The Anatomy of an At Bat

When you are watching sport you inevitably at some point ask yourself ‘What would I have done if it was me?’ or you say ‘They shouldn’t have done that, they should have done this’.  To me this comes from hours of watching games which puts knowledge in your head and you develop your own instinct.  This happens with players, coaches, managers and pundits as well.  What baseball offers in comparison to a large number of other sports is the chance to look at the numbers and the outcomes to see if our preconceptions are right or wrong.

Now there are lots of factors that take part in the decisions that a batter will make, who the pitcher is, what the score is, how many are on base, what is the stadium? But for this analysis we are going to talk about everything as the average at bat, meaning average pitcher, batter, stadium, pressure.  I took all of the MLB Statcast data from 2017 which gave me 725,641 moments to analyse and generate what would happen on average for each pitch count.

I have created three stats here all base around wOBA (weighted on base average), the reason why I have used wOBA over AVG or OPS as it is the most accurate representation of a players batting ability. They are called Outcome wOBA, Fair Ball wOBA and Through wOBA.  Outcome wOBA is defined as the average wOBA if the at bat ends on that pitch count, Fair Ball wOBA is the average wOBA if the at bat ends with a fair ball (i.e flyouts, groundouts, singles, doubles etc.) and Through wOBA is average wOBA of all at bats which have that count at some point of the at bat.  For all counts which cannot end in a walk or a strikeout Outcome wOBA and Fair Ball wOBA will be the same. Below is the wOBA weighting for each positive batting outcome in 2017 and the average wOBA for an at bat in 2017 was 0.321.

Pitch Count 0-0

Every at bat starts here and what is remarkable is that over 70% of these pitches are not swung at and that 31.8% of all first pitches are taken for a strike.  These percentages are second only to the 3-0 count and based on the figures for all the other counts there is premeditated approach here not to swing on the first pitch.  The Outcome wOBA is fair bit better than the Through wOBA which says to me people should try and put the ball in play more often then they currently are, some of those called strikes are hittable.  Taking that first strike could be costly as the 0-1 count is on average 49 points worse for expected wOBA going through that count.

Approach – if it is going to the edges of the zone take it, otherwise swing away.  Swing more than you do now.

Pitch Count 0-1

Just shy of 50% of MLB at bats in 2017 ended up here after the first pitch.  Although this is described as a pitchers count if the batter gets the ball in play on this pitch then they will do better than average.  If you take a ball here and get to 1-1 you are output on average will be just below 0-0 but if you get a strike you go way below average.  The main change from the 0-0 count is far fewer pitches are taken for strikes, batters are seeing those pitches and attempting to put them in play.  37% of batters that swing get in play but 25% miss and 80% of those who don’t swing get a ball, so a good read of the pitch makes a huge difference here.

Approach – No major change just slight increase in swinging at those called strikes.

Pitch Count 0-2

Down 0-2 is the worst place to be for a hitter to be and 20% of batters end up here which is a larger percentage than I was expecting, also 3.5% of all at bats get here by watching 2 strikes.  Which means that at the average game last year one player for each team would watch two pitches and end up 0-2.  The called strike rate drops to 3.7% and the foul ball rate is lower than quite a few of counts below, this is a sign to me that the batter has his eye in by now.  So they are good at leaving here but they are bad at putting the ball into play, 34% of swings put the ball in play which is the lowest of any count.  And even if you do get the ball in play then the wOBA for those hits is lower than any other count.

Approach – Take ball if you can, foul the ball if you swing and repeat till you get a ball.  Little value on ball in play best approach is to get out of the count.

Pitch Count 1-0

So they took a ball on the first pitch, well done to the 40% who get here.  57.6% of batters still don’t swing here, remember they didn’t swing on the first pitch, but only 60% of them get the ball that they were hoping for.  So looking for a second ball seems to be the preferred approach but of those who do swing 38% put it in play with a .411 wOBA which is excellent.  The called strike rate of 23% is too high for me and all but confirms a premeditated approach to the at bat.  Yes, as we will show later that 2-0 is better than 1-1 but good hittable balls are missed here.

Approach – if it is going to the edges of the zone take it, otherwise swing away.  Swing more than you do now.

Pitch Count 1-1

Ok, so the batters that have a ball and a strike but they are not that much worse off than 0-0, through wOBA down by 15 points.  But it looks like players treat this at bat very differently.  Batters swing on 53.4% of these pitches compared to 28.7% at 0-0 with most of that difference coming from called strikes (11.2% compared to 31.8%).  But when they do swing they put it in play 37.7% with a wOBA of .397, compare that to 0-0 (37% and .399) and you will see that they in fact are very similar.  To me what you have here is what would happen on the first pitch if the batter didn’t go in with the idea of taking the first ball.

Approach – No major change just slight increase in swinging at those called strikes.

Pitch Count 1-2

This is the most likely pitch count to be in after 3 pitches. 56% of at bats get to a fourth pitch and of them 49% end up here at 1-2. This is the second worst outcome to be in during an at bat, all three wOBAs are slightly better for this count compared to 0-2 but not for any other counts.  In this count, compared to 0-2, batters are slightly more aggressive with 7% more pitches being swung at, they make contact more often and get the ball into play more often. It still only just over 20.7% but it is larger than the strikeout rate of 18.3%. Your more likely to be hit by pitch in this count that any other but it is still less than 0.5%.

Approach – Take ball if you can, foul the ball if you swing and repeat till you get a ball.  Little value on ball in play best approach is to get out of the count.

Pitch Count 2-0

With a good eye and some defensive/errant pitching 13.5% of batters end up here and here you get a ‘free’ pitch. All the wOBAs are very similar which shows that what a batter chooses to do here has little to no impact on the outcome. More than half of the batters here still choose to leave the pitch but like the 0-0 pitch over around 44% of them take a strike. Which is part of why this pitch count sees a higher percentage of strikes than all other counts, bar 3-0. As all other counts with no strikes there is a premeditated approach of not swinging. This is one of only two counts where the ball is more likely to go into play than be fouled off.

Approach – No suggested change, numbers paint everything as even.

Pitch Count 2-1

At 2-1 in the count the batters is in a better position than where they were when they started but this is a very pivotal moment in this at bat. A strike puts you at 2-2 and decreases the through wOBA by 89 points but a ball takes you to 3-1 propels you through wOBA up by 100 points. That difference is the biggest and it changes the batters approach compared to all other counts where you cannot be walked or struck out. The batters are more inclined to swing, only 2-2 and 3-2 have higher swing rates. While the in play wOBA is a very respectable .416, batters swing and miss 21% of the times.

Approach – If it is going to the edges of the zone take it, otherwise swing away.  Swing less than you do now.

Pitch Count 2-2

37% of at bats face a 5th pitch and 55% of those pitches are at 2-2. But remarkably that only accounts for 2/3 of the 2-2 pitches, thanks to foul balls on two strike counts there was a further 21,000 pitches with the count at 2-2 after the 5th pitch. This isn’t a great position for batter to be in and like all of the other two strike counts the swing rate and the ball in play rate are high, only 3-2 is higher. Of the those batters who swing 38% put the ball in play but the fair ball wOBA is the 3rd worst.

Approach – Take ball if you can, foul the ball if you swing and repeat till you get a ball.  Little value on ball in play best approach is to get out of the count.

Pitch Count 3-0

Only 4.2% of at bats in 2017 get here, they haven’t swung in the previous 3 so it should be no surprise that 89.4% of the time they don’t swing here. With a called strike rate of 56.8% I was all prepared to rip into this approach say they should swing more. The fair ball wOBA is a very high .514, highest of all at counts, but that is worse that value of a walk, .693.  While they do they make contact better than any other count, only 14% of swings miss.  But the scenario you should aiming for here is the walk, that on average is where the value is. And 56% of all batters who get to 3-0 eventually will be walked, with 32.6% doing it on this count and a further 23.5% via 3-1 and 3-2.

Approach – None. I cannot believe I am stating this but maybe they should swing less. Some people have to swing otherwise pitchers would just always through strikes but unless the pitch comes right into a batter’s wheelhouse they should leave it.

Pitch Count 3-1

70% of batters come to this count from 2-1 and there isn’t too much difference in approach here from the 2-1 count (<1% change in swing rate) but there is a massive change from 3-0. The swing rate increase from 10.5% to 57.3%, that is much more standard approach after more than likely taking a strike in the 3-0 count. The contact rate and fair ball wOBA are only second to the 3-0 but still fall short of the wOBA for a walk. Chance of a walk in this at bat are now 37.4%.

Change to Approach – No major change just slight decrease in swinging, which would lead to more walks.

Pitch Count 3-2

We have reached the full count and here is where batters are the most aggressive, swinging 72.5% of them time. Like the other 3 ball counts the swing and miss much lower than average, this is due to the pitcher also having to be fairly aggressive as well. Any ball means a walk. What is remarkable here is that fair ball wOBA is very close to the other two meaning that there is no predetermined method which would give better results. A larger percentage of pitches end up as foul balls here with players fighting to avoid the strike out.

Change to Approach – No major change just slight decrease in swinging, which would lead to more walks.

Pitch Count 4-2

Wait, what do you mean 4-2 surely that is already a walk? Yes, but you have to remember that every human makes mistakes and baseball is still controlled by humans. On the 8th August 2017 Yoan Moncada managed to earn himself a 5 ball walk. In the bottom of the 4th with 2 outs and nobody on, Moncada worked through a 7 pitch at bat against Dallas Keuchel and draw his 4th ball but instead of walking to first he stayed, fouled off one pitch and then took another for a ball to earn his walk. We see one of these on average every 2 years.

Now this is all a simple statistical view as we haven’t looked at the type of pitches but I hope it gets you thinking the next time you are watching a game.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.