Can Spring Training Statistics Predict Breakout Power Seasons?

March 25, 2014
 

Over the last 10 or so seasons, this has been the time of the year when we’ve listed a group of players we expect to see a power breakout in the upcoming regular season. That list is based on a study of players from 1996-2001 that saw players who saw an increase in their spring slugging percentage of at least 100 points over their career slugging percentages entering the season then increase their slugging percentage in the regular season relative to their previous career levels.

Here are the results of that original study:

Spring Slugging Percentages (1996-2001)
(minimum 50 spring, 250 regular season,
and 750 previous career at-bats)

Spring 100+ Points Better
Than Previous Career
Regular Season Better
Than Previous Career
Pct
307 220 71.7

 

Based on our qualifications, 71.7 percent of players who enjoyed those spring power surges went on to increased power in the regular season. Moreover, a similar study of that trend on the team level showed similar results.

In recent seasons, we’ve started to see some feedback from independent studies suggesting that trend may no longer hold true. We were curious to investigate, so we ran the same study with the same qualifications on the data from the most recent five seasons, 2009-2013. Here are those results:

 

Spring Slugging Percentages (2009-2013)
(minimum 50 spring, 250 regular season,
and 750 previous career at-bats)

Spring 100+ Points Better
Than Previous Career
Regular Season Better
Than Previous Career
Pct
192 70 36.5

 

What a stark turnaround! For me, the initial study seemed to show a clear trend that indicated spring slugging spikes likely predicted regular season power increases. The same study on recent seasons shows the exact opposite results that, if anything, suggest spring training power spikes indicate a likely decrease in power in the regular season.

In light of the dramatic difference in results between the two samples, I will of course not recommend potential power breakouts based on hitters showing less power this spring. At this point, I’m leaning toward an opinion that the only thing spring stats might predict are which young players may earn playing time.

That said, I can think of reason why the original study and the new study both might have held predictive significance in their times rather than simply showing two small-sample blips that would regress over larger samples. The original study was run from 1996-2001, the height of the steroids era. It was not unusual for players who had never shown power tendencies to suddenly start hitting for power. It stands to reason those changes would show up in spring, following offseasons when players would have had an opportunity to train. Meanwhile, the current study from 2009-2013 is in the presumed post-steroids era, when power across the sport has been on the decline. It will be interesting to continue to track this data in future seasons.

 
 

COMMENTS (8 Comments, most recent shown first)

flyingfish
Seems to me people often say that in spring training, they are "working on something." If that's true, then you wouldn't expect much correlation between spring stats and regular season stats. On the other hand, you would expect Miguel Cabrera to demolish baseballs at any time of the year....Interesting, though, and thanks for reporting on it.
3:49 PM Apr 3rd
 
doncoffin
Old Backstop--Little value, yeah. No value at all, well, that's probably too strong a statement. Spring training stats are (guessing) probably 80% noise and 20% signal.
9:09 PM Apr 1st
 
OldBackstop
Meaning to say, I think all those factors, plus things like who gets to go to what blow out split squad games, would lead to very inconsistent data, which it apparently has. I think the coaches are looking at the players' abilities and the fans are looking at the numbers.
10:47 AM Mar 30th
 
OldBackstop
Am I stupid or....wait, let me start again. Aren't spring training offensive stats totally whacked from the kiddie pitchers getting their shot and the split squad games and the smaller fields, etc?

It seems to me that slugging .570 by bouncing the ball off the 338 Hooters sign at Tradition Field against Journeyman Jim and an 18 year old prospect and Joe Bagadonuts who is trying to develop a splitfinger would have small relation to Stephen Strasburg tomorrow afternoon at Citi.
10:42 AM Mar 30th
 
MarisFan61
(Sorry, have to clarify again -- we really need an edit function here!
Or at least I do. :-)
In the post below, that first portion was regarding what John said for the earlier period -- i.e. the spring trainings and regular seasons from 1996-2001.) The second portion was regarding the later period, 2009-2013.)
1:22 AM Mar 28th
 
MarisFan61
John: Regarding your proposed explanation at the end, I easily understand the first part, i.e. why the increased power in the spring might well have been predictive for the coming season.

But this:
"Meanwhile, the current study from 2009-2013 is in the presumed post-steroids era, when power across the sport has been on the decline" -- I don't see why this would at all explain an inverse correlation between spring and regular season. No correlation, sure; but INVERSE? It seemed like you felt there was something intuitive or obvious in there, but I don't see it.

I could take wild guesses, mainly something like "Home Run Derby theory": that in the spring they either wore themselves out, or "changed their swing" to try for more power and it worked briefly but then what they were left with was a messed-up stroke. But, just wild guesses. I do believe them about Home Run Derby (and please nobody start showing me data that this doesn't happen; I'm talking about whatever guys for whom it does happen), but I'd have trouble thinking that it would apply for spring training.
1:18 AM Mar 28th
 
doncoffin
I just looked at the relationship between spring training stats and regular season stats using 2013 data. What I found was that spring training isolated power was more closely correlated with regular season ISO than was the case for BA, OBA, SLG, or OPS.
(http://isolatedpower.blogspot.com/2014/03/do-spring-training-stats-predict.html)
8:35 PM Mar 25th
 
rgregory1956

Sometimes, it's nice to know that what we'd expect to see actually shows up in the data. It's fun to find "new" things, but it's also comforting when our long-held suppositions are confirmed.

7:08 PM Mar 25th
 
 
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy