Cobb. Lynn started off with a poor debut and had a few other rough starts in his first month, but overall, his FIP is actually beating expectations. Since his first start, he’s put up a 4.65 ERA and 4.05 FIP, the latter beating projections by a wide margin. He’s definitely walking more batters than might be expected, and that might be a sign of some rust, but he’s also inducing more ground balls and giving up fewer homers. He’s been hurt some by a .337 BABIP that makes his ERA worse than his performance might dictate.

Alex Cobb has pitched poorly, but he hasn’t pitched as poorly as his 7+ ERA shows; his .374 BABIP and 59% left-on-base rate aren’t entirely his fault. Even so, there’s not much to be done to sugarcoat Cobb’s performance. He does have a 4.68 FIP in his last nine starts, which is pretty close to the projections so there’s that. Maybe we could attribute those poor first starts to rust, but Yu Darvish and Jhoulys Chacin had poor first starts, too. Getting players into camp and on schedule with the big-league team and acclimated to hitting and pitching is more likely than not to be beneficial, but in terms of the performances we’ve seen this season, missing spring doesn’t seem to have been a major impediment to players performing up to the same expectations as their fellow free agents.

Craig Edwards can be found on twitter @craigjedwards.

Newest Most Voted
Inline Feedbacks
View all comments
5 years ago

I’m begging you: learn how to use a student’s t-test

5 years ago
Reply to  000031

Been a while since I calculated a t-stat, but I get -8 wrc+, so Craig’s estimate of significance of 15 wrc+ seems to be a good measure.

I know the league has only played ~62 games so far so small sample caveats, but Craig’s analysis shows that 14 of the 22 hitters (8 of 12 significantly) and 7 of the 12 pitchers are failing to meet expectations. There’s plenty of time this season for most of them to turn things around, but this is not a good sign for those teams who gave free agents lagging their expected results multi-year contracts. If they have a down year this year, it’s less likely that the FA’s will get better in subsequent years as they age.

Maybe Craig could do this analysis for past years as well, but here’s more evidence that FA is a bad way to build a playoff team and that teams are right to be reluctant in signing FA’s.