April 18, 2014

The Diff: Cleveland Indians embrace the strikeout

Changing the topic this week back to my original favorite sport, baseball. Thanks again to all for your really positive feedback on last week’s edition on franchise-building in the NBA.

The Diff

“The best way to improve your team is to score more runs and allow fewer runs. I’m not as concerned at how we do it, but our focus is to score more and allow fewer.” Those were the inspiring words of Cleveland Indians general manager Chris Antonetti in an AP story on Dec. 12. Yet, despite that pretty elementary proclamation, all it seems anyone wants to talk about this offseason is strikeouts.

The Cleveland Indians were not very good offensively in 2012. That much, at least, shouldn’t be too difficult for most to comprehend. They overall were not a good baseball team — their 68-94 record should indicate that. So, to a point, Antonetti is dead right and there’s no argument. That’s all that matters: scoring runs, allowing fewer.

But per the Moneyball-obsession of emphasizing on-base percentage as the Bible of stats, somewhere in the last 5-10 years, strikeouts became the enemy. If someone wasn’t getting on-base, then it was bad. Thus, if someone wasn’t even putting the ball in play — meaning there was no opportunity at all to even get a hit — then it was worse.

Sure, overall, strikeouts have never generally been cast as a great thing for any player or team. But in the context of getting on-base most importantly, it was a big deal for a while. More emphasis was placed on players that just didn’t strike out as much.

So my goal today is to place the Indians’ polarizing offseason acquisitions in a little bit of context. I wanted to look at MLB production in 2012, compared side-by-side to strikeout rates, as well as look at the Indians roster for 2012 and 2013, with how it compares to the rest of the AL Central.

For starters, I wanted to be clear about one simple manipulation: In all of the following statistics, I removed pitchers from the equation as much as I could. Of course, pitchers suck at batting. Looking at the 5,913 plate appearances by MLB pitchers in 2012, they batted .129/.162/.166 and struck out in 37.1% of plate appearances. That’s dreadful and skews everything. So it’s truly not fair to include them in any samples, which inherently hurts NL teams.

Thus, with that intro over with, here’s a quick look at the American League:


Team Runs K%
Mariners 3.82 20.7%
Indians 4.12 17.4%
Royals 4.17 16.7%
Rays 4.30 21.6%
Twins 4.33 17.1%
Orioles 4.40 21.3%
Athletics 4.40 22.3%
Blue Jays 4.42 20.5%
AL Average 4.45 19.2%
Tigers 4.48 18.0%
Red Sox 4.53 19.3%
White Sox 4.62 19.6%
Angels 4.73 18.1%
Yankees 4.96 18.8%
Rangers 4.99 17.6%

 

One could immediately cherry-pick the data — as I’ve been known to do — and point out that Cleveland’s AL Central division friends Kansas City and Minnesota were No. 1 and No. 2 in the AL in least strikeouts. The Indians were No. 3. All three teams were well below league-average in runs scored per game. Jon pointed this out already this offseason in a great tweet.

Yes, that’s undoubtedly accurate. And a great argument against strikeouts mattering. But generally speaking, it’s not the full story. In fact, only one team that had above-average run-scoring also had a K% more than 0.1% worse than the league average. So that kind of shows how strikeouts do kind of matter, and that one should avoid them, right? Kind of. Let’s keep going.

On to the National League:


Team Runs K%
Astros 3.60 21.9%
Marlins 3.76 19.0%
Cubs 3.78 19.7%
Dodgers 3.93 18.0%
Mets 4.01 20.1%
Padres 4.02 19.0%
Pirates 4.02 21.4%
Reds 4.13 19.7%
NL Average 4.22 19.2%
Phillies 4.22 16.5%
Braves 4.32 19.8%
Giants 4.43 16.6%
Nationals 4.51 20.7%
Diamondbacks 4.53 19.9%
Rockies 4.68 18.5%
Cardinals 4.72 17.6%
Brewers 4.79 19.1%

 

First key observation: The strikeout rate is the exact same among non-pitchers in both leagues. Isn’t that fascinating? That regardless of pitching and talent level and everything else, strikeout rates for all non-pitchers in both the AL and the NL were right at 19.2% in 2012. That’s just fascinating to me. Inherently, it makes sense, but it’s just amazing.

Next: No matter what I do with removing pitchers’ plate appearances, the runs scored still are a lot lower in the NL. Nothing I can do there. Yet, the Indians — and Royals and Mariners — still scored less than the NL average. So that means they’re pretty bad.

Finally: The same patterns from above don’t work as cleanly as above. The Nationals ranked No. 3 in most strikeouts, with the Diamondbacks and Braves not far behind, yet all had above-average offenses. But the fact only one below-average offense had a K% lower than 19.0% is also interesting.

Here are then the correlation levels between strikeout rates and runs scored per game in the respective leagues:

American League: -0.187
National League: -0.368

OK, then. So is there a potentially negative relationship? Yes, or so indicates the data simply from 2012. But the correlation levels aren’t by any means significant, so for now, statistically, it’s relatively inconclusive.

I could keep going on and on about other years and bigger data sets, here are my two takeaways before I move onto just the Indians and their new players: 1) Strikeouts don’t matter. Or, at least that much, in either direction. 2) Strikeout levels are pretty consistent across both leagues, with the average for both at 19.2% in 2012.

Next topic: The Indians and their roster turnover.

In fact, that’s how we got to this point anyway. Because Chris Antonetti was talking about scoring more runs, no matter how the team does it. Yet all anyone wants to talk about this offseason is strikeouts. First, let’s start with what the 2012 Indians produced:


Name Status SO/PA PA OPS+
Shin-Soo Choo Gone 21.9% 686 131
Jason Kipnis Staying 16.2% 672 103
Asdrubal Cabrera Staying 16.1% 616 115
Carlos Santana Staying 16.6% 609 122
Michael Brantley Staying 9.2% 609 113
Casey Kotchman Gone 9.8% 500 73
Jack Hannahan Gone 19.8% 318 86
Shelley Duncan Gone 22.3% 264 90
Travis Hafner Gone 17.9% 263 121
Lou Marson Staying 18.7% 235 84
Jose Lopez Gone 15.6% 224 79
Johnny Damon Gone 12.1% 224 73
Ezequiel Carrera Staying 22.2% 158 99
Lonnie Chisenhall Staying 17.9% 151 108
Jason Donald Gone 29.6% 135 50
Brent Lillibridge Gone 32.5% 123 75
Aaron Cunningham Gone 22.9% 109 41
Russ Canzler Gone 22.7% 97 96
Matt LaPorta Gone 28.3% 60 68
Cord Phelps Staying 29.4% 34 52
Vinny Rottino Gone 25.0% 32 6
Thomas Neal Gone 25.0% 24 46
Juan Diaz Staying 29.4% 17 80
Luke Carlin Gone 21.4% 14 40

 

At first, that seems like a lot of roster turnover. More on that in a little bit.

Secondly, doesn’t it seem like a lot of the guys that are no longer on the 40-man roster struck out much more often? And doesn’t it seem like a lot of the players with fewer PAs had worse OPS+ statistics? Again, more on both those concepts in a moment.

What I just wanted to present was the entire team in order, sorted by plate appearances in 2012. It’s pretty sad, clearly, that only six players on the Indians had more than 320 plate appearances last year. In fact, there were 253 players with 320+ PAs last year; so that’d average to about 8.5 per team, when not factoring in trades and the like. So yes, officially, the Indians had lots and lots of roster turmoil and injuries going on last year.

So next, here’s what happens when we segment the 2012 Indians roster into a few different categories. Whether or not they are still on the 40-man roster as of Jan. 30, 2013, or whether or not they had at least 200 PAs in 2012:


Segment # players %/PA SO/PA OPS
Total 24 100.0% 17.4% 0.705
Gone 15 49.8% 19.2% 0.673
Staying 9 50.2% 15.7% 0.737
<200 PA 12 15.5% 24.9% 0.618
>200 PA 12 84.5% 16.1% 0.722

 

Those are the points I hinted at above. Only about 50% of PAs are returning from 2012 to 2013. That seems like a huge amount of turnover, and I’ll place that in a little bit of context first. Most notably though, the returning players had a 15.7% SO/PA rate and a .737 OPS. Both are much better than league average. The players who are no longer on the roster had a perfectly average 19.2% SO/PA rate, with a slightly below average .673 OPS.

A similar pattern also follows for the players that received at least 200 PAs in 2012. To a certain extent, obviously, that shows that the Indians were kind of right with their plate appearance distribution last season. It seems inherent, but the team also deserves some credit: They gave more plate appearances, or at least as many as possible, to the players that were better. And they also kept many of those players. Obviously, Shin-Soo Choo was a clearly above average offensive performer, but besides him, none of the other departing players were that good anyway.

While still on this topic, I wanted to just share this statistic briefly of comparing the Indians’ roster turnover to the rest of the AL Central:


Team PA staying SO/PA SO/PA stay
Chicago 78.1% 19.6% 19.9%
Cleveland 50.2% 17.4% 15.7%
Detroit 79.6% 18.0% 17.7%
Kansas City 86.5% 16.7% 17.0%
Minnesota 71.6% 17.1% 18.3%

 

Wow. The Indians roster turnover clearly is the most in the AL Central. No other team lost more than 29% of its plate appearances from the 2012 season going into 2013 — and that second-most team, Minnesota, traded away two of its best players (Denard Span and Ben Revere) in rebuilding moves. Cleveland, on the other hand, because of the turmoil in 2012 and one trade this offseason, is looking to replace over 49% of its plate appearances.

But here’s the notable thing again about those returning players: They only struck out in 15.7% of their plate appearances in 2012. That’s well below league average. Michael Brantley was under 10%, while Jason Kipnis, Asdrubal Cabrera and Carlos Santana all hovered around 16%. Those marks all are below the established mark of 19.2%.

So this means that even with the upcoming changes for 2013 — and I’ll finally get to those now — the Indians would have to strike out an absolute ton to make it all the way to even a team mark of 19.2%. Doing some simple math, if you hold over 50% of plate appearances at 15.7%, the remaining 50% of plate appearances would have to strike out 22.7% of the time. And only one-fifth of the 253 players with 320+ plate appearances struck out that often last season. Yet two of those guys are heading to Cleveland.

Here are the new additions to the 40-man roster that didn’t have a plate appearance in Cleveland last year:


Name SO/PA Level
Nick Swisher 21.3% Majors
Drew Stubbs 29.3% Majors
Mark Reynolds 32.6% Majors
Mike Aviles 14.5% Majors
Yan Gomes 22.0% Minors
Mike McDade 22.9% Minors
Chris McGuiness 19.5% Minors
Tim Fedroff 16.6% Minors

 

Stubbs and Reynolds are two of those well-above average strikeout guys from 2012. Plus Swisher, who also was above the mark of 19.2% last season. Aviles has been fairly good at avoiding them in his MLB career, while the other four players have yet to establish enough of a big league track record. Their numbers listed above are their minor league marks.

So last but not least, I hoped to now estimate out plate appearances for the Indians in 2013, thus estimating strikeout rate based on career averages, and then make a few final comments about the team’s ability to at least score more runs.

Here’s my best shot at estimating 2013 plate appearances based on the current 40-man roster:

Name PA est. SO/PA career
Jason Kipnis 650 17.4%
Asdrubal Cabrera 640 16.3%
Nick Swisher 630 21.3%
Carlos Santana 620 18.0%
Michael Brantley 610 12.2%
Drew Stubbs 580 29.3%
Mark Reynolds 550 32.6%
Lonnie Chisenhall* 500 15.4%
Mike Aviles 300 14.5%
Lou Marson 300 21.7%
Yan Gomes* 225 22.0%
Ezequiel Carrera* 150 15.6%
Mike McDade* 125 22.9%
Chris McGuiness* 75 19.5%
Tim Fedroff* 75 16.6%
Cord Phelps* 50 16.5%
Juan Diaz* 50 19.8%
Total 6130 19.9%

 

A couple notes first:

– The players marked with a star do not have at least a full year’s worth of MLB experience, so I’m using their career minor league numbers. Obviously, those probably are on the low end of what they’d do in the big leagues. But many of the big-leaguers then are at lower levels — because of better patience, such as Michael Brantley — than their career marks, so I’m estimating that these two forces practically will cancel each other out.

– I’ve tried to estimate this to as accurate a pattern as possible. The average AL team had 6,128 plate appearances in 2012. Also, the average AL Central team had about 67% of their plate appearances by their top seven guys — I’ve estimated the Indians for about 70%, which obviously is on the high-end and assumes relative health for all of those regulars. But it also shows how many at bats the other guys always will get, no matter what happens.

So the Indians, in my estimate, should have about a 20% strikeout rate in 2013. This obviously is a rough estimate because baseball projections are always insanely difficult, but the three true outcomes — walks, strikeouts and home runs, as this Indians-based fantasy projection detailed — are generally consistent from year-to-year for specific players.

