Justin Upton, the Braves, and their Strikeouts

As of Thursday, an awful lot of people are happy. The Diamondbacks are happy, because they were able to move Justin Upton for value. The Braves are happy, because they were able to get Justin Upton for a good but not outstanding package. And Justin Upton is happy, because he gets to go play in Atlanta, instead of Seattle or somewhere else. Upton was hoping for a destination like this, if not this destination specifically, and now he can put his alleged malcontent days behind him and play alongside his brother. This is a move that felt like it had to be made, and now it’s been made, officially.

But from the Braves’ perspective, there are not zero concerns. For one thing, now they’re down a Martin Prado, and their third-base situation could be stronger than it is. And then there’s the matter of Upton, and the strikeouts that he brings. He’ll share an outfield with B.J. Upton, who strikes out, and Jason Heyward, who strikes out too. There’s been some degree of concern that the new Braves might strike out too often, which means there’s concern that the Braves’ offense might underachieve.

Immediately, one should note that Justin Upton isn’t exactly strikeout-prone. His strikeout rate the last two years is about 19%. The league-average strikeout rate the last two years is about 19%. Upton shows a lot of strikeouts if you look at his strikeouts as a counting stat, but that’s because he gets a lot of plate appearances, and all we should really care about are rates. Other people have struck out way more often than Justin Upton has.

But we can look at the Braves, and strikeouts, and runs scored, anyway. Don’t forget that the Braves have Dan Uggla, and at present they stand to play a lot of Juan Francisco. There’ll be whiffs. If the Braves’ offense were to underachieve on account of its strikeouts, we’d think their runs scored wouldn’t match up with their overall team wOBA. First, let’s examine the recent relationship between wOBA and team runs scored. We’re doing this very simply: raw runs vs. wOBA, 1996-2012, post-strike and eliminating the abbreviated 1995 year. We have a sample size of 506 individual team seasons.

runswOBA

If we wanted, we could get a little more detailed, but this does well enough. There is, obviously, a very strong relationship between wOBA and team runs scored. Nothing about that should come as a surprise. Teams fall on either side of the best-fit line, but no one’s that far off. There’s never been any such thing as a team with a productive offense and a bad wOBA, or a team with an unproductive offense and a good wOBA. Post a decent wOBA and the runs scored will follow.

But, all right, let’s examine now the difference between runs and “expected runs”, where expected runs are calculated using the best-fit line equation. There’s a suspicion that the Braves might score fewer runs than another team with an identical wOBA, because of the strikeouts. Here’s the difference between runs and expected runs, plotted against team strikeout rate:

diffstrikeouts

There’s hardly any relationship at all. There’s a very slight downward slope, as strikeout rate increases, but it’s practically meaningless. There’s no evidence here that a team with a high strikeout rate will underachieve to a great degree. We can look at things another way. Over the window, 50 teams posted a strikeout rate of at least 20%. They averaged 1.3 more runs scored than expected runs scored. Meanwhile, 57 teams posted a strikeout rate below 15%. They averaged 6.3 more runs scored than expected runs scored. If there’s an effect here, it’s so small that it barely even matters. Strikeouts are not offensively crippling.

You know what shares a better relationship with the difference between runs and expected runs? Baserunning. Which isn’t a shock, since baserunning matters, but it isn’t factored into wOBA. Here’s the difference between runs and expected runs, plotted against baserunning value:

diffbaserunning

It’s still a fairly weak relationship, but it’s a far stronger relationship. The better your team baserunning, the more likely it is you outperform your team wOBA. We can also examine this in a table, breaking the teams into four ~even groups:

Group R-ExpR R ExpR K% wOBA BsR
1 33 799 767 17% 0.329 2.4
2 8 773 765 17% 0.329 1.4
3 -9 754 763 18% 0.328 -0.5
4 -31 731 762 17% 0.328 -3.3

The groups are in descending order of difference between runs and expected runs. By wOBA, the groups are virtually identical, and by strikeout rate, they’re virtually identical, too. We see a small pattern in the final column, though, corresponding to baserunning. So there’s some effect there, even if a lot of the difference between runs and expected runs can probably be chalked up to luck, or, if you prefer, unsustainable timing.

Why does this matter for the Braves? Justin Upton, historically has been an above-average baserunner. So has B.J. Upton. To whatever extent the Uptons might squander the occasional productive-out situation, they’ll also contribute on the basepaths, presumably offsetting any effect and maybe then some. We know that baserunning matters, here. More in one direction than strikeouts seem to matter in the other direction.

The Braves, as a team, are probably going to post a strikeout rate that’s above the league average. But whatever their team wOBA turns out to be, the runs scored will follow, and the strikeouts won’t make much of a difference. At times, the Braves’ strikeouts might be frustrating, but they’re still going to score, and they’re still likely to be one of the very best teams in the league.




Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.

49 Responses to “Justin Upton, the Braves, and their Strikeouts”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Forrest Gumption says:

    The Oakland A’s hitters led MLB in strikeouts last year and made the playoffs, that’s all I’ll ever need to think about on the topic.

    Vote -1 Vote +1

    • Mark says:

      They didn’t just lead MLB, they trashed the AL record (although 2013 A’s will have their say soon).

      Vote -1 Vote +1

    • Tom says:

      And on the flip side the 3 AL teams with the fewest strikeouts all lost at least 90 games and scored below the AL average. KC, Min, Cle.

      Vote -1 Vote +1

  2. Jesse says:

    Reading the title of this article, I clicked on it getting ready to troll the notion that strikeouts mattered. However, I was pleasantly surprised by this article, nicely done, thank you!

    Vote -1 Vote +1

    • Balthazar says:

      ‘Preciate the post Jeff, and I found the baserunning evidence particularly interesting. And not factored into wOBA? Time for a re-forumulation there. I paricularly find interesting questions of the “How does X beat projections?” sort. I remember back time was inquiries of that sort really brought the positive effect of OPB, better defense, and GB rates all into the foreground of discussions of value. None of that diminished the importance of slugging or pitchers K rate for example, but gave a more complete and nuanced view. Baserunning is one more thing to build in, and it’s something which, say, the Upton’s and Bourn too bring to the table as a singificant and _consistent_ part of their game, something one can count on at least at present. More to think on . . . .

      Vote -1 Vote +1

  3. TKDC says:

    So even the slight downward slope might be explained by the correlation of high strikeouts and bad base running? The Braves as a whole are probably around average in terms of base running, as their awful base runners (McCann, Freeman, 3rd base) are cancelled out by their above average to great base runners.

    Vote -1 Vote +1

    • Steve 1 says:

      No, that plot has no correlation whatsoever. Jeff was being generous when he said ‘hardly’. That’s a pure scatter plot, plain and simple.

      Vote -1 Vote +1

    • Wil says:

      To be fair McCann isn’t an “awful” base runner, he is below average but so are most catchers. He’s a better base runner than Matt Wieters and Posey if the metrics are to be believed.

      Freeman is also barely below average.

      In general the Braves will probably be an above average base running team because none of their runners are truly awful, at least not awful enough to offset the excellent base running by the entire outfield

      Vote -1 Vote +1

      • Anon21 says:

        Re: McCann, that surprises me. My eyes suggested he was just atrocious on the basepaths–not that he made bad decisions, but that he was preposterously slow, such that gappers routinely turned into singles or outs at second base.

        Vote -1 Vote +1

      • TKDC says:

        I agree completely (“awful” base runner was a bad way of putting it – low value on the base paths is better). I watch almost every Braves game and McCann looks like one of the slowest human beings in the sport of baseball. I do love when he sneaks a stolen base. It almost always involves him taking off before the pitcher even starts towards home. He’s 23/30 for his career.

        And yes, he’s not really that bad for a catcher.

        Vote -1 Vote +1

  4. eastsider says:

    Does this ruin the idea of a productive out? A strikeout is not productive. If it doesn’t really matter at what rate you have unproductive outs then does it mean that it doesn’t really matter at what rate you have “productive outs?”

    Vote -1 Vote +1

    • olethros says:

      By itself, no. There’s a difference between “subtracting value” and “failing to add value.” Just because Ks are neutral doesn’t mean that other types of outs don’t have a positive effect. I sincerely doubt that they do, but the possiblity exists.

      Vote -1 Vote +1

    • LTG says:

      There are other ways to have unproductive outs (pop-ups, ground balls to the left, shallow fly-balls, unlucky line-drives). More data would have to be studied before considering that inference.

      Vote -1 Vote +1

    • wilt says:

      With a runner on first and no outs, you might say that the hitter should at least put the ball in play and move the runner over, and that a strikeout does nothing. But double plays are also a possibility, and also a force-out in which the hitter simply takes the runners place at first– effectively the same as a strikeout.

