Relevance of First and Second Half Stats

Coming into the 2014 season, Jean Segura’s 2013 production still baffles me (I have two more pieces on him coming up in the future). He had a an amazing first half (11 HR, 27 SB, .325 AVG) and then his production dropped off in the second half (1 HR, 17 SB, .241 AVG) which no obvious signs of change. Other players showed the same decline like Chris Davis (.461 wOBA to .365) and Jason Kipnis (.387 wOBA to .317).  Players exist on the other end of the spectrum. They seemed to get better in the second half of the season. Elvis Andrus went from hitting .242 with no home runs to .313 with four dingers.

Having more data is always better. Using three to fours years worth of data make the best projections. It is just human nature to remember the most recent results and quickly react off of them. Whether it is fantasy baseball or the economy, people jump immediately to action. I will look at probably the smallest set of data (besides pitch velocity) which fantasy expects use: half season data.

I will start with the easy answer: No, first and second half stats are not as important the entire season for pitchers and hitters.

If you have a real life and want to get on reading about the Astros outstanding bullpen, go on. For the two or three others, here is a longer math filled response with a little more grey area.

I took a player’s 1st half (1H) and 2nd half (2H) stats and full season season data. Then I calculated the r-square (a value of 0 means no correlation and a value of 1 is perfect correlation) for the three data sets compared to the next season’s stats. I will start with the hitters who I set a min of 200 PA in each full season.

AVG OBP SLG ISO K% BB% BABIP
Season 1 to Season 2 0.195 0.339 0.352 0.504 0.754 0.593 0.136
1H Season 1 to Season 2 0.109 0.237 0.239 0.393 0.680 0.481 0.075
2H Season 1 to Season 2 0.136 0.253 0.266 0.412 0.688 0.520 0.089

The table has two spectrums. On one end is Plate discipline (K% and BB%) which correlates nicely with both full and half season data. BABIP, which barely correlates, is on the other end. All the other other stats are some combination of the two. AVG, which is heavily BABIP driven doesn’t correlate as good as OBP which includes a player’s walk rate.

You Aren't a FanGraphs Member
It looks like you aren't yet a FanGraphs Member (or aren't logged in). We aren't mad, just disappointed.
We get it. You want to read this article. But before we let you get back to it, we'd like to point out a few of the good reasons why you should become a Member.
1. Ad Free viewing! We won't bug you with this ad, or any other.
2. Unlimited articles! Non-Members only get to read 10 free articles a month. Members never get cut off.
3. Dark mode and Classic mode!
4. Custom player page dashboards! Choose the player cards you want, in the order you want them.
5. One-click data exports! Export our projections and leaderboards for your personal projects.
6. Remove the photos on the home page! (Honestly, this doesn't sound so great to us, but some people wanted it, and we like to give our Members what they want.)
7. Even more Steamer projections! We have handedness, percentile, and context neutral projections available for Members only.
8. Get FanGraphs Walk-Off, a customized year end review! Find out exactly how you used FanGraphs this year, and how that compares to other Members. Don't be a victim of FOMO.
9. A weekly mailbag column, exclusively for Members.
10. Help support FanGraphs and our entire staff! Our Members provide us with critical resources to improve the site and deliver new features!
We hope you'll consider a Membership today, for yourself or as a gift! And we realize this has been an awfully long sales pitch, so we've also removed all the other ads in this article. We didn't want to overdo it.

Here is a look at Kipnis’s and Segura’s 2013 seasons which saw 2H dropoffs. Both of them saw a drop in BABIP with Kipnis’s going from .351 to .326 while Segura’s went from .349 to .285. A decent drop for each hitter. With BABIP taking so long to correlate, it is tough to take much credence in either value. Now looking at the plate discipline stats, Segura had a higher strikeout percentage in the season’s 2nd half.

Kipnis: BB%, K%
1H: 12%, 22%
2H: 11%, 21%

Segura: BB%, K%
1H: 4%, 12%
2H: 4%, 16%

Kipnis barely saw his eye numbers barely improve while Segura saw a decent jump in his strike out rate. While I should use the complete 2014 stats, I would worry more about Segura bouncing back from his 2H troubles because of the increased strikeout rate jump which stabilizes quickly.

AGain, use the yearly stats over 1H or 2H stats, but a person may need to take some credence in 2H plate discipline based stats (K%, BB%).

Now onto the pitchers. For them, I used 40 IP as my cut off point and here are the results.

WHIP ERA FIP K/9 BB/9 HR/9 BABIP
Season 1 to Season 2 0.158 0.112 0.223 0.549 0.315 0.080 0.030
1H Season 1 to Season 2 0.093 0.063 0.134 0.448 0.235 0.036 0.013
2H Season 1 to Season 2 0.106 0.070 0.176 0.489 0.224 0.056 0.021

Pitcher plate discipline stats, like the ones for hitters, correlate the best with K% being the only one over 0.5 and BB the net highest. Let’s look at two pitchers, R.A. Dickey and David Price in 2013. Both struggled in the 1H and then turned their season around in the 2H.

Dickey
1H ERA: 4.69
2H ERA: 3.56

Price
1H ERA: 3.94
2H ERA: 2.87

Now here is how they each made their improvement.

