Flyover: Minnesota Twins

When I started this series last month, my intent was to review each team's performance over the past 30 days, alphabetically. I got through Arizona, Atlanta, Baltimore, Boston, Cincinnati, the Cubs, and Cleveland before I realized that I was going to run out of time and that it'd be more timely to focus on the teams that are in contention for postseason play. That's what I've done next, hopscotching among the Tigers, Brewers, A's, Angels, Cardinals, Pirates, Dodgers, and Giants. Let's take a day off from good teams, though, and instead discuss one of the most interesting teams of the year: The Minnesota Twins.

How Are They Doing Lately? Terrible. The Twins are 8-19 over the last 30 days, the worst record in the American League, and tied with the Diamondbacks for the worst record in the league. That's a 114-loss pace over a full season.

So why write about them? Well, put it this way: After the Tommy John surgery epidemic, replay, and that Yankee shortstop retiring, what's the biggest story this year? The decline of offense. Teams are scoring an average of 4.08 runs per game this year. That's the lowest average since 1980. and a run per game down from the highs of 1999 (5.08) and 2000 (5.14). 

Take the American League. Since the All-Star Game, American League games are featuring 8.05 runs per game, scoring 4.06 runs and allowing 3.99. (The discrepancy is due to interleague games.) But if you exclude the Twins from the total, that figure drops to 7.91 runs per game, 4.03 scored and 3.88 allowed. Twins games have featured 10.02 runs! How's this: The last time the American League averaged less than eight runs per game, as it's done excluding the Twins since the All-Star break, it was 1972, and the league responded by instituting the designated hitter. The last time the league averaged 10.02 runs per game, as the Twins have since the All-Star break, the league batted .276 and Pedro Martinez was the only pitcher with an ERA below 3.70; this year the Tigers are the only team batting above .276 (by .001), the league ERA is 3.83, and there are 24 pitchers with ERAs below 3.70. The Twins have been a throwback to the Steroid Era!

Actually, the pitchers have been more throwbacks than the hitters of late. Over the past 30 days, the Twins are scoring 4.3 runs per game, fifth in the league, while giving up 6.1. No other team has allowed more than 4.7 runs per game.

What's Going Right? The Twins are batting .263 over the past 30 days, fourth in the league, with a fifth-best .399 slugging percentage and a sixth-best .316 on base percentage. Their .329 batting average on balls in play, compared to a league average of .297, indicates that they've probably been a little lucky, since wide deviations don't usually persist. They strike out a lot (23% of plate appearances, third in the league), but their walk rate, 6.8% of plate appearances, isn't much off from the league average of 7.2%. sixth;hey'be been good at getting on base (.284 batting average, .336 on base percentage, both second in the league) over the past 30 days. Their .421 slugging percentage is fifth, as they've hit only 20 homers, which is tenth in the NL. Their batting average on balls in play is .330 compared to a league average of .303, suggesting that their batting average and on base percentage have been a little lucky, but only 7.5% of their fly balls have gone over the fence compared to a league average of 9.6%, suggesting that their slugging percentage and home run totals have been a little unlucky. The Giants have been the toughest team in the NL to fan, striking out in only 17% of plate appearances compared to a league average of 21%. They've had problems with their bullpen, as I'll explain below, but the starters have been best in the league at getting hitters to swing at pitches outside the strike zone (33%), they've struck out the most batters (23% of batters), and walked the fourth fewest (6%).

What's Going Wrong? The Twins pitchers have been remarkably consistent over the past 30 days: 5.78 ERA overall, worst in the league; 5.78 starters' ERA, worst in the league; 5.77 relievers' ERA, worst in the league. You know the On The Field Of Play pitching mantra, It's good to throw strikes, but it's better to get strikes? As it turns out, the Twins have been pretty good at throwing strikes (51% of pitches in the strikes zone, third most in the league) and getting strikes (batters swinging at 32% of pitches outside the striek zone, fourth highest in the league). So why haven't they been successful? Because the pitches they've thrown, inside and outside the strike zone, have been hittable. Twins pitchers have the third-slowest fastball in the league (91.6 mph average), the third-slowest slider (83.0 mph average). They've allowed the highest batting average on balls in play (.334, league average is .293), the second highest percentage of fly balls going over the fence (11.7%, league average is 9.2%), and the most baserunners to score (35%, league average is 27%). Those three figures, widely different from the league averages, probably indicate some bad luck, but that's not all of it. You throw hittable pitches, you get hit.

