Observational Sabermetrics

May 23, 2014

Observation 1

 

                Is it just me, or does Bert Blyleven in 2014 look exactly like Herbert Hoover?

 

Observation 2

 

It’s funny the things you can miss.    I’ve seen Bob Meusel’s batting record in a baseball encyclopedia 1,000 times; I’ve studied him, evaluated him.  I had Bob Meusel on a Ballpark team for years. . .and I never before noticed that his strikeout to walk ratios are absolutely terrible.

You can miss something like that because strikeout to walk ratios have changed so much over time that a strikeout to walk ratio that was terrible in the 1920s no longer looks off today.   In the 1920s a good hitter usually had more walks than strikeouts.  You look at Meusel. . .72 strikeouts, 20 walks in 1920, 88-34 in 1921, you don’t realize how bad those numbers are because they’re not bad compared to Corey Patterson, Peter Bourjos or J. P. Arencibia.  

I got interested in the question of hitters who have terrible strikeout to walk ratios given

a)  their own era, and

b)  their quality as hitters.

In other words, good hitters with bad strikeout to walk ratios.

I set up a method to identify the worst of all time.   I took a spreadsheet with the batting records for all players, and eliminated from that spreadsheet:

                a.   All hitters with less than 400 plate appearances, and

                b.   All hitters in seasons in which one league or another was not counting strikeouts (a hundred years ago and more.)

The remaining seasons, I ranked the players by

                a.   Runs Created per 27 outs, and

                b.   Strikeout to Walk Ratio.

In 1920, for example, there were 105 players with 400 or more plate appearances.    Babe Ruth ranked first among them in Runs Created per 27 outs, or, as CNN would say:   BREAKING NEWS, BABE RUTH BEST HITTER OF 1920.    I stated this as a percentage, .995. . .the actual formula was this:

 

Formula

 

In other words, for Babe Ruth, ranking first out of 105 hitters, I divided  0.500 by 105.500, then subtracted that from One.   The result is .99526.

This number I multiplied by two and subtracted one, creating a scale which runs from +1 to -1; the best hitters are near +1; the worst hitters are near -1.

Then I did the same thing for strikeout to walk ratios, only I reversed the polarity on the strikeout to walk ratio.  Of course, generally speaking, good hitters have good strikeout to walk ratios.   Tris Speaker, who was the third-best hitter of 1920 (behind Ruth and Sisler) had the best strikeout to walk ratio of the season, and Shoeless Joe, who was the fourth-best hitter, had the second-best strikeout to walk ratio.  Babe Ruth had the 29th-best strikeout to walk ratio of 1920, still well in the "positive" range when we convert the numbers to a +1 to -1 scale, except that I reversed the polarity of the strikeout to walk ratio so that the good strikeout to walk ratios would show as negative numbers.

Ruth is thus a non-instructive example; I don’t know why I started with him.    Ruth’s outfield teammate, Bob Meusel.  Meusel, a rookie in 1920, was the 16th best hitter in the major leagues.     He hit .328 as a rookie with some power, creating 6.98 runs per 27 outs.  16 out of 105 is .853 by the little formula above, which becomes +.706 when we convert it to a +1 to -1 scale.

                However, Meusel had the worst strikeout to walk ratio of the season, ranking 105th out of 105 hitters.  That makes him .0047 on the strikeout to walk scale, which becomes negative .991 when we convert this to a +1 to -1 scale, which becomes +.991 when we reverse the polarity so that bad strikeout to walk ratios are good (because that’s what we’re looking for—good hitters with bad strikeout to walk ratios.)  Bob Meusel is what we’re looking for:  a good hitter with a bad strikeout to walk ratio.

                We "score" the season by multiplying these two numbers by one another, then converting that to an integer equivalent.   In other words, for Bob Meusel, he ranks at .706 on the "good hitter" scale and .991 and the "bad strikeout to walk ratio" scale.   We multiply .706 times .991, and we get .699.  

                This is the highest figure of the 1920 season, although not by much.   There is a wrinkle here.   George Cutshaw scores at .682.    George Cutshaw ranks as the 101st best hitter of the 1920, out of 105—the fifth worst hitter—but has the 13th best strikeout to walk ratio.   That makes him negative on both scales, and when you multiply a negative times a negative, you get a positive, so Cutshaw scores at +.682, just behind Meusel.  

                We’ll filter those guys out in a moment; that’s not what we’re looking for.    But first, how many players would you guess would have "positive" numbers here?    A player has a positive number if

                a)  He is a good hitter but has a bad strikeout to walk ratio, or

                b)  He is a bad hitter but has a good strikeout to walk ratio.

                In other words, he has a positive number if the two disagree.   What percentage of players would you guess have positive numbers?