Dickey
Half: K%, BB%, HR/9, BABIP
First: 17%, 9%, 1.4, .260
Second: 22%, 6%, 1.4, 271

Price
Half: K%, BB%, HR/9, BABIP
First: 21%, 4%, 1.0, .322
Second: 20%, 3%, 0.6, .281

Dickey’s 2H improvement is based on plate discipline while Price’s is based on batted ball distribution. People, should use their full season stats to determine their overall value, but I may give Dickey’s late season improvement a bit more credit.

Just remember doing the upcoming fantasy season, people are more likely to remember and act on the last set of data the most which includes first and second half stats. With the exception of plate discipline stats, it is best ignore any first and second half centered data and and instead look at data from a full season or more.





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.

17 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Richie
12 years ago

How dare you say I have no life!?! (hurts a lot more when it’s true, so be more careful in those situation-types from now on)

rotorankers.tumblr.com
12 years ago

Great stuff. Segura and Kipnis are MY BOYS and this article will help me in my daily arguments about them.

Jared
11 years ago

How did this article on Kipnis and Segura work out for you in 2014?

Detroit Michael
12 years ago

Both the batter and pitcher tables indicate that for every statistic (1) second half performance correlates better with next year performance than first half and (2) whole year performance correlates better than second half performance. That’s not very surprising, at least to me.

Isn’t the real question whether second half performance provides any useful prediction information beyond merely looking at whole year performance? If so, how much? Seems like a regression analysis might answer those questions.

Chris
12 years ago

Exactly.

And what’s interesting is that the data in the article suggests than second-half splits are probably the *least* useful for hitter K%.

1H K% correlates almost as well as 2H with future performance (.680 v .686). That says to me that once you have full-season data for K% there is no point in looking at 1H/2H splits. Jeff, what leads you to conclude otherwise?

RainmakerMember since 2016
12 years ago

Agree with Michael above. But a few other points:

First, I think its safe to say that looking at all player stats in the aggregate generally doesn’t yield particularly interesting results. In order to really get at something interesting, you probably need to look at a cross-section, set off with particular factors, because we’re not dealing with random samples here. For example, Kipnis came into 2013, with contact and power skills, just a tick above league average. In terms of game planning the Indian, I don’t think many pitchers/scouts were spending a lot of time studying Kipnis. Fast forward to the ASB, where Kipnis has 13 HRs, a .213 ISO, and .301 AVG, and stopping Kipnis is suddenly key #1 to beating the Tribe. Hence, teams react and you can see in Kipnis’ Pitch f/x data — teams expand the zone more on him, and found a weakness away with him.

Second, looking at any stats in a vacuum doesn’t in itself make for a predictive factor. I get the line gets fuzzy when you open statistical analysis to subjective considerations. Matt Cain for example put up a 5.06 ERA in 1st half only to put up a 2.36 ERA in the second; when you consider his career baseline talent level, its more likely that the his 1H was the outlier not his 2H, right? But a statistical analysis in vacuum, disconnected from that baseline doesn’t account for that baseline.

RainmakerMember since 2016
12 years ago
Reply to  Jeff Zimmerman

Yeah, that’s just silly talk.

Just something that popped into my head, but there could be a bias in scouting that works for “fast-starter” types that typical have big first halves. I’d always say that is the explanatory factor behind the enigma that is Dexter Fowler — the guy absolutely hammers anything in. And it seems like early in the season, maybe scouting reports start fresh for the new season and teams make the mistake of pitching him in; then he goes on a tear, puts up massive April/May, scouting adjusts by June and he fades.

It seems like a real outcome, but I’d hardly bet on it…

lester bangs
12 years ago
Reply to  Jeff Zimmerman

Geesh. Strawman.

Joe m
12 years ago

For a guy like segura you have to point out that he is a rookie and rookie that got very hot. Once he got hot, every pitcher was forced to look at video to find his weaknesses and usually a rookie has weaknesses especially based on segura’s minor league numbers. It will be interesting to see if he can make the adjustments this year.

kipmonk
12 years ago

Kipnis wOBA went from .387 to .317 and his BABIP went from .361 to .326. You made errors at the top and half way through. Good piece though

kipmonk
12 years ago

I see now that your wOBA numbers were from 2012 on Kipnis, which I’m guessing isn’t what you intended

Malon
12 years ago

I think you typo’d your thesis statement.

Doug Anderson
12 years ago

I agree that 2nd half stats can be ignored for established players. Even legendary 2nd half player Adam LaRoche failed us in 2013. I do think there is something to take from them though when we’re looking at young players. Did the player figure things out and get comfortable in the 2nd half? Did pitchers figure out a weakness in a young hitter who tore up the 1st half?

lester bangs
12 years ago
Reply to  Doug Anderson

I was going to say the same thing. Agreed. At least, let’s study this specific thing.

LRG
12 years ago

The 2nd half numbers of Segura don’t concern me as much as the first half numbers intrigue me. Yeah pitchers began adjusting to him and his production tailed off in 2nd half but he did show that he is capable of being an absolute beast. It’s nice to see that in a rookie. If he got called up in June and dominated thru the end of the year, everyone would be talking about him as being the next big thing.