This shouldn't be blamed exclusively on the pitchers, though. The Twins fielders have converted 67% of balls in play into outs this year, the lowest percentage in the league, and the team does poorly in advanced defensive measures. 

Who's Hot? The Twins have managed to compile those awful pitching numbers despite having Phil Hughes, who'll get a few Cy Young Award votes. Over the past 30 days, he has a 3.44 ERA to go with 33 strikeouts and just one walk over 36.2 innings. His strikeout to walk ratio is 11.3, which is the highest in baseball history, topping Brett Saberhagen's 11.0 in 1994 and Cliff Lee's 10.3 in 2010. (h/t: D.J. Short at Hardball Talk) He's making me look smart when I wrote in December

Hughes isn't an elite pitcher. But he's probably better than what he was last year, and the move out of the Bronx should help him keep more balls in the park. By the end of the year, I wouldn't be surprised if his numbers were better than league average.

Actually, that isn't all that smart, because he's been a lot better than league average, 21st in ERA, eighth in WHIP, and a 15-10 record for a team that's won 61% of the games that he's started (99-63 pace) but only 38% of games started by others (61-101 pace). 

The Twins also acquired Ricky Nolasco over the winter, and while he's been OK over the past 30 days (3.93 ERA), he's been a disappointment, going 5-12 with a 4.29 ERA, making me look not-so-smart when I wrote back in November

I'd expect an average or slightly better-than-average ERA and 200 or so innings, but nothing spectacular.

The league ERA 3.83 and Nolasco's pitched 153 innings.

No hitter has been really outstanding, though rookie centerfielder Danny Santana has hit .290 and is seven-for-eight as a basestealer over the past 30 days.

Who's Not? Rather than name names, let's just say that Twins starters not named Hughes or Nolasco have an 8.00 ERA over the past 30 days, giving up more homers in their 77.2 innings pitched (16) than the Indians (15), Padres (15), Rangers (13), Royals (12), Braves (11), and Diamondbacks (11) have hit. No reliever has been particularly good, and closer Glen Perkins had more blown saves (three) than saves (two) and more than twice as many home runs allowed (five) than strikeouts (two) to go with a 12.27 ERA before revealing a season-ending nerve injury. 

Rookie DH Kennys Vargas leads the team with five homers and 15 RBI over the past 30 days, but otherwise has looked kind of lost at the plate, striking out in 29% of his plate appearances (the league average is less than 21%) and failing to draw a single walk. Rookie right fielder Oswaldo Arcia is batting .228 with an even higher strikeout rate than Vargas's, 32%. 
  
What's the Outlook? You'll note that every hitter I mentioned is a rookie. The Twins also have two of the most highly-rated prospects in the game, consensus No. 1 Byron Buxton, an outfielder, and third baseman Miguel Sano. Sano is 21 and Buxton's 20, and they both missed all (Sano) or almost all (Buxton) of the 2014 season with injuries, so they won't be headed up to Minnesota until sometime next year at the earliest. But there are bats on the way.

The pitching this year has been terrible, of course, but there are plenty of young arms in the farm as well. I could see the Twins being an exciting team at the plate in coming years, but my jury's still out on the pitching staff. The Twins have had a pitch-to-contact focus in recent years, while the rest of the league's gone for strikeouts. The idea is that the Twins's approach holds down pitch counts, preserving pitchers' arms, and lets the players in the field turn batted balls into outs. Well, it hasn't worked. The fielders aren't that great, and the Twins will finish the season with by far the lowest strikeout rate in the league (16.5% of plate appearances, league average is 20.1%). They were last, by far, in 2013, 2012, and 2011 as well. They weren't in 2010, which is also the last time they played in the postseason. I've heard Twins fans called for manager Ron Gardenhire's head, as this will be the team's third last place finish in the past four years, but it seems to me that someone in player development--specifically, pitcher development--is more to blame. The Twins aren't drafting, or aren't developing, or both, the type of pitchers who are successful in contemporary baseball. The Twins have used 23 pitchers this year, and the only ones with above-average strikeout rates are Hughes, Perkins, and 27-year-old rookie Aaron Thompson, who's pitched only 6.1 innings. Until the Twins embrace Ks for their hurlers, it's hard to see them competing, even in the generally weak American League Central. But their efforts to single-handedly keep run production across the league alive, I suppose, is admirable.

