Do Certain Hitter Profiles Increase Year-To-Year Consistency?
As for now, I can’t find any find any predictability to year-to-year hitter consistency once adjusting for plate appearances. For the readers looking for a short article, stop now and move on to Paul’s thesis on starting pitchers. For the stubborn ones, here is what I’ve additionally found out after previously investigating the subject.
On Monday, I could not find any predictability for hitter being consistency. That is not entirely true, I did find that the more plate appearances a hitter accumulates, the more likely they are to reach their true talent level. And if given the opportunity to be closer to their talent level, the more consistent their output.
The one factor I thought might point to year-to-year consistent play in a player’s statistical profile. Are power hitters inconsistent because a few gusts of wind could make a difference in a half-dozen home runs? Do high-walk hitters see their stats as being more consistent since walks stabilize faster? Basically, are certain hitter types more consistent on a yearly basis.
To measure of year-to-year consistency, I will again use the standard deviation of OPS. Using a three-year sample (min 100 PA in each season), I compared the constancy value to several metrics. Here are the r-squared values for players with the three-year average number of PA listed.
| PA/R-squared | BB% | K% | Contact% | GB% | ISO | PA |
|---|---|---|---|---|---|---|
| >100 | .003 | .036 | .025 | .012 | .011 | .063 |
| >250 | .003 | .034 | .024 | .012 | .013 | .053 |
| >400 | .011 | .025 | .019 | .013 | .023 | .031 |
| >500 | .016 | .013 | .011 | .016 | .015 | .016 |
| >600 | .000 | .032 | .020 | .000 | .000 | .020 |
| 100 to 300 | .003 | .071 | .056 | .018 | .000 | .003 |
| 300 to 500 | .003 | .033 | .028 | .019 | .044 | .011 |
The largest r-squared in the whole table is .071 with several zeros scattered around. Not for being related.
The truth is that as a player accumulates more and more plate appearances, they get closer to their true talent. As Bill Petti found at the Hardball Times, day to day volatility is dependant on the type of hitter.
Hitters that tend to hit the ball in the air for power tend to produce in a more volatile fashion, while groundball hitters with higher on-base skills appear to produce more closely to their average on a daily basis.
His concept makes sense with some hitter’s power coming in bunches over a whole season. But once a hitter gets to 600 PA for several seasons in a row, we know who we think they are.
Even the most stubborn stats (e.g. BABIP) begin to stabilize within a few seasons and a hitter’s talent level is revealed. If an owner wants a reliable hitter, find one with the most plate apperances in the past three or four seasons.
I figured I’d find some consistency like Petti saw. I was wrong. For those owners looking for some level of year-to-year consistency, start at the top of a multi-year plate appearance leaderboard.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
Wouldn’t the obvious thing be health track record? If you are on the dl you can’t produce and even if you are only like 80% fit or slightly hurting that might hurt too.
Of course health is no guarantee for future health but isn’t a bad health record somewhat predictive (at least of games played and dl stints, we can’t see if a player is playing though something)?
Have you checked on That? How predictive are past injuries for future missed games and underperformance?
Could the Pa thing be a measurement for good health?
“Hitters that tend to hit the ball in the air for power tend to produce in a more volatile fashion, while groundball hitters with higher on-base skills appear to produce more closely to their average on a daily basis.”
Isn’t this the mother of all no-brainers though?
Elite home run hitters are still hitting homers at a rate of one every three to four games (40-50 HR pace). What are they doing the other days of the week? Not hitting home runs, and probably, since they are fly ball power guys, striking out frequently.
The HR is also a higher variance play. It is more runs compressed in a single play so a few less homers hit means a whole bunch of runs less while a couple less singles have less of an impact of run production.
But it is of course both ways, the hr guy also has more upside, a few bombs more and a whole bunch of extra runs are scored.
Most of the work I’ve seen on “consistency” assumes that “true talent” is a stable concept. Give a guy enough PAs and he will be what he will be. But I haven’t seen much that takes into account changes in “true talent”. The most obvious example of a true talent level change is playing with injury. For example, is Miggy’s “true talent” a 100 wRC+ or a 150 wRC+?, doesn’t it vary? In other words, instead of there being 1 Miggy, who will eventually always end up with a 142 wRC+ if given enough time, aren’t there 2 Miggys? 1 of whom will always put up a 100 wRC+ over a large enough sample and one who will put up a 150 wRC+ over a large enough sample? If we look at it this way, then consistency isn’t just, who gets the most time, but rather, who has the fewest or most similar different versions of themselves? Who will be hurt the least is certainly a large part of this, which factors into both playing time and rate performance, but I’d like to see an attempt at the other factors here. Unfortunately, I assume this will require a granular PA by PA analysis in an attempt to identify behavioral changes on a micro level.