Your Annual Spring Training Stat Reminder

Spring training games start on Wednesday. There are still several openings left on the roster which players will be fighting for. After a long winter, there’s also the desire to create stories where they might not otherwise exist. That means we’ll be hearing a lot about spring training statistics over the next few weeks.

Of course, there’s a significant problem with this: spring training statistics do not matter*. I ran through some previous seasons of batting statistics and pitching statistics before last year’s spring training started. Eric Stephen also recently pointed to the fact that spring training record isn’t important. Rather than completely rehash those posts, let’s go through last year’s results specifically.

First, the hitters. Qualifiers for this list had at least 30 plate appearances. Here are the top 4 and bottom 4 offensive performers from last year’s spring training, sorted by OPS:

	Spring Training				Regular Season
Player	PA	BA	OBP	SLG	PA	BA	OBP	SLG
Justin Turner	42	.389	.477	.528	322	.340	.404	.493
Alex Guerrero	35	.300	.400	.500	13	.077	.077	.077
Andre Ethier	48	.333	.417	.476	380	.249	.322	.370
Miguel Rojas	34	.387	.441	.452	162	.181	.242	.221

	Spring Training				Regular Season
Player	PA	BA	OBP	SLG	PA	BA	OBP	SLG
Tim Federowicz	35	.067	.200	.100	78	.113	.158	.197
Yasiel Puig	49	.167	.173	.229	640	.296	.382	.480
A.J. Ellis	45	.158	.289	.158	347	.191	.323	.254
Carl Crawford	46	.154	.283	.179	370	.300	.339	.429

Some of these almost look predictive. Justin Turner showed up at the top of the best performances list before his break-out season. Tim Federowicz and A.J. Ellis both showed up on the worst performances list before having difficult offensive seasons. However, what about Rojas showing up on the top table? Ethier had a good spring before a major down year at the plate.

On the poor performers side, Puig shows up with a .402 OPS, right before having a five win season. Remember when people were trying to use his spring performance from 2013 to say that he was major-league ready? Was last year’s performance taken as evidence that he wasn’t?

Onto the pitchers. This one is a bit more difficult to quantify, since last year’s spring training was so abbreviated by the Australia trip. Only five pitchers threw more than ten innings for the Dodgers in spring training last year. Here they are, listed by ERA with peripheral stats added.

	Spring Training				Regular Season
Player	IP	ERA	K/9	BB/9	IP	ERA	K/9	BB/9
Red Patterson	11-2/3	0.77	5.40	3.09	4-2/3	1.93	1.93	5.79
Seth Rosin	11	1.64	9.82	1.64	4	6.75	6.75	2.25
Hyun-jin Ryu	16-1/3	2.20	5.51	1.65	152	3.38	8.23	1.72
Dan Haren	12	6.00	3.75	2.25	186	4.02	7.02	1.74
Clayton Kershaw	14-2/3	9.20	6.13	3.07	198-1/3	1.77	10.85	1.41

These are all over the place. Patterson had a good ERA but his spring peripherals weren’t very good. Rosin had both the ERA and peripherals, but spent most of the season struggling in the Phillies’ minor league system. Kershaw had poor peripherals and an unthinkably bad ERA, then he went on to win the Cy Young award and MVP.

So, there you have it. The annual reminder. When somebody points out that Andre Ethier is having a better spring than Joc Pederson (or something similar), keep these numbers in mind. There are many factors which would lead to roster decisions (and, with Pederson, it could very well be service time games). The actual numbers a player produces during spring training shouldn’t be the main one.

—

*Caveat: Spring stats may matter, a little bit. A presentation at yesterday’s Sloan Analytics Conference in Boston showed a bit of importance to the stats. Dan Rosenheck of The Economist gave the presentation, and you can review his slides here. A few of the main points:

• What we already know (projection systems) still matters
• Statistics which don’t stabilize quickly still aren’t important. ERA, BABIP, and such aren’t trustworthy.

I do have some issues with the conclusions he draws (why only compare to one projection system? why use outdated stats like qERA? missing info?) but it’s an alternate point of view. At the very least, one point remains the same: you have to be extremely careful when reviewing spring training stats. I still wouldn’t cite them as the only source of information, and I’d trust a well-informed scout a lot more.