      If I remember correctly, there are VERY small differences in WPA between a strikeout and groundout in certain situations, but nowhere to the degree that it would ever become a concern unless maybe you had a team full of Mark Reynolds(es) and Adam Dunn(es)

      Vote -1 Vote +1

    • Wil says:

      To steal from CAC:

      “In The Book: Playing The Percentages In Baseball, Tango shows (Table 11) that value of a strikeout is worth -.301 runs, while a non-strikeout is worth -.299 runs”

      So there is little difference between a K and any other out. Especially true because any non K out is always subject to the possibility of multiple outs, while a K can only result in one out.

      Vote -1 Vote +1

      • Balthazar says:

        On the other hand, putting the ball in play raises the possibility of an error, which may show up in the .002 difference between K and non-K outs. Regardless, too minor an issue to be consequential in a talent acquisition evaluation.

        But there are Ks and there are _Ks_. Too me, the issue is less the absolute number than what the batter is swinging at. If the guy doesn’t know the strike zone, or has a major hole, THAT matters more than his willingness to take a bit cut and accept he’ll miss at times. What kind of K really does matter in assessing a guy; where is he vulnerable? It would be more intersting in someways to evaluate a group, like the present Atlanta roster, and see if they have an across-group profile of where their ‘K weakness’ lies. I doubt that would have a huge effect on their runs scored, but it might mean that some pitching staffs line up with them particularly well, and that’s worth knowing.

        The minituae which makes the fan salivate . . . .

        Vote -1 Vote +1

    • rogue_actuary says:

      You can have productive outs, but it’s more of a consequence than a skill.

      Teams/players are not “better” at hitting fly balls with runners capable of advancing than they are at hitting fly balls in general.

      I didn’t think that we were talking about “productive outs” anymore.

      The problem with comparing the value of a strike-out to the value of “other” outs, is that it doesn’t isolate the run expectancy of a specific “productive” out situation vs. just having your guy swing away.

      There are very few situations in which a team should voluntarily surrender an out.

      Vote -1 Vote +1

    • Braves11 says:

      And McCann is the only slow player in the league? Every team has slow base runners, that doesn’t make them a bad base running team. It’s the decisions they make that matter.

      Vote -1 Vote +1

  5. Phantom Stranger says:

    I think there is a widespread belief amongst front offices that a lineup full of high-K, low-contact hitters is more prone to being dominated by elite pitching and certain types of pitchers. Where would that supposed effect show up? In the playoffs…

    Vote -1 Vote +1

    • Undocorkscrew says:

      Certain types of pitchers? Wouldn’t any kind of lineup tend to be dominated by “certain types” of pitchers? Wouldn’t “elite” pitchers stand more of a chance of dominating any lineup than a “non-elite” pitcher?

      Vote -1 Vote +1

    • Jonnie K says:

      Wow. Well put. I truly believe that comment about playoff pitching. And remember – statistics are legitimate – but do NOT tell the whole story. I disagree that high strikeout rates don’t adversely affect teams. It’s counterintuative. Next thing they’ll say is that walks don’t factor into ERA…

      Vote -1 Vote +1

  6. Jim says:

    What’s the p-value on the slope of the line for your baserunning analysis? From the graph, it doesn’t look like baserunning is statistically significant, even if it is *less* insignificant than K-rate.

    Vote -1 Vote +1

  7. dcs says:

    This has been studied by many people. The difference between a K and a regular batting out is around .01 runs, give or take. So if the Braves next year have 150 more K than an avg team, they should only expect to lose about 1.5 runs.

    Vote -1 Vote +1

  8. Scott says:

    The only extra effect a K has vs. a BIP out is the ability to advance runners. Since advancing a runner usually entails an out your run scoring probability will rarely go up from a “productive” out. (this is why sac bunting is usually bad strategy unless playing for 1 run late in the game).

    basically, the only productive outs come w/ a runner on 3rd and less than 2 outs or potentially a player moving from 1st to 2nd with less than 2 outs. This would also be true for bases loaded/<3 outs (but the DP possibility may negate this base/state) 1st and 2nd w/ less than 2 outs, 2nd and 3rd with less than 2 outs, and 1st and 3rd with less than 2 outs. That is at best 6 of the 24 base/out states and some of the less common ones and many are double edged swords of DP probability (can't hit into a DP w/ a K).

