Checking in on late-signing free agents

Picture the man or woman of your dreams. Now imagine that beautiful, paragon of a human standing on the altar beside you, gazing lovingly into your eyes. All your friends and relations are there to witness the happiest day of your life. But instead of saying, “I do,” your soulmate simply smiles, turns, and walks out of your life forever. At that same moment, your boss crashes the ceremony just to fire you. The caterer under-cooked the chicken, and now half your loved ones have food poisoning. Also, your chinchilla died.

The good news for Greg Holland is that it’s only his career that’s gone sideways, not his entire life (hopefully). Not too long ago, a league leader in saves and games finished would have struck oil in free agency. In fact, Mark Melancon had recent black ink in those exact categories when he became a free agent following the 2016 season. He signed early in the free agent period, inking a four-year, $57 million contract with San Francisco on December 5, 2016.

Holland and Melancon aren’t perfectly analogous, but the latter’s contract should’ve been a good gauge for the former’s. MLB Trade Rumors projected Holland to sign for four years, $50 million after he rejected his $15 million option and $17.4 million qualifying offer. Of course, the market didn’t materialize the way it had in the past. The entire off season went by, as did Spring Training. On March 31, 2018, Holland signed the final major league free agent contract of the off season, a $14 million single-season deal with the Cardinals.

The injury of losing out on millions of dollars was further insulted. Since joining the team on April 9th, Holland has looked nothing like his former self. He’s walked 11 of 55 batters faced this season, while striking out only eight. His xFIP is a ghastly 6.82, nearly three runs worse than last year. His average fastball velocity dropped from 93.8 to 91.7 mph. If he owns a chinchilla, it’s time for a wellness check.

Naturally, the easy thing to do is blame the lack of a Spring Training. After all, he didn’t sign until the regular season was already underway. In Holland’s case maybe that’s true; no one really knows except for him, but he wasn’t the only free agent to sign late into the off season. Now that we’ve had a month and a half of regular season baseball, let’s see whether signing late had any impact on this free agency class.

Analyzing the Free Agents

Doug Fister signed the first major league contract of the winter on November 28, 2017. We’ll call this Day One of free agency. Holland was the last player to sign, of course, putting pen to paper on Day 124. More than 70 other players came to terms somewhere in between.

Not all free agents are equal, so it wouldn’t be fair to compare them to each other. However, we can compare each free agent to himself. For batters, we’ll look at the difference in wRC+ from this year to last year. For pitchers, we’ll do the same with xFIP.

In general, most players regress from one season to the next. Obviously some improve, but most go the other direction. What we’re looking for here is a correlation between signing date and rate of regression.

One note on this: it’s surprisingly difficult to pin down a good list of free agent signings. The best three are found at FanGraphs, MLBTR, and MLB.com. However, these trackers each omit certain players for whatever reason, and disagree on signing dates. At any rate, we’ve got a decent enough sample of players to use for comparing this season to last.

Results

Much like Greg Holland, there are anecdotal examples across baseball of players who “suffered from a shortened spring.” Neil Walker signed with the Yankees on Day 105 of free agency. His wRC+ dropped from 114 last year to 54 in 2018. However, Jonathan Lucroy signed the same day and boosted himself from 82 to 110.

All in all, batters did show a correlation of 0.2 between later signing date and increased regression. That’s a pretty small correlation score, but it’s enough that there could be some fire with the smoke.

Pitchers were a different story. For every Greg Holland there’s a Trevor Cahill, who’s xFIP improved from 4.16 to 2.27 despite signing on Day 110. The correlation between xFIP regression and signing date was actually -0.2. In other words, pitchers performed better the later they signed!

This doesn’t really make any sense. The shortened spring was supposed to be worse for pitchers than batters. Yet the late-signing pitchers are outperforming both their early-signing counterparts as well as the late-signing hitters.

The total correlation score of hitters and pitchers combined is -0.06. This tells a more accurate story, which is that there’s no story at all. Sure, there might be individual cases like Holland or Walker that might have been hurt by a later signing (or maybe not!), but there’s no evidence it impacted the free agent class as a whole.

Of course, the financial impact is another matter entirely. But that’s not a regular season conversation. The baseball world will surely revisit that topic next winter when Holland becomes a free agent once again.