Throwing Strikes, Getting Strikes - May Update

Over the winter, I did a series of articles about how pitchers' ability to throw strikes affects outcomes. The research is here, here, here, and here. The basic conclusion is this: throwing strikes is good, getting strikes is better. 

Last year, these pitchers threw the largest percentage of pitchers in the strike zone:
     Pitcher           Zone%   W-L    ERA  WHIP
  1. Bartolo Colon     58.5%  18-6   2.65  1.17
  2. Cliff Lee         56.9%  14-8   2.87  1.01
  3. Bronson Arroyo    56.1%  14-12  3.79  1.15
  4. Jordan Zimmermann 55.0%  19-9   3.25  1.09
  5. Jose Fernandez    55.0%  12-6   2.19  0.98
  6. David Price       54.0%  10-8   3.33  1.10
  7. Julio Teheran     53.2%  14-8   3.20  1.17
  8. R.A. Dickey       53.1%  14-13  4.21  1.24
  9. Shelby Miller     53.1%  15-9   3.06  1.21
 10. Travis Wood       53.0%   9-12  3.11  1.15
Data from FanGraphs. Zone% = percentage of pitches in the strike zone.

And these pitchers got batters to swing at the largest percentage of pitches outside the strike zone:
     Pitcher        O-Swing%   W-L    ERA  WHIP
  1. Hisashi Iwakuma   36.5%  14-6   2.66  1.01
  2. Cole Hamels       36.5%   8-14  3.60  1.16
  3. Patric Corbin     35.3%  14-8   3.41  1.17
  4. Adam Wainwright   34.5%  19-9   2.94  1.07
  5. Homer Bailey      34.1%  11-12  3.49  1.12
  6. Andy Pettitte     34.1%  11-11  3.74  1.33
  7. John Lackey       33.8%  10-13  3.52  1.16
  8. Felix Hernandez   33.6%  12-10  3.04  1.13
  9. Dillon Gee        33.4%  12-11  3.62  1.28
 10. Jordan Zimmermann 33.2%  19-9   3.25  1.09

Data from FanGraphs. O-Swing% = percentage of pitches outside the strike zone at which batters swung.

For a frame of reference, the average ERA last year was 3.87 and the average WHIP was 1.30. In case you were wondering, NL Cy Young Award winner Clayton Kershaw was 35th in Zone% and 13th in O-Swing%. AL winner Max Scherzer was 16th in Zone% and 48th in O-Swing%. 

Obviously, these aren't perfect measures, but they're not bad. The pitchers listed above are all really good. How about this year?

     Pitcher           Zone%   W-L    ERA  WHIP
  1. Bartolo Colon     58.5%   2-5   5.36  1.37
  2. Phil Hughes       56.9%   4-1   3.92  1.28
  3. Zach McAllister   56.1%   3-3   3.89  1.30
  4. Nathan Eovaldi    55.0%   2-1   2.86  1.11
  5. Robbie Ross       55.0%   1-4   5.04  1.52
  6. Alfredo Simon     54.0%   4-2   2.89  1.08
  7. Henderson Alvarez 53.2%   2-3   3.33  1.40
  8. Jordan Zimmermann 53.1%   2-1   2.92  1.32
  9. Cliff Lee         53.1%   3-3   3.64  1.29
 10. Jesse Chavez      53.0%   2-1   2.47  1.03
Data from FanGraphs. Zone% = percentage of pitches in the strike zone.

     Pitcher        O-Swing%   W-L    ERA  WHIP
  1. Masahiro Tanaka   40.0%   5-0   2.57  1.00
  2. Felix Hernandez   38.0%   3-1   2.73  1.10
  3. Phil Hughes       37.0%   4-1   3.92  1.28
  4. Jose Fernandez    36.7%   4-2   2.44  0.95
  5. John Lackey       36.4%   5-2   3.57  1.21
  6. Tim Hudson        36.3%   4-2   2.09  0.81
  7. Francisco Liriano 35.4%   0-3   4.64  1.45
  8. Stephen Strasburg 35.3%   3-2   3.42  1.35
  9. Roberto Hernandez 35.2%   2-1   4.08  1.46
 10. Dallas Keuchel    34.8%   3-2   3.68  1.18

Data from FanGraphs. O-Swing% = percentage of pitches outside the strike zone at which batters swung.