                The number is much higher than I would have guessed.   It’s 38%.   I would have guessed that 80% of players would "agree", thus have negative numbers, and that only 20% of players would "disagree", like Bob Meusel and George Cutshaw, and thus have positive numbers.     But in fact it’s 38%.  It’s amazing what we don’t know.

                Anyway, we now get rid of the players with negative numbers.   I started with a spreadsheet that has 52,841 lines of player data in it, and eliminated those players with less than 400 plate appearances.     That left me with 17,568 lines of data.   I eliminated the data for those seasons in which some leagues were not counting strikeouts; that left me with 15,015 lines of data.    Now we eliminate those players who have negative scores—the good hitters with good strikeout to walk ratios and the bad hitters with bad strikeout to walk ratios.    That leaves us with 5,703 lines of data.

                Half of those, though, are bad hitters with good strikeout to walk ratios.    We’ll eliminate those; that leaves us with 2,877 lines of data, each representing a season by a hitter who hit well but had a bad strikeout to walk ratio. 

                A few of those, however, are so weak as to be irrelevant.  Going back to Bob Meusel in 1920, Meusel had a score, you will remember of +.699.   We multiply this by 100 and take the integer, and Meusel gets 69 "points" for this season.     The maximum points you can get for a season is 99; if you were the best hitter in the major leagues but had the worst strikeout to walk ratio in the major leagues, you would get 99 points for that.   It’s never happened. 

                Of the 2,877 relevant seasons, however, 230 have "positive" numbers but score at zero, so we’ll eliminate those.   That leaves us with 2,647 player/seasons which are relevant to our study—2,647 good hitters who had bad strikeout to walk ratios.  Bob Meusel in 1920 actually ranks 22nd among those seasons.   These are the top 25 seasons, the top 25 players who had good seasons but bad strikeout to walk ratios (in the context of their era.)

                25.   Al Simmons, 1929.   Simmons hit .365 with 34 homers, 157 RBI, but ranked 98th out of 106 regulars in 1929 with a strikeout to walk ratio of 38-31.   69 points.

                24.   Bob Horner, 1979.   Hit .314 with 33 homers, 98 RBI, but ranked 180th out of 188 regulars with a strikeout to walk ratio of 74 to 22.    69 points.

                23.  Hank Greenberg, 1934.   Hit .339 with 63 doubles, 26 homers, 139 RBI, but ranked 97th out of 109 regulars with a strikeout to walk ratio of 93 to 63.   69 points.

                22.  Bob Meusel, 1920.    69 points.

                21.  Walt Dropo, 1950.    Drove in 144 runs as a rookie, but ranked 103rd out of 110 regulars with a strikeout to walk ratio of 75 to 45.   70 points.

                20.   Reggie Jefferson, 1996.   Hit .347 with 19 homers in 386 at bats, but ranked 193rd out of 205 regulars with a strikeout to walk ratio of 89 to 25.   70 points.

                19.    Tom Brown, 1891.   Hit .321, stole 106 bases  and scored 177 runs—the second-highest total ever—but ranked 92nd out of 104 regulars with a strikeout to walk ratio of 96 to 70.   (By "regulars" we mean players with 400 plate appearances.)  70 points.

                18.  Kirby Puckett, 1986    Hit .328 with 31 homers, 96 RBI and forced Bob Costas to name his kid after him, but ranked 161st out of 180 regulars with a strikeout to walk ratio of 99 to 34.  71 points.

                17.  Jeff Heath, 1941.   Hit .340 with 123 RBI and was one of five players ever to hit 20 doubles, 20 triples and 20 homers in a season, but ranked 104th out of 117 regulars with a strikeout to walk ratio of 69 to 50.   71 points.

                16.  Dave Kingman, 1979.   Hit .288 with 48 homers, 115 RBI, but ranked 172nd out of 188 regulars with a strikeout to walk ratio of 131 to 45.  71 points.

                15.  Cito Gaston, 1970.    Hit .318 with 29 homers, 93 RBI and set himself up for a career as a batting coach, but had the worst strikeout to walk ratio in the major leagues in 1970:  142 to 41.  72 points.

                14.   Willie McGee, 1985.   Won the NL MVP Award and led the league in hitting at .353, 18 triples, but ranked 164th out of 184 regulars with a strikeout to walk ratio of 86 to 34.  73 points.

                13.   George Watkins, 1930.    His .373 batting average as a rookie is still a record, but he ranked 100th out of 105 regulars with a strikeout to walk ratio of 49 to 24.   75 points. 

                12.   Andres Galarraga, 1988.    Hit .302 with 42 doubles, 29 homers, but ranked 178th out of 186 regulars with a strikeout to walk ratio of 153 to 39.  75 points.

                11.  Joe Medwick, 1935.    Hit .353 with 23 homers, 224 hits, 46 doubles, 132 runs scored and 126 RBI,  but ranked 109th out of 116 regulars in 1935 with a strikeout to walk ratio of 59 to 30.  75 points.

                10.   Bob Meusel, 1921.   Followed up on his rookie performance with numbers similar to Medwick in 1935--.318 with 40 doubles, 16 triples, 24 homers, 135 RBI, but ranked 111th out of 113 regulars with a strikeout to walk ratio of 88 to 34.    76 points.

                9.   Kirby Puckett, 1988.   Had a season much like Medwick and Meusel, hitting .356 with 42 doubles, 24 homers, 121 RBI, but ranked 170th out of 186 major league regulars with a strikeout to walk ratio of 88 to 34.  76 points.

                8.  Dante Bichette, 1995.   Hit .340 with 40 bombs, 128 RBI in Colorado, but struck out 96 times with 22 walks, ranking 167th out of 172 major league regulars.   76 points.

                7.   Ryan Braun, 2007.   A rookie as many of these players were, he hit .324 with 34 homers, 97 RBI, but ranked 204th out of 213 regulars with a strikeout to walk ratio of 112 to 29.   77 points. 

                6.  Pete Reiser, 1941.   Hit .343 and was second in the National League MVP voting, but ranked 108th out of 117 major league regulars with a strikeout to walk ratio of 71 to 46.   78 points.

                5.   Wally Berger, 1933.   Like Bob Meusel, I’ve written quite a bit about Wally Berger, but never before realized that he had poor strikeout to walk ratios, because they don’t look poor by today’s standards.   In 1933 he hit .312 with 27 homers, 106 RBI, 37 doubles, but ranked 108th out of 114 major league regulars with a strikeout to walk ratio of 77 to 41.   80 points.

                4.  Carlos Gonzalez, 2010.    Hit .336 with 34 homers, 117 RBI, but ranked 192nd out of 205 major league regulars with a strikeout to walk ratio of 135 to 40.   81 points.

                3.  Hal Chase, 1916.    Returned to the National League after Chasing his dream in the Federal League for two years, and led the NL in hitting at .339, but ranked 97th out of 102 regulars with a strikeout to walk ratio of 48 to 19.  81 points.

                2.  Kiki Cuyler, 1924.   Another rookie, Cuyler launched a Hall of Fame career with a .354 season, good line-drive power, but ranked 104th out of 105 major league regulars with a strikeout to walk ratio of 62 to 30.   81 points.

                1.  Andres Galarraga, 1993.   Andres hit .370 in 1993, hitting 22 homers and driving in 98 runs in just 120 games, his season ended by a broken hand on July 28.  He had a 1.000 OPS for the only time in his career, but ranked 188th out of 202 major league regulars with a strikeout to walk ratio of 73 to 24.    82 points. 

                The "negative image" Andres Galarraga, if you’re curious, is Bob Lillis in 1965.   Lillis was an absolutely wretched hitter, hitting .221 with no homers in 439 plate appearances, a .522 OPS which is among the worst of the 20th century for  a player with 400 plate appearances, but he drew 20 walks and struck out only 10 times.    May not have had a lot of bat speed.  

                Those are the single-season leaders.   Many of those, however, were players having unsustainable success, guys like George Watkins, Cito Gaston, Dave Kingman and Walt Dropo, who were having career years.    This is the list of the top 25 good hitters with bad strikeout to walk ratios in a career:

                25.   Vada Pinson, 1958 to 1975, 243 points.

                24.  Greg Luzinski, 1970 to 1984, 247 points.

                23.  Dick Allen, 1963 to 1977, 248 points.

                22.  Hack Wilson, 1923 to 1934, Hall of Fame, 257 points.

                21.  Kiki Cuyler, 1921 to 1938, Hall of Fame, 265 points.

                20.  Hank Greenberg, 1930 to 1947, Hall of Fame, 268 points.

                19.  Andre Dawson, 1976 to 1996, Hall of Fame, 271 points.

                18.  Kirby Puckett, 1984 to 1995, Hall of Fame, 272 points.

                17.  Frank Howard, 1958 to 1973, 282 points.

                16.  Jeff Heath, 1936 to 1949, 283 points.

                14-15 tie.  Joe Medwick, 1932 to 1948, Hall of Fame, 294 points.

                14-15 tie.  Juan Gonzalez, 1989 to 2005, 294 points. 

                13.  Tony Perez, 1964 to 1986, Hall of Fame, 300 points.

                12.  Alfonso Soriano, 1999 to the present, 312 points. 

                11.  Wally Berger, 1930 to 1940, 313 points.

                10.  Bobby Bonds, 1968 to 1981, 315 points.

                9.  Bob Meusel, 1920 to 1930, 323 points.

                8.  Jim Rice, 1974 to 1989, Hall of Fame, 343 points.

                7.  Orlando Cepeda, 1958 to 1974, Hall of Fame, 345 points.

                6.  Lou Brock, 1961 to 1979, Hall of Fame, 348 points.    Brock and Bobby Bonds are the only leadoff men in the group.

                5.  Roberto Clemente, 1955 to 1972, Hall of Fame, 356 points.

                4.   Reggie Jackson, 1967 to 1987, Hall of Fame, 402 points.

                3.  Andres Galarraga, 1985 to 2004, 442 points.

                2.  Al Simmons, 1924 to 1944, Hall of Fame, 472 points.

                1.  Willie Stargell, 1962 to 1982, Hall of Fame, 520 points.

                Going into the study, my guess as to who would be the all-time champion good hitter/bad strikeout to walk ratio. ..my guess was Roberto Clemente.   It turned out it wasn’t Clemente, it was his longtime teammate Willie Stargell.   Stargell as a rookie in 1963 hit only .243 with 11 homers, 47 RBI, in part because of a strikeout to walk ratio of 85 to 19.    That season doesn’t contribute toward his career total of 520 points because he didn’t have 400 plate appearances and also wasn’t a good hitter, but the next season he played more, hit .273 with 21 homers—and his strikeout to walk ratio got even worse, 92 to 17.    He had the third-worst strikeout to walk ratio in the majors that season.   He never had a season in which he had 400 plate appearances and his strikeout to walk ratio wasn’t comfortably in the bottom half of regulars.  He only had one season, 1973, in which he wasn’t in the bottom third of all regulars—and that barely, and that only because he was having a monster season and was intentionally walked 22 times.   Even in 1979, when he was the David Ortiz of his era, a 39-year-old who was still one of the best hitters in baseball, he ranked 158th out of 188 major league regulars in strikeout to walk ratio.   He shared the National League MVP Award that season.  

 

 

Observation 3

We appear to be headed to the time at which every major league hitter without exception will have more strikeouts than walks.    Until doing this study, I hadn’t realized had close we had come to that point.   In the 1920s 74% of major league regulars finished the season with at least as many walks as strikeouts.   This changed very little for the next two decades; it dropped to 68% in the 1930s, 69% in the 1940s, but no real movement.  From 1950 to 1952 it was still 67%.

It dropped some after 1952 and began to drop dramatically in the late 1950s; by the end of the decade the percentage was 53%.    In the 1960s positive strikeout to walk ratios became a distinct minority, at 20%, but the lowering of the pitcher’s mound (1969) helped the number for the 1970s to recover to 29%.     74, 68, 69, 53, 20, 29; this is our pattern from 1920 to 1979.

The percentages trended downward during the 1980s; for the decade as a whole the percentage was 23%--28% for the first three years of the 1980s, 22% for the rest of the decade.    The number for the first half of the 1990s was 23%; for the second half, 16%, and for the decade as a whole, 19%.  

As recently as ten years ago, there was still a significant population of hitters who walked at least as often as they struck out, 25 of 218 in the 2004 season.  But in recent seasons, we are trending rapidly toward zero.  In 2009 there were 14 major league regulars (400 PA) who had at least as many walks as strikeouts.  In 2010 there were 7.   In 2011 there were 11, but in 2012, only six (Carlos Lee, Jose Reyes, Chipper Jones, Joe Mauer, Joey Votto and Prince Fielder.)   In 2013 there were four—Norichiki Aoki, Marco Scutaro, Edwin Encarnacion and Alberto Callaspo.   Alberto Callaspo is keeping the ship afloat.   It now seems inevitable that within a few years, there will not be a single major league player who walks more often than he strikes out. 

 

Observation 4

The Hitting Chain

If you think about the sequence of joints that leads from the ground to the bat in a batter’s hands, you’ve got

1)  Feet (not that the foot is a joint, but that the juncture of the foot and the ground is a kind of a joint),

2)  Ankles,

3)  Knees,

4)  Mid-Section or Pelvis,

5)  Shoulders,

6)  Elbows,

7)  Wrists,

8)  Hands.

All hitters use all eight of these elements, but the extent to which they use them is different.  Willie Stargell and David Ortiz are alike in many different ways; they’re at the same point on many different scales.  Stargell won an MVP Award when he was 39; I wouldn’t bet a lot of money that David won’t.   David and Stargell are alike in that both of them are able to diagnose a pitch exceptionally quickly, and then, both of them being enormously strong men, both of them are able generate tremendous bat speed in a fraction of a second after a relatively late trigger.  Both of them get their upper body in position to hit early—generating relatively little power out of their lower halves—but then both of them, once turned, whip the bat with quite exceptional force.

But there is this difference:  that David is generating most of his power around his elbows—that is from the muscles above and below his elbows—whereas Stargell, I am quite certain, had more power in his wrists and hands than anyone I’ve ever seen.    He almost pinwheeled the bat; everything came from his wrists, which were extraordinarily quick and strong.  Most guys who hit with their wrists are singles hitters, like Carew and Gwynn and Willie Wilson; of course, Aaron had those fantastic wrists, and Aaron centered the bat on the ball more often than Stargell did, but Stargell had much more bat speed than Aaron. . ..more bat speed than anybody except Ortiz and Bonds, Sheffield and McCutchen, Mantle and Trout, those kind of freaks of nature.    Gwynn would rotate his mid-section much more than Carew did.

I was thinking about this. ..McCovey and McGriff were in the same group as Ortiz and Stargell, but McCovey had that wide, wide stance and thin torso, and generated a lot more power out of his mid-section and his shoulders than Ortiz or Stargell.   McCovey whipped the bat in a long arc. . .I would guess the longest arc I’ve ever seen in a great hitter.   Sometimes bad hitters have very long arcs.   McGriff, too, was doing more with his mid-section than David does; David’s basically just positioning himself with his mid-section and then cutting loose with the shoulders and elbows.

My point is, if you can see these things, you should be able to study them; you should be able to teach yourself to see them in a more organized way.   John Wathan hit more with his ankles than anyone else I’ve ever seen; that is, he kind of jumped at the ball.    Casey Blake jumped at the ball quite a bit, not as much as Wathan.    Kevin Youkilis did a lot with his knees; he was exploding out of a crouch as the ball got to home plate. . .of course, Rickey was in a crouch, and Bagwell was in a  crouch.    Some guys have a long stride and are getting some power out of shifting their weight, but most power hitters keep their weight back until very late in their swing and have a shorter stride.    Gary Sheffield hit with his wrists, of course—everybody remembers the rocking back and forth of the bat—but he stood straight as a post, getting nothing out of his knees or mid-section, then kicked his left leg forward as he hit, using his feet to generate a little forward momentum.  George Brett had a wide set, no stride, and completely transferred his weight to his front leg by the point of contact; a lot of guys with a wide set are a little bit off-balance a quarter of a second before they need to swing.  Pedroia uses his shoulders more than almost anyone, swings from his shoulders.   Jackie Bradley has a lot of swing in his mid-section; he is struggling right now, and it looks bad when pivots his mid-section like that and misses, but then everything looks bad when you’re struggling.   Pedroia looked like hell until he started hitting; people were always telling us he would have to cut down his swing.    Bobby Richardson was like that; he was a small man who cut loose like a lumberjack with his arms and shoulders, but he was able to make contact.  

Some scouts will tell you that "everybody’s in the same position when they swing the bat", but I don’t think that’s exactly right.    I know what they mean; what they mean is that a lot of what we see in the way hitters set up in the batting box, their stance, where they hold the bat, etc.. . ..a lot of it isn’t relevant and you shouldn’t focus on it, because everybody goes to the same point before they actually swing.    I know what they mean, but I don’t think it’s exactly right; I think that can be a way of excusing yourself for not seeing things that you should be seeing.

Is the "back" an element of this?   Not sure. . .obviously hitters USE the muscles in their back, some more than others, but there’s no joint in your back; you don’t fold your body at your spine.   The back muscles are "pulling" the joint below, the mid-section, and the joints above, the shoulders.  It may be that the back is an independent variable; some guys have stiff backs, whereas other guys are more flexible.   Also, I’m not sure whether one can actually see the difference between a hitter using his wrists and using his hands; I know that many good hitters have huge, strong hands, and I am sure that they are using the hands to help them generate power; I’m just not sure that you can see it.  I was thinking that there should be a way to quantify how much power a player is getting from each of these.  Like this:

 

BodyMatrix

 

                Just a sketch. . .obviously need a lot more work.

 

 

Observation 5

Functionally Random

All sports depend upon people attempting to do things that are so difficult to do that no one can do them with absolute consistency.   Shooting a free throw, for example, is one of the easiest things to do in sports—but no one hits 100% of their free throws.   Even the best free throw shooters miss 10% of their shots or something, and when a shot will be missed is functionally random.    Shooting free throws is not a random activity; some people are much better at shooting free throws than Shaquille O’Neal, and you get a better result if you do things right than if you do things wrong.  However, if you stood at the free throw line with either Shaquille O’Neal or Steve Nash, you could not predict whether he will miss or make the next free throw with any more accuracy than you could predict the next random number.     It is functionally random. 

The free throw is one of the very few activities in a team sport which has no defensive component at the moment of the shot; another is the kick off in football.  In almost all sporting activities, there are two elements of functional randomness:

1)  The inherent difficulty of the task itself, and

2)  The fact that one player is trying to prevent the exact outcome that the other player is trying to create.

Outcome, or array of outcomes.   Sometimes you get asymmetrical competition, in which one player is trying to prevent a different outcome than the other player is trying to create; for example, a defender in basketball may be jumping to block a shot, but the offensive player may be faking a shot, when in reality he intends to pass.   In baseball the pitcher may be trying to prevent a deep fly ball, but the hitter might be focused on getting on base, might be more than willing to take the walk, or he might be just intending to foul that pitch off and force the pitcher to throw some other pitch.  Offensive strategy often consists of creating asymmetrical competition, because the defense often fails when it is attempting to prevent the wrong thing.  

In general, though, offense and defense are in symmetrical competition.   Whatever one player is most trying to do is what the other player is most trying to prevent.   This creates a "double randomness", in that there is functional randomness both on the part of the hitter (or shooter), and on the part of the defender (or pitcher). 

In hitting a baseball, there are three more levels of functional randomness:  there is the plane, there is the timing, and there is the direction of the ball.    I’m thinking about modeling the problem; it’s not truly random, but it appears to be truly random because there are so many levels of functional randomness that the player-determination elements are entirely masked.   The three additional levels of functional randomness on the part of the batter:

a)  plane,

b)  timing,

c)  direction.

If the batter swings at exactly the right height—that is, on precisely the right plane—then the outcome may be good for the batter, but it is so difficult for the hitter to swing on precisely the right plane that more often than not he cannot do it; more often than not he will swing an inch too low or an inch too high or a quarter of an inch too high or a quarter of an inch too low.    Whether or not he can get the bat in the right plane is a matter of skill, not luck, but whether or not he will get it precisely right on this particular swing is so difficult that no one could predict it.  It is functionally random.

Even if he gets the bat on precisely the right plane, he still gets a negative outcome unless his timing is absolutely perfect—and again, while timing the pitch is skill, not luck, getting absolutely perfect timing is SO difficult that it is beyond skill, and requires some luck as well, which is to say that that also is functionally random.

And even if the batter times the pitch perfectly and also swings on precisely the right plane at precisely the right moment, the ball may go directly to a fielder or it may go in some direction where there is no fielder; this, again, is functionally random.    A good hitter can direct the ball to a place where there is no fielder to a certain extent, but this is so difficult to do that it is functionally random whether he can do so or not on this particular swing.

In sabermetrics we deal constantly with cause and effect relationships which should be there, but which cannot be found in the data.   The fan perceives that a hitter is "hot" or that he is "cold"; the analyst concludes that the patterns are random.    The fan perceives that the batter hits well in the clutch or that he does not; the analyst finds that there is no evidence for such a proposition.   The fan perceives that the on deck hitter is "protecting" the hitter; the analyst finds that such effects are so small as to be negligible.

In sabermetrics we are constantly asked to look for cause-and-effect relationships that should be there.  A pitcher should lose effectiveness in the next inning after he has to run the bases in the previous half-inning.    When you face a knuckleball pitcher and then you come back to face a hard thrower the next day, that should mess you up.    You should hit better on your birthday, because it’s your birthday, or you should hit better with a fast runner on first base, because you will see more fastballs. 

One comes to know immediately, from having done dozens of similar studies, that that effect isn’t going to be there when you look for it.   The reason it isn’t there is that for that cause and effect relationship to manifest itself in the statistics, it has to fight its way through layer after layer after layer of functional randomality;  I mentioned three layers of functional randomality for the hitter, but there are at least as many from the pitcher’s standpoint.   It is not that these cause and effect relationships do not exist, exactly; it is that they do not connect from end to end.   There are too many screens in the way. 

Backing off and running at this in a different direction. . . .Hitting skill as it functions in a swing can be reduced to three components:

a)  plane,

b)  timing,

c)  bat speed.

Scouts always talk about Bat Speed, and I get frustrated—and many times the scouts get frustrated—by the fact that bat speed is overrated; there are very good hitters who have very little bat speed, like Ichiro and Tony Gwynn, and there are absolutely terrible hitters who have tremendous bat speed, like Wily Mo Pena.    Wily Mo had bat speed like David Ortiz, like Stargell, like Sheffield; he just couldn’t hit.

Bat speed is certainly overrated in scouting, but the thing is:  it’s not random.   It’s there or it isn’t.   You can see it, you can observe it; you can be 100% correct in evaluating it.   Whether a hitter can master getting his bat on the right plane is unknown; whether he can master getting the timing exactly right is unknown.   Bat speed is observable.   In any kind of competition of pieces of information, the knowns will crush the unknowns, even if the unknowns are actually more important than the knowns. 

Trying to back off to the beginning of this, again. . .it struck me that there is something here that we could know, but that we do not know.    Only a tiny percentage of swings succeed.    Miguel Cabrera is the best hitter in baseball; he swung the bat 1,206 times in 2013 and got only 193 hits.   Even the best hitters get relatively few hits per swing.    80 to 90% of swings fail.

Whenever a swing fails, it fails for one of two reasons:  either the plane is wrong, or the timing is wrong.   (Occasionally a hitter swings at a pitch so far off the plate that he couldn’t reach it even if his bat was on exactly the right plane at exactly the right instant, but that’s a relatively rare outcome, and we can deal with that.)  Normally, either the plane is wrong, or the timing is wrong.

If you watch the pitch, you can almost always say whether that was a timing failure or a "plane" failure.   It is observable which outcome has occurred.   If a hitter swings too early or too late, he will almost always foul the pitch off, although sometimes, if he fails to read a changeup, for example, he will swing SO early (or so late) that he will miss the pitch entirely.    If a hitter swings under a pitch he will pop the pitch up; if he swings over it he will hit it into the ground.   It is relatively easy to observe which of these has happened.

Sometimes, of course, the batter hits the ball on the nose but it goes right to a fielder.   That’s not really a failure; that’s an unfortunate outcome.   Sometimes a hitter hits a ball that trickles in front of the plate and he reaches first.   That’s not really a success; that’s a fortunate outcome.

We could study at bats, then, and sort each swing into one of five outcomes:

a)  Perfect swings with good outcomes,

b)  Perfect swings with unfortunate outcomes,

c)  Bad swings with fortunate outcomes,

d)  Bad plane, and

e)  Bad timing.

We can ignore the first three of those; we know how many good outcomes each hitter has, and we have as much information as we need about hitters hitting in good luck.   What I am focused on here is the bad plane/bad timing thing.   Every swing that fails, fails either because it was on a bad plane or there was bad timing, or both.

But since the results of these two are the same in the final statistics, we know nothing more about the subject.   There is a raft of information there that we could have, if we studied the issue, but do not have.    Obviously Miguel Cabrera and Mike Trout have very good timing and swing on an optimal plane, but let’s take. …Mark Kotsay, or Martin Maldonado, or Rob Brantly.   Someone who is not quite such a good hitter.   Does this hitter fail because his timing is bad, or does he fail because he doesn’t get the bat on the right plane?  

Obviously, both, but is it all the same for every hitter?   That would seem unlikely.    It would seem likely that there would be some (not good) hitters who are extremely good at timing the pitch, but extremely bad at putting the bat on the right plane, while there are other not good hitters who are good at putting the bat on the right plane, but bad at timing the pitch.    Travis Snider, for example. . .I would guess, watching him hit, that his "plane" is good but that his ability to time the pitch is terrible.   It’s a guess; I don’t know.   I don’t imagine anybody knows.  I don’t imagine his hitting coach knows.   There is information there that we could have, but we don’t. 

We analyze failure in terms of mechanics.   When a player is struggling and you talk to the hitting coach, you will get something about mechanics; he’s opening up too early, or he’s pulling off the inside pitch, or he’s lunging at pitches, or he is not picking up the movement out of the pitchers hand, or something.   Nobody actually knows whether that stuff is insight or gibberish; it all sounds the same.  Jackie Bradley Jr. is struggling.  Is he struggling because he isn’t timing the pitch, or because he is on a bad plane?

Well. ..it’s obvious that he’s not picking up off-speed pitches at a major league level at this point, and this results in bad plane; when you swing at a pitch in the dirt, that’s always a bad plane.   My point is, though, that when a given hitter is struggling, the hitting coach might say that he is opening up too early, but if you had actual, factual information, you might look at the data and realize that he’s not having a "timing" problem, at all; he’s having a "bad plane" problem.   Information is always better than guesswork.   There is information there that we could have if we did the work.

 
 

COMMENTS (11 Comments, most recent shown first)

slh112358
re: Blyleven/Hoover. I can't see it. Looking at photos of Hoover through Google image search, Hoover strikes me as a better match for Gardenhire!
8:33 PM Jun 1st
 
moscow25
This is way too math-y, but could we approximate bad plane/bad timing by considering hitters' results against pitch types & location?

I know Bill is loathe to incorporate pitch f/x data, and for good and understandable reasons, but this is an extension of "if he swung on a pitch in the dirt, that's bad plane"?

Though I'm immediately stumped. If a player fails to catch up to a high fastball... that's timing right? Or a bit of plane, too...
7:44 PM May 26th
 
evanecurb
Bill: I found out about the relationship between timing and plane the hard way - went 1 for 8 in a double header; seven groundouts to the second baseman and a ground ball single to right.
7:20 PM May 26th
 
DEK1966
I remember Reggie Jefferson as just another failed prospect. Forgot (or just failed to notice) that he had a 19-HR, .347 season as a two-thirds time player. Wow.
3:56 PM May 26th
 
bjames
Evanecurb--that's correct. I probably should have put in a sixth option, for "combined failure".
9:55 AM May 24th
 
evanecurb
Timing and plane are interrelated in some (not all) cases. If a left handed batter with a slight uppercut swings too early on an outside and low pitch, his bat will be slightly above the plane of the ball when he makes contact, and he will ground out to the second baseman. If he swings late at an inside fastball, he will be below the plane of the ball when he makes contact with the handle of the bat, and he'll hit a pop up or foul off the pitch.
8:32 PM May 23rd
 
smbakeresq
Riceman1974 might be on something with "greatest bad player I ever saw." I read that and immediately started thinking of players. Wily Mo Pena I think is the defining limit, a little better and you would play a lot more. The 2 players who I think of right away who were a little over the limit are below.

I grew up in the 70's in Baltimore, so the two players I saw that jumped right out were Lee May and Cliff Johnson. They were not bad players, but Lee May had 2 things, the ability to pull the ball with power sometimes and a great, friendly, hard-working disposition. You would never think he was an athlete, but he knew what he had and was good at it.

Cliff Johnson I would see all the time against Baltimore and he always seemed to kill them. Now of course I know that while he wasn't a great ball player, he could have had much better career doing what Lee May did, hit the ball with power. Johnson however could walk some and struck out less, today he would probably hit 500 homers some where as an Adam Dunn type.
12:19 PM May 23rd
 
Riceman1974
Good use of Gwynn and Willy Mo as the antipodes of bat speed. Watching Gwynn swing a bat made you wonder how he ever made his high school team. Willy Mo was a tremendous natural athlete who likely swung a bat in excess of 90 mph, but rarely did it ever make contact. But the few times is did, what glory.

Willy Mo was the greatest bad player I ever saw. He would look incredible at times, but those times were so fleeting, literally for an inning or two. He would strike out all the time, make bone-headed plays in the field (routinely turning singles into doubles and was completely ignorant of the concept of the cut-off man), but then he'd hit a 475-foot monster dinger and throw a guy out at third from a distance Clemente wouldn't even attempt. God bless Willy Mo.
11:26 AM May 23rd
 
chuck
The George Scott 1968 question comes to mind with this.
He led the league in batting average on balls in play in 1967, then was near the bottom of the league in 1968. One of the noticeable things in his '68 line is that his ground out-to-fly out ratio went up to 1.4. That suggests to me that
a) pitchers had made an adjustment to him and were throwing him lower pitches or more offspeed pitches with downward movement.,
b) and/or he was swinging on a bad plane,
c) and/or he was unable to lay off the low pitches out of the zone.

One place perhaps to start studying the issue of plane vs timing might be with a batter's heat map. Look at his swings in the worst of the 9 boxes and figure whether the batter is usually off on plane or timing there.
10:39 AM May 23rd
 
taosjohn
Re: Blyleven/Hoover-- I don't think they actually look so much alike-- but at certain angles that Bert is often shot at, the resemblance can be remarkable. It is not just you.

One of the things that will make proceeding on the timing/plane thing tricky is that bad timing on a breaking pitch, as you sort of point out, will appear to be bad plane. But given the number of times contact is actually made, it should be possible to get a partial picture without any picture at all-- just from counts of popups and tricklers vs fouls outside third or first.

When Carlos Gonzales is fooled it is usually plane. When Troy Tulowitzki is fooled it is usually timing; or so it looks to me...
9:52 AM May 23rd
 
greggborgeson
Bill, love the idea of studying bad plane and bad timing (I think there would be another category, bad plane & bad timing on the same swing). Data on this would yield really important new information to help remove the random factors from batter (and pitcher) analysis. What I'd be most interested to see is if batters go through patterns (streaks) where their frequencies go up and down significantly. I've always felt that the studies that discount streaks were flawed because of the highly random influences on the outcome of an at-bat, when measured within the small sample sizes inherent in a streak. You've given us a means to look at the issue more precisely.
6:26 AM May 23rd
 
 
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy