Monday, May 6, 2013

Hitting, Hitting Attitudes and Small Ball Over Time

Yesterday I wrote about how runs scored per game has fluctuated over time, and Twitter user @guruE6 asked:
Looks as if the current era correlates well with the Golden Era. I wonder if this is the case with other stats.
Immediately, two thoughts popped into my head:
2. Good question, and I'm always looking for something to research. If I understood the question correctly (and in a 140-character world, that's ALWAYS a big if), this post will show the changes in hits, approaches to hitting and the use of bunting and stolen bases over time.

This chart shows the percentage of hits that were singles, doubles, triples and home runs from 1901-2012:

I'll contrast 1901 with 2012 to demonstrate what it shows. In 1901, around 78% of hits were singles, 14% doubles, 6% triples and 2% home runs. Read anything about the hitters of that time, they had a completely different approach. The balls were used to within an inch of their usefulness, and the spitter was a legal pitch, so by the 7th inning, it would be like hitting a blackened beanbag. It wasn't as much the Dead Ball Era as the Destroyed Ball, with the added bonus that pitchers would abuse those balls in ways that would make Gaylord Perry blush.

Fast-forward to 2012 and singles were 66% of hits, doubles 20%, triples 2% and home runs 12%. Smarter people than me can discuss why the triple has essentially disappeared, but I suspect one reason would be the value of taking that added base has decreased over time with respect to the out if the batter is thrown out. To explain that convoluted statement, let's assume that back in the day, players were much more reckless on the base paths. In the modern era, especially with no outs, a runner on 2nd will score about 63.7% of the time (using the Tom Tango Run Expectancy Matrix), whereas a runner on 3rd with no outs will score 85.3%, an improvement over 64%, to be sure, but contrasted with the probability that a runner will score after being thrown trying to stretch a double into a triple--0%. The reward of that marginal base is outweighed by the risk of an out, and I strongly suspect this type of thinking is behind watching players stroll to second on line drives into the gap. Of course, other factors like the hitter (was it Paul Konerko or Jacoby Ellsbury? Jason Giambi or Carl Crawford? Is it a running team like the Rays or a bang-it-out team like the Yankees?), the game situation, etc. all come into play.

This chart shows the increase in the prevalence of home runs:
It's a two-axis chart, with the left axis (the columns) being hits per game, which is relatively steady over time. This is no surprise, since even back in the Dead Ball Era, teams were scoring runs, so hitters had to be getting on base somehow. The red line is home runs per game, with some interesting and pronounced changes over time. The beginning of the Lively Ball Era in 1920 is clearly demarcated, there's a pronounced dip during World War II and a really nasty dip from 1969-1973 that had a direct impact on the creation of the designated hitter. I'm not sure what happened from around 1985-1993 to cause that rather pronounced trough, but in addition to "enhanced" offense, the new park building binge that began with Camden Yards certainly contributed to the upsurge in home runs that began after the strike year of 1994. That peak crested around 2000, but even today home runs are being hit at a historically high rate.

This chart might look familiar:
This is the picture for my Twitter account, and I can't even begin to explain it. Right up to 1950, walks AND strikeouts were relatively rare, pitchers being careful to throw strikes and hitters finding "shame" in striking out. I'm stunned by the lack of movement over time of the walk line. 

I strongly suspect that the strikeout per game number is at its peak and very near to moving back down--to where, I can't say, but probably in that 6 per game range. The age of "it's just another out" is over, and the three outcome hitter (paging Adam Dunn and...well, they're all gone now except for him) is a dinosaur. When middle infielders like Jose Hernandez were hitting 20+ home runs (and in his case, striking out 180+ times) a year, perhaps the risk/reward matrix did view the strikeout as just another out, and face it, it COULD be worse--the hitter could bat into a double play. I won't know for about five years or so if I'm correct because this will have to begin with minor league instruction, but I have a feeling I'm right. The days when we marvel instead of yawn over a 40 home run season have returned.

I hate the term small ball. When the White Sox won the World Series in 2005, everyone heaped praise on Ozzie Guillen for playing National League ball in the American League. Somehow this notion that they bunted and stole their way to a World Series took hold despite the facts that:
1. They hit 200 home runs, 5th in the majors.
2. They were 4th with 137 stolen bases...and 1st in CAUGHT STEALING with 67. You're not allowed to lead the league in caught stealing if you don't also lead in stolen bases--this means you are BAD at stealing bases.
3. They had lights-out pitching in the playoffs.