Comparing 20% to last year’s established average of 19.2%, obviously I’m guessing the Indians will strike out a bit more than league-average this year. That’s a huge jump from the 17% mark in 2012. But it doesn’t necessarily mean anything — the Diamondbacks, Braves, Nationals and White Sox all have above league-average marks with about 20%+ strikeout rates.

And when you look at the changes — swapping one good player in Choo along with scrubs like Hannahan, Kotchman, Duncan and Jose Lopez for three regulars in Swisher, Stubbs, Reynolds, plus more proven depth in Aviles, Gomes, etc. — then I think you have to believe the Indians improved in what matters most: Scoring more runs, no matter how that occurs.

While overall I don’t think the Tribe will be significantly more improved this season — unlike Jordan Bastian, my guess is that 81 wins is likely the best-case scenario — I think the offense should be notably better. Even though we’ll most definitely see a few dozen or hundred more strikeouts along the way.

  • mgbode

    you bring up alot of good points and i’ll try to get back to some of those later in the day.

    one thing to note: I prefer to look at SO’s as a ratio in baseball. I fully subscribe to the “3 true outcomes” method of hitting (walks, so’s, and HRs). much like a pitcher, the SO isn’t as important as the BB:SO ratio and then you have to factor in how often the hitter gets a HR as well (cannot be a straight ratio there).

  • Natedawg86

    I looked into this a little at the past Indian teams (when they were good at hitting) and other teams to look at how many total SOs the teams had year by year. I was going to make a joke about how the indians could potentially strike out +1200 times this year ( Thats 7 per less than 9 innings (assuming we sometimes don’t have to bat in the bottom of the ninth). That is funny that you came up with an estimated 1220 SO projected. Take a look at some of the successful indians teams recently (2007, and 90s) and you will find that they ended up with quite a few SO.

  • JacobWFNY

    Very fair. BB/SO ratio is judged by many to be the single-most important factor in the game, then throwing in HR and those are the only things an individual batter or pitcher can truly control.

    And I think, if I were to go into estimated team OPS — or BB, HR rates or anything — that you’d see that go up for ’13 based on the new regulars we added. I kind of just wanted to focus on strikeouts for now to get that argument out of the way. But there’s so much more that could be a factor in this debate.

  • dwhit110

    Fascinating. Thanks for the analysis!

  • JacobWFNY

    One other point you’ll find fascinating that I didn’t even get to in the article was this post: http://www.fangraphs.com/blogs/index.php/home-run-rates-in-1998-and-2012/.

    Jon RT’d it about three weeks ago. Just incredible to see that debate presented in the context of the three true outcomes.

  • The_Real_Shamrock

    Yuck!

  • WFNYJon

    I’m not sure I totally follow, mg, but K% is my favorite “ratio” to use when looking at Ks. It’s the ratio of Ks per PA, which really should be the denominator for most pitcher-batter stats (K/9 is not nearly as appropriate for all sorts of good reasons; here’s a nice pice on that: http://www.beyondtheboxscore.com/2012/10/8/3451856/mlb-pitching-k9-bb9-plate-appearances-sabermetrics

    The problem is we all got so used to the per nine stats that it’s hard to break old habits. I’m as guilty as anyone.)

    Where was I? Oh yeah. K% is best, which is what Jacob uses here.

    K% is best, of course, IF you believe that K% has some bearing on runs created. Jacob does a fantastic job of showing that’s not really the case though.

    So back to the 3-true-outcomes thing. What do you mean? Like, does that mean a player like Billy Butler gets no credit for all those doubles? I understand that a player like Adam Dunn IS a TTO guy, but not all guys are, so I don’t know why we’d want to use the same rubric for all of them?

    I may be wildly missing your point, by they way.

  • Ezzie Goldish

    Ugh. Even as I try to swear off baseball, you suck me back in with these good posts…

    What I don’t get is why a team that seemed to be so wisely ridding itself of K’s would then turn around and grab some of the league’s worst offenders?

    I can think of three options, curious what you think:

    1) They believe they can afford the Ks because of the relatively low Ks elsewhere, and that these are otherwise good players.

    2) They believe they can teach them to change significantly, which would make them really great players.

    3) They believe that Ks was too small of a factor since it didn’t help them much this past year, so they’re giving up on some aspects of their advanced analysis.

  • JacobWFNY

    I would say 1 is the most-right answer, then kind of 3 a bit too.

    Pretty much, Ks don’t matter. Runs do. Strikeouts aren’t significantly correlated with runs. So it’s not a big deal, they’re not going to necessarily try to change them significantly (crossing out 2).

    I think they just think these players are much better than what they had. Even 20% is not that drastically awful — it’s pretty darn close to league average. So it’s really not a big deal in the slightest. And if the overall batters are better at the things that clearly do matter — walks, hits, homers, etc. — then we’re better than we were.

  • Natedawg86

    More Ks aren’t always a bad thing too. If they are good offensively, they will have more Ks because they have more ABs. One interesting thing to compare would be swinging Ks vs K’d looking (although much more hard of numbers to come up with)

  • mgbode

    somewhat missing my point, but it’s because I typed it up too quickly on my way out to lunch and poorly stated it (and I never meant to imply to use per9 stats).

    analyzing hitting is complicated as BABIP doesn’t normalize as easy because you need to factor in LD%, FB%, GB% and if we can find a normal rate for each, then we can see what a hitters BABIP may look like (if they can hit the same %’s).

    but, as you know, there are 3 outcomes of a PA where BABIP doesn’t matter. Ks, BBs, and HRs. 2 of those are good. One of those is bad. to find a hitters true value to a lineup, we need to first find what their expected BABIP, then subtract their K% and add in their BB% and HR% to more fully understand their expected impact.

    what I have found as a “quick and dirty” tool to utilize is to use a BB%:K% (missing the ‘%’ above) to figure out if a hitter is going to significantly hurt a lineup with his SOs (or help with his BBs). I know I should put the HR% in there somewhere, so if it’s good, I usually just add in “plus he’s a power hitter” or “and he’s not even a power hitter” into it.

    for instance, our new friend Nick Swisher had a 12.3BB% and a 22.6SO% in 2012 for a 0.55BB:SO ratio. he was much better in 2011 when he had a .76BB:SO ratio. (note: and in this particular example his ISO power dipped when his ratio dipped but that is not the case for alot of examples and isn’t the case when you look back further at Mr. Swisher).

    sorry to confuse the point.

  • mgbode

    that’s a fantastic table on that page.

  • mgbode

    I believe you are correct. I think the new additions add more than what was taken away. I just wanted to add in the other things to consider.

    The part that is always hard for me to figure out is how to factor in “good” contact (LD%) vs. bad contact (GB%, FB% – w/ HR/FB factored in).

    When I put the example of Swisher above though, it really made me pause as he actually seems to have a loose correlation between his BB/SO ratio and his BABIP. His BBs are fairly consistent. But, in years he’s striking out more, he’s getting a much better LD% and, as a result, a much better BABIP. Another layer to this onion.

  • Kmandingo

    The teams with higher strikeout rates have far more power than the Indians. having banjo hitters striking out below the league average is worthless if their OBP is below average as well. Our new acquisition (Stubbs) had an OBP below .300 and Swisher’s in the playoffs was below that! So, we are going to strikeout more AND have a lower OBP. Thanks Ant.

  • Kmandingo

    Bases loaded, Indians down by one in the 9th. Who would you rather have up to the plate, Stubbs, Reynolds, or anyone in a big league uniform that K’s less than 15% and has an OBP above .350? The issue is OBP and Ks. If you K a lot and can’t get on 35% of the time you are an issue. We inherited two issues and can’t attract good talent because of our front orifice.

  • mgbode

    that’s easy. if it’s a RHP, then you just pinch-hit with Hafner. oh wait.

    seriously though, you are applying one specific scenario to an overriding factor. that’s not correct or fair.

    and we got Stubbs through a trade, pretty sure we didn’t have to “attract” him to get him.

  • JacobWFNY

    Kmandingo — You do bring up a good point about the impact of strikeouts in clutch situations such as that. But in the end, we’re not comparing apples to apples here.

    Between a guy with a .350/.450/.800 OBP/SLUG/OPS guy with a 10% strikeout rate and a .350/.450/.800 OBP/SLUG/OPS guy with a 30% strikeout rate, of course I’d take the first one. But these aren’t the same players as last year — they’re better overall hitters, despite (or regardless) of the strikeouts.

    As you can see above, the 15 players that had a plate appearance for the Indians in ’12 but are no longer on the team had a .673 OPS last season.

    Stubbs: career OPS of .698, just had worst career season, but only 28
    Reynolds: career OPS of .807, worst season ever still was .750+, just 29
    Swisher: career OPS of .828, four straight seasons of at least .822, but is 32