There are two ways you could look at this year's leaders. With the obvious caveat that we're not a fourth of the way through the season, so this isn't a large sample, you could look at that list and see some misfires. (All data are through Sunday, by the way.) I mean, Colon and Ross have been legitimately bad, and Liriano and Hernandez have been...how would you put it...not very good? So has the throwing strikes is good, gettng strikes is better rule broken?

Or is there a better explanation - that maybe Colon et al have pitched better than their record might indicate, and they could be positioned to perform better, and that the perhaps surprising performance of Hughes and Alvarez and Keuchel might be sustainable?

I'm guessing that's the case, but we'll check in later in the season for an update. 

How High-Quality is a Quality Start?

One of the easier-to-understand new metrics is the Quality Start, a statistic invented by Philadelphia Inquirer sportswriter John Lowe in 1985 (according to Wikipedia). (I suppose something that's been around 29 years isn't all that new.) A starting pitcher gets credit for a quality start if he pitches six or more innings and allows three or fewer earned runs. That's it.

The criticism you hear of a quality start is that three runs in six innings is a 4.50 ERA, which isn't very good. Of course, a perfect game's a quality start too. Do quality starts indicate pitching excellence or mediocrity?

I looked at quality starts in 2013. Here are a few tidbits:

• There were 2,431 games last year, so 4,862 games started. There were 2,556 quality starts. That means about 53% of starts were quality starts. That doesn't sound that impressive. But wait.
• The leaders: Clayton Kershaw and James Shields had 27, Adam Wainwright had 26, Cole Hamels and Max Scherzer had 25. By percentage of starts: Kershaw had quality starts in 82% of his starts, Shields 79%, Scherzer 78%, and Cliff Lee, Bartolo Colon, Chris Sale, and Matt Harvey all had 77%. Those are all good pitchers, including both Cy Young Award winners. That sounds good.
• Teams that got a quality start won 65.6% of their games last year. That's the equivalent of going 106-56. That tells you that a quality start gives a team an excellent chance to win. 

All well and good, you're thinking, but do pitchers in quality starts actually pitch well? It turns out the answer is: Yeah, do they ever.

The major league average ERA and WHIP for starters last year were 4.01 and 1.306, respectively. The six-inning, three-run outing with a 4.50 ERA cited above would thus be below-average.

But that's not the norm. In their 2,556 quality starts last year, pitchers compiled a 1.93 ERA and 0.986 WHIP. The closest to those figures: Kershaw's 1.83 ERA and Jose Fernandez's 0.979 WHIP. Yes, it's true that a starter can get a quality start with a 4.50 ERA. That happened 213 times last year, 8.3% of quality starts. But on average, you're getting something like Kershaw and Fernandez when you get a quality start.

What Happens When a Pitch is Thrown: Part 2 - Strikes and Balls

Here is the story behind this series.

In the first entry, we determined that just under 45% of pitches are in the strike zone, and pitchers who get it in the strike zone regularly, particularly those who exceed 46%, do better than those who don't. But I noted that throwing into the strike zone isn't the same as getting a strike, because the umpire can get the call wrong in or out of the zone and the batter may swing at a pitch outside the strike zone. So now we're going to just look at strikes.

How often is a pitch a strike? Per Fangraphs, of the 709,916 pitches thrown last season, 451,242 were strikes. That's 63.6%.

Is there a difference between the leagues? A little bit: 63.4% in the American League, 63.7% in the National League. I'm guessing here, but I assume the difference is due to pitchers batting. We've all seen pitchers flailing away at pitches outside the strike zone.

Is there a difference between starters and relievers? Starters get strikes on 63.7% of pitchers, relievers 63.4%. That surprised me since, as we saw in the last post, relievers throw more pitches in the strike zone. Digging deeper, AL starters and relievers are about the same: 63.4% starters, 63.3% relievers. The larger difference is in the NL: 63.8% starters, 63.4% relievers. Again, I assume that's because starters are more likely to face the other team's pitcher than relievers, who are more likely to face a pinch hitter with better strike zone command

Is there a difference between right-handed and left-handed pitchers? Not enough to care about. It's 63.4% for lefties, 63.5% for righties. 

Does it matter? I broke all pitchers into ten groups, based on the percentage of their pitches that were strikes. The top group, Group 1, had strike percentages between 67.7% (NL Cy Young winner Clayton Kershaw) and 83.3% (the Phillies' John McDonald, who got 10 strikes on the 12 pitches he threw - the most for a guy you've heard of is 73.9% for Boston's closer Koji Uehara). The lowest 10%, Group 10, got strikes on less than 60.5% of their pitches. Here's how the 10 groups fared in terms of ERA, WHIP (walks and hits per inning), and percentage of batters struck out, walked, or allowed a home run:
   Group ERA  WHIP   %K    %BB   %HR
     1  3.16  1.10  22.0   4.8   2.4
     2  3.39  1.17  21.8   6.4   2.5
     3  3.80  1.23  20.3   6.5   2.7
     4  3.52  1.24  20.4   7.1   2.3
     5  3.72  1.25  20.9   7.8   2.7
     6  3.86  1.32  19.0   7.3   2.5
     7  4.00  1.33  19.8   8.4   2.5
     8  4.14  1.39  20.0   9.6   2.5
     9  4.61  1.47  17.4   9.3   2.5
    10  4.64  1.55  16.9  12.3   2.5

This is a lot more dramatic than the chart in the last entry, which divided pitchers by their ability to pitch in the strike zone. Throw out the freak ERA for Group 3 (or, more accurately, blame it on CC Sabathia, Dan Haren, R.A. Dickey, and Joe Blanton), and every successive group is worse than the one before it. Getting strikes is clearly important, the more the better. The top 10, Group 1, has starters like Kershaw, Cliff Lee, David Price, and Jordan Zimmermann, and relievers like Uehara, Mariano, and Kenley Jansen. The best pitchers get the most strikes.

A fair question to ask: Why is getting strikes a bigger determinant of pitching success than getting pitches in the strike zone? Ah, that's where things start to get interesting. Next post.

What We Know So Far: The more strikes a pitcher gets, the better he fares.

What Happens When A Pitch is Thrown: Part 1 - The Strike Zone

This is the first of a series of posts. I'm not sure how many - we'll see where it goes.

Baseball can be described as a series of events. In the last game of the World Series, for example, Boston won 6-1. On a more detailed level, the Red Sox took command by scoring three in the third and three in the fourth, the biggest hit being Shane Victorino's bases loaded double in the third. But at the most granular level, the game was determined by the 284 pitches thrown in the game, and what happened with each of them. We're going to look at how that works.

When I say we're going to look at this, I'm not using the editorial we. I'm going to ask a lot of questions and try to provide answers. Most of the time, I'm finding out the answers for the first time. We're going to find out things together.

So here goes.

How often is a pitch in the strike zone? There are several places to find this. I used Fangraphs, which has customizable leaderboards. I looked up the percentage of pitches in the strike zone. The answer, in 2013, is 44.7%. This isn't the same thing as the percentage of pitches that are strikes. (I'll get to that next.) Sometimes a pitch in the zone is erroneously called a ball. Sometimes a pitch outside the strike zone is erroneously called a strike. Often, a batter will swing at a pitch outside the strike zone. All those can turn pitched strikes into balls and pitched balls into strikes. Here, we're just looking at the percentage of pitches that cross the plate in the strike zone. The answer, as I said, is 44.7%.

Is there a difference between the leagues? Not much of one: 44.7% in the American League, 44.8% in the National League. Maybe it's a little higher in the NL because pitchers don't have to get cute with the strike zone when the other team's pitcher is at the plate; they can just throw it in the zone and there's a good chance the pitcher will make an out. I'm not going to worry about it; the difference is too small.