    Let's guesstimate that those base/out states come up ~10% of the time. If a player K's at 1.5x the league average rate (that'd be ~30% which is generally the tops you will see in MLB) then he is moving runners along 50% less often in 10% of his PA's….I just don't see how that effect could add up to much more than a few runs over the course of a season if that….

    Vote -1 Vote +1

  9. JChang says:

    If strike outs don’t matter to hitters, then strike outs (and maybe FIP) shouldn’t matter to pitchers.

    Vote -1 Vote +1

    • brendan says:

      I think you are correct, that pitchers prevent runs at almost the same rate with a BIP out vs. a K.

      However, when it comes time to give credit for getting the out, a pitcher gets much more credit for the K. That’s why FIP matters to us.

      Vote -1 Vote +1

      • kevinthecomic says:

        Your comment is probably true 90% of the time. The only exception that I can think of would be with a team that has a brutal defense. For example, the Tiger’s infield is manned by a tree (Cabrera), a lamppost (Infante — he sways in the breeze a little bit) and two fire hydrants (Peralta, Fielder). A ground ball pitcher can struggle with these guys behind him, which is probably one of the reasons the Tigers are trying to move Porcello, while a high K guy like Verlander is an absolute necessity.

        Vote -1 Vote +1

      • jfree says:

        So in the Tiger infield, who is the dog who is in mark your territory heaven?

        Vote -1 Vote +1

      • kds says:

        In the outfield, chasing balls.

        Remember that K rate, combined with BB rate and HR rate, e.g. FIP, is better at projecting next years run rate than this years run rate. At the season level the batted ball data for pitchers is mostly noise.

        Vote -1 Vote +1

    • Kris says:

      Strikeouts do matter for pitchers. This is a weird situation we find ourselves in – caring about them for pitchers but not for hitters.

      The reason is because we are creating a false dichotomy with hitters when we investigate a strikeout vs a BIP out. The true dichotomy for a hitter is a strikeout or a BIP, which may or may not be an out, and in fact will NOT be an out about 30% of the time. If a pitcher allows balls in play, they will go for hits 30% of the time, while strikeouts will lead to baserunners far less than 1% of the time.

      Why do we not care as much about strikeouts for hitters? Because we can’t make an assumption of all else being equal. We assume that strikeouts are the result of a hitter’s approach to maximizing his hitting value. That is to say, if a hitter makes a conscious effort to reduce his strikeouts, it will come at the cost of some other aspect of his offensive performance (usually assumed to be his power). Right now, the best assumption we can make is that a hitter has already optimized his cost-benefit analysis at his current K-rate. We don’t care about his Ks because given his skills as a hitter, this amount of Ks is simply the tradeoff we make for the amount of power/AVG/OBP/whatever he provides. It would absolutely be better for him to make more contact, all else being equal, because that extra contact would translate directly into more hits in the same number of PAs. We just can’t posit the all else equal part.

      Vote -1 Vote +1

  10. dcs says:

    In my post at 3:44 I only addressed the direct run difference between a K and an in-play out. But people often point out the ‘esthetic’ drawbacks to a higher k approach. Yeah, seeing a batter swing and miss is not pretty. But it does have a good aspect. It keeps the PA going (unless it’s strike 3) so that the next pitch might be a better pitch to hit, or a ball on the way to a possible walk. On the negative side, a strikeout precludes putting the ball in play with 2 strikes, and the possibility of a ball falling in for a single. Overall, these things pretty much cancel out for MLB batters, and K rate is only rarely worthy of mention. Looking at Ks separately from other outs might have value at the lower levels of baseball…

    Vote -1 Vote +1

  11. beckett19 says:

    Would have to agree with dcs there. Error% is probably higher at lower levels. I would conjecture that at a high school or college level, the combination of less fielding experience and more powerful line drives due to a larger sweet spot on aluminum bats would make a strikeout way more costly. Would be great to see some numbers on that as a followup.

    Vote -1 Vote +1

  12. John says:

    Guys. Baseball is the coolest.

    Vote -1 Vote +1

  13. JD says:

    Getting down to the nth degree (what the ‘n’ in fangraphs stands for anyway), the value of a strikeout on wearing down the opposing pitcher should also be considered, no? On the Productive Out Continuum that has first-pitch gidp as its lowest point, the strikeout must rate much higher. I didn’t look up the average number of pitches that it takes to strike someone out, but it must be above 6. You don’t mind a few of those early in the game if it means the stud starter is on the bench by the fifth inning.