The Sox had Paul Konerko (40 HR), Jermaine Dye (31), Carl Everett (23) and Joe Crede (22) and five more players with double-digit homers. They played at US Cellular Field, which had a park factor of 103 that year--they didn't play small ball, they played SOFTball. And don't even get me started on Scott Podsednik--he was the guy starting all the small ball hoopla, batting (a really soft) .290 with 53 stolen bases (and a league-leading 23 caught stealing). What people DON'T remember about his year was that he drove in 25 RBI--in 568 plate appearances! Even for a leadoff hitter, that's pretty rare territory, only 36 players since 1901 having that few RBI with that many plate appearances. Special mention must be made of Enzo Hernandez's 1971 season for the Padres--he had 12 618 plate appearances.

Whew, I feel MUCH better now--almost need a cigarette, but that's probably too much information. This chart shows stolen bases per game and the successful steal percent:

It's another two-axis chart, with stolen base attempts (which are stolen bases AND caught stealing) the blue line and the success rate in the red columns. I wish data went back prior to 1920, because that first data point shows successful steals at just over 50%, which is just pathetic by modern standards, but in an era in which runs were harder to score (not necessarily more scarce), greater risks were taken. You can see that the Lively Ball took a couple of years to make the stolen base both much less prevalent and more successful, since Babe Ruth, Lou Gehrig and others introduced the notion of station-to-station baseball and waiting for the three-run homer. Earl Weaver would have loved that era.

I don't know what caused that upward movement in the 70s of the stolen base--was it nature (i.e., better athletes with more speed) or nurture (a change in strategy)? As always, it's probably both, but it's heartening to see that the increase in stealing attempts had a corresponding increase in the success rate. So, they did it more often AND got better at it. Nobody asked me, and I'll write about it at some point, but in my mind, Lou Brock is one of the worst Hall of Fame selections in recent memory, and I even thought it at the time it happened. But that's just a teaser.

The stolen base is coming back. Just as more home runs made it less of a tool in the Lively Ball Era, the same happened in the "Enhanced Offense" Era--but that era is over. Whether base stealing will creep back up to the level of the 70s and 80s, I suspect not, but it's going to get close.

This chart shows the use of the sacrifice bunt over time:

I thought about this while mowing the lawn today and truly expected to see different values. In the DH Era, half of the major league has a decreased use for the sacrifice but, and yet the incidence didn't drop off the table like I would have expected. 

There's a shortcoming with this table--it only shows SUCCESSFUL sacrifice bunts and doesn't show unsuccessful bunts--those are just recorded as outs, but I feel that's an important stat to test whether skills like bunting have atrophied over time as its use is diminished. I can do that with play-by-play data, but I only have four years at this point, so I can't make grand pronouncements. However, I can use data found in the splits data at to tease out some thoughts.

Data that breaks out bunts only goes back to 1988, and this chart is another two-axis chart, but with three values--sacrifice hits, at-bats with bunts and the batting average on bunts:
 This is tricky, so I'll do it by bullet point:
1.  Sacrifice bunts are not counted as an at-bat. As such, they have no impact on batting average, etc.
2. This means that any at-bat with a bunt is either:
     a. an attempt to get on base with a bunt
     b. a FAILED sacrifice attempt
3. Bunting for a hit is good strategy--teams bat around .350 when bunting, so it works.  
4. I can't tease out of this data what are sacrifice situations and what aren't, and that's pretty important.

The number of sacrifices is pretty constant, around 1400-1500, obviously skewed more toward the National League than the American. There's a pronounced dip in 2000-2001, which coincides with the increase in home runs, but that's history. 

Play-by-play data suggests (and I'm going from memory here) that about 75% of bunts occur with a runner on base, a sacrifice situation. Obviously, it depends on the hitter (I'm sure David Wright and Prince Fielder aren't bunt-crazy). Research I did some time back (and I won't reference it, because it needs to be updated) suggests that teams are successful in bunting (either advancing the runner or getting a hit) about 80% of the time. What I haven't measured is how often that sacrifice bunt actually leads to a run scoring--I've just started doing that, and that should be interesting. I have done a similar analysis with stolen bases, which you can read here.   

The game is always changing, and we're lucky enough to be witnessing a new era, one that's going to be really short, because as pitchf/x and fieldf/x data becomes more incorporated (and it already is at the majors, probably the minors as well), it will be a brand new day. When teams can quantify to what spot on the diamond the ball goes on a given pitch, how the fielders react, etc., it will change the game to a new era--the Metric Era. 

No comments :

Post a Comment