Is there a difference between starters and relievers? A slight one. Starters are in the zone 44.7% of the time, relievers 44.9%. Again, that's a pretty small difference. I think we can attribute it to two factors. First, relievers come into the game fresh and aren't fatigued. Second, relievers often come into pressure situations where a base on balls is more damaging than, say, a walk in the second inning with one out and nobody on. Anyway, the difference isn't large.

Is there a difference between right-handed and left-handed pitchers? Lefties are in the strike zone on 44.9% of their pitches compared to 44.7% for right-handed pitchers. I'm not going to hypothesize yet on the reason for this small difference. 

Does it matter? This is a key question when looking at any statistic: Context. For instance, last year the Dodgers won 88% of the games that they led after seven innings. Does that mean that if you want to win against the Dodgers, you need to score early, because their bullpen locks down leads after seven innings? No. In fact, that 88% success rate was one of the lowest in the majors. A figure doesn't tell us anything by itself, we need to know what it means.

So what does throwing strikes mean? To answer the question, I borrowed a method from the Bill James Handbook 2014 and divided all pitchers into ten roughly equal-sized groups, based on the percentage of their pitches that were in the strike zone. The top 10% ranges from Wesley Wright, who, in 53 2/3 innings split between the Astros and the Rays, got 48.4% of his 910 pitches in the strike zone, to the Cardinals' Rob Johnson, who threw only four pitches all year with three in the zone. Familiar names in this group include Dodgers closer Kenley Jansen (55.1%), Phillies starter Cliff Lee (53.3%), and Reds workhouse Bronson Arroyo (50.0%). The bottom 10%, all of whom got 41% or less of their pitches in the strike zone, includes Yankees ace Hiroki Kuroda (38.8%) and A's closer Grant Balfour (39.6%). So if there are good pitchers in the best and worst groups, how important is this?

Pretty important, it turns out. Here are ERA, WHIP (walks and hits per inning), and percentage of batters struck out, walked, and allowed home runs for each group:
         ERA  WHIP   %K    %BB   %HR
     1  3.67  1.23  20.2   6.5   2.6
     2  3.66  1.23  20.8   7.1   2.7
     3  3.65  1.25  21.2   7.8   2.4
     4  3.73  1.27  21.1   7.9   2.6
     5  3.88  1.30  18.2   7.2   2.4
     6  3.89  1.30  20.8   8.1   2.5
     7  3.94  1.32  19.2   7.8   2.5
     8  4.00  1.33  20.2   8.9   2.4
     9  4.08  1.37  19.2   8.8   2.6
    10  4.17  1.40  17.9   9.3   2.4

In the table, the first row is the top 10% of pitchers, the second is the next 10%, etc. This says that pitching to the strike zone is important, though not as dramatically as we might've thought. There really isn't much of a difference among the top 30%, other than that pitchers who throw more pitches in the strike zone walk fewer batters. Once you get past them, though, the ERAs and WHIPs start climbing. The dividing line's right around 46% of pitches thrown in the strike zone--below that, and performance starts to deteriorate. The one exception is home runs, which don't appear to be dependent on the percentage of pitches thrown in the strike zone. In fact, the correlation between home run rate and pitches in the strike zone is positive, indicating that pitchers in the zone tend to allow more homers.

What We Know So Far: Pitchers who throw more pitches in the strike zone are generally better pitchers than those who don't.

Matt Harvey, 2013

Let me just say this.

Pitchers with 191 or more strikeouts, 31 or fewer walks in a season (yes, I know they're arbitrary endpoints, just play along):

Player	Year	SO	BB	Age	Tm	Lg	G	GS	CG	W	L	W-L%	IP	H	R	ER	ERA
Matt Harvey	2013	191	31	24	NYM	NL	26	26	1	9	5	.643	178.1	135	46	45	2.27
Cliff Lee	2012	207	28	33	PHI	NL	30	30	0	6	9	.400	211.0	207	79	74	3.16
Roy Halladay	2010	219	30	33	PHI	NL	33	33	9	21	10	.677	250.2	231	74	68	2.44
Cy Young	1905	210	30	38	BOS	AL	38	33	31	18	19	.486	320.2	248	99	65	1.82
Cy Young	1904	200	29	37	BOS	AL	43	41	40	26	16	.619	380.0	327	104	83	1.97

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 8/26/2013.

(via Corey Seidman @CoreySeidman on Twitter)