    Vote -1 Vote +1

  14. Hurtlockertwo says:

    Strikeouts matter when you watch the games. Having 6-7 guys in your lineup that strikeout 130-165 times each is not my idea of exciting baseball.

    Vote -1 Vote +1

    • Wil says:

      It is if you are a fan of the pitchers team!

      Vote -1 Vote +1

    • Anon21 says:

      If they put lots of runs on the board (which this Braves outfield should), you learn to live with it.

      Vote -1 Vote +1

    • LHomonacionale says:

      I would always prefer to see a guy strikeout trying to score the guy from first than see him ground into a double play trying to move both of them one base. The majority of double plays occur with a man on first. It would seem wise to put a high power, high K% guy immediately behind a productive singles/walk producer because his outs will be less liable to eliminate the runner at first.

      Vote -1 Vote +1

      • TKDC says:

        Or a guy with high K%, high power, high speed (and lefty, and high BB%) like Jason Heyward, who grounded into 4 DPs last year in 136 chances. The perfect #2 hitter.

        Vote -1 Vote +1

  15. Ian says:

    I wonder how team K% correlates to feast & famine offenses. Perhaps plot K% vs. standard deviation of runs scored per game?

    While the A’s carried a high K% to a playoff berth, the Giants carried a pretty low K%, low HR lineup to another championship. (There are other differences between the teams of course, but their season runs scored totals were nearly identical.)

    Vote -1 Vote +1

  16. Wil says:

    Did you guys snag the idea for this article from Capitol Avenue Club? They had a great article arguing basically the same thing posted yesterday.

    Vote -1 Vote +1

  17. dcs says:

    —-”Strikeouts matter when you watch the games. Having 6-7 guys in your lineup who strike out 130-165 times is not my idea of exciting baseball.”

    Sure, but making outs is almost never exciting. You think that watching a guy ground out routinely to short, or pop-up on the infield, or hit a routine fly to center is exciting? Perhaps a line drive out, and long fly caught at the top of the fence, are mildly exciting, but those are rare. The fact is, most outs are routine boring plays. For me, a swing at the first pitch resulting in a routine out is much less interesting than a strikeout where the batter also worked the pitcher for a called ball or two.

    Vote -1 Vote +1

    • Hurtlockertwo says:

      If you hit the ball you at least have a chance to get a hit, error, move the the runner up, score the runner from third. Strikeouts just put pressure on the next hitter, and frustrate the fans.

      Vote -1 Vote +1

  18. Jimbo says:

    Another factor to consider is that the period of this study includes the steroid era when there was significantly more power hitting than there is now. I wonder if a strikeout has a different impact on run scoring in the current environment than this study suggests.

    And I do not think it is accurate to compare the value of a strikeout to the value of any other out. I think it is more correct to compare the value of a strikeout to the value of an AB that ends in any other outcome. To compare a strikeout to any other out assumes that an AB that does not end in a strikeout would end in some other type of out, which is not the case.

    Vote -1 Vote +1

    • Baltar says:

      Surely you are not suggesting that all statistics from “the steroid era” should be ignored in studies.
      How about stats from “the pitchers’ era” of the late 60′s?

      Vote -1 Vote +1

  19. Antonio bananas says:

    I think it’s silly to correlate 2 things. There are sooo many other factors. You’d have to do a lot to neutralize all that. I like the braves. Balanced lineup with lefties and righties. Solid power up and down. Patches of speed. Mostly young so they should collectively get better. Great defense at short and right, good at first (arguably great, freeman looks great, I don’t care much about a 1B range), the upto a are both good. Uggs and Francisco are bad and McCann is solid. Their 8 starters all have value. Honestly at this point, I’m more worried about their SP depth.

    Vote -1 Vote +1

  20. Homey D says:

    Here’s a situation where a productive out by making contact and no stiking out matters:

    Runner on second with no outs

    Are you gonna tell me that striking out isn’t bad here? if the hitter chokes and slaps the ball the other way, the runner is on 3rd with one out, with a greater chance of scoring.
    I think there are situations where striking out hurts the team more often than others.
    If this team becomes too homer-conscious and forget about playing small ball when called for, watching them might not be as fun as we might think.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *