Do players outperform their projections in walk years?

Every spring, when baseball pundits begin making their predictions for the season, there's one statement that we seem to hear over and over, so much so that it feels like it has become part of the game's "conventional wisdom."

"Player X is going to have a great year because he is a free agent after the season."

The thinking behind this statement is that players are extra motivated to do well in their final year before free agency (also known as their walk year), since a good season can lead to extra millions of dollars in free agency. With extra motivation, players could theoretically perform even better than before (at least for one season), just in time to earn a nice contract.

I'm sure that all of us can think of anecdotal evidence of players having career years right before free agency. Perhaps these players worked extra hard throughout the year, beginning with them showing up to Spring Training in the best shape of their lives. Often, we assign narratives such as this to explain unusual performances, but in reality, we usually have no idea if our reasoning is correct. After all, players have career years at other times as well, and we cannot point to their upcoming free agent status as the reason for their breakout.

With that in mind, I thought I would test this piece of conventional wisdom by looking at this year's upcoming free agents to see if there is a pattern of players performing better than expected. Previous research on this topic appears to be inconclusive, as Jake Fisher concluded that players (hitters in particular) have a statistically significant peak in production during their walk year, while Phil Birnbaum found no statistically significant difference between walk year and non walk year production. As each of these writers has shown, there are a few different ways to examine this issue. I may eventually expand my analysis to include several years of data, but for now, I thought I would start small by looking at the latest data available from this year's upcoming free agents.

To set a baseline for performance, I decided to use preseason projections. Projections are emotionless by nature and do not take factors like free agent status into account, other than indirectly through their recognition of the number of seasons a player has been in the majors. But of course, they don't know if the player has signed an extension or anything like that. They also adjust for age and regress unsustainable performances. As a result, projections are very helpful for knowing what we can reasonably expect from a player in any given year and what would constitute exceeding expectations.

I looked at data from all players who are set to be free agents after this season, with the exception of those players who were free agents last season and signed one year deals. I decided to exclude these players from this investigation, since being in a walk year seems to imply that the year in question is different from prior years. If a player is signing one-year deals, then he is in the same contract situation each time and will not have any extra motivation from one year to the next.

I also excluded players who have club options for next year, since those players do not have the same opportunity to sign a big contract in free agency if they have a good season. However, I did include Alex Gordon in my sample, since he has a player option which he can choose to decline. In theory, he has the extra motivation that an impending free agent would have, since he can decline the option and sign a big free agent contract if he has a good year.

Once these changes were made, I had a group of 76 players total who are set to be free agents after this season. I compared each player's preseason projected WAR (from ZiPS) to their actual fWAR at this point in the season (with rest of season ZiPS projections added in).

Of the 76 players who are currently in their free agent walk year, 29 overperformed their projections, 44 underperformed their projections, and three hit their projections right on. One one end of the spectrum, players like Yoenis Cespedes and David Price are on pace to outperform their preseason projections by just over two wins. On the flip side, Matt Joyce and Ian Desmond have had rough years and could fall nearly three wins short of their preseason projections. On average, players were underperforming their projections by a quarter of a win.

These results are relatively surprising, since we would expect no significant overperformance or underperformance from a group composed of a random sample of players. It's possible that injuries could be a big factor in the underperformance of several players. I decided to see what the results would look like if I excluded players who had spent time on the DL this season. Predictably, the number of underperformers and the average level of underperformance decreased as a result. Even so, there were more slightly more underperformers than overperformers, and the players in the sample were, on average, still underperforming their projections by around .1 WAR.

(It is also important to keep in mind that projection systems account for injury history and do their best to estimate the amount of playing time each individual player will get. Some players get injured and play less than they are projected to, but some players who have been injured in the past may overperform their projection simply by being healthy. In theory, these factors should balance out, which means that excluding only injured players from this analysis, as I did above, could skew the results.)

* * *

While this is only a preliminary look at a relatively small group of walk year free agents, I think these results can at least make us a little skeptical of the idea that players tend to do better in walk years. I understand the psychology behind this idea, as it is human nature to work extra hard when such a huge reward is on the line. However, free agent walk years are not the only times that players can increase the amount of money they will make. Players who perform well prior to free agency can make a good amount of money in arbitration. Teams also value consistency from players when they sign them as free agents. While it always helps to have a great season just prior to free agency, having more than one great season prior to free agency is even better, as teams will be more willing to invest in a player who has shown that he can repeat his level of success. While having a good year at the right time can lead to an extra several million dollars in some cases, major league players will almost always see some sort of financial benefit to performing well, regardless of whether they are an impending free agent or not.

Major league players are the best of the best at what they do and have already had to put in an incredible amount of work just to get to the level that they are at. While there are many reasons why players sometimes exceed expectations (adding a new pitch, adjusting one's swing, etc.), I find it hard to believe that players can improve by simply flipping a switch and trying harder because they are motivated by their contract status.

. . .

Nick Lampe is a contributor at Beyond the Box Score and Viva el Birdos. He is in his walk year of college and is unsure how this will affect his level of productivity. You can follow him on Twitter at @NickLampe1.