How Much of an effect does pitch tipping have on a pitcher’s performance?
Pitch Tipping is one of the oldest flaws in the game. Nevertheless, it appears to rear it’s head each postseason, with this year being no exception. Of course, this makes sense – the postseason is a time where intense scrutiny is placed on each player as teams try to gain a competitive advantage over each other.
You will likely recall that pitch tipping was used to explain the early postseason struggles (and subsequent revivals) of two prominent pitchers for the Boston Red Sox, Craig Kimbrel and David Price. It is against this backdrop that I want to explore the question: ‘to what extent does pitch tipping effect a pitcher’s performance?’
To start, let’s make sure we all have a common understanding of what is meant by the term Pitch Tipping. For the purposes of this exercise, Pitch Tipping refers to anything in a pitcher’s preparation or delivery that allows a hitter to determine the coming pitch type. This is distinct from the concept of sign stealing (also a prominent issue this postseason), where the catcher’s signs are deciphered and relayed to the hitter. While I think the overall effects of these two phenomena are likely the same, I will focus simply on Pitch Tipping here.
Before moving forward I want to quickly differentiate between two types of pitch tipping, which I will define as such:
- Selection – a tip that signals a particular pitch type is coming
- Elimination – a tip that eliminates one particular pitch type for the coming pitch.
This is a subtle distinction that I think is important for understanding this phenomena. To help grasp this, let’s work through some simple examples. Let’s say that a given pitcher has a devastating change up. And let’s say that during his delivery, this pitcher usually taps the ball in his glove multiple times during his warm up – except when he throws his change up.
If we put ourselves in the position of the hitter, if we see this pitcher not tap his glove during his wind up, then we know that the pitch will be a change up – we can select change up. Now let’s say in this same scenario as a hitter, I notice that the pitcher does tap his glove during his wind up then I can eliminate it from consideration.
Let’s now try to answer the question we posed above – what effect does pitch tipping have on a pitcher’s performance? We’ll start with the selection type of Pitch Tipping.
Recall that selection allows the hitter to determine that a single pitch type will be thrown. This same phenomena is observed commonly in a 3-0 count! In 2018, fastballs accounted for 95% of the pitches thrown in a 3-0 count. This seems like a pretty reasonable proxy for the selection type of pitch tipping.
So then what effect does this have on a pitcher’s performance? A simple way we can measure this is by comparing the league batting average in a 3-0 count relative to the league’s batting average overall.
In 2018 MLB hitters held a batting average of .248. In 3-0 counts however, that number jumps to .373
While batting average is a flawed measurement of performance, it does give us a good first insight into the effects of pitch tipping – how much more often does it lead to a hit? As a you can see above, a hitter who successfully selects the coming pitch is likely to get a hit an extra 12.5% of the time. This is huge! Over the course of a season, this works out out to something like an extra 65 hits. It’s the difference between being a league average hitter and winning the batting title.
So a tip that allows the hitter to successfully select the coming pitch type puts the hitter at a significant advantage.
What then of the elimination type of pitch tipping? The effect of this elimination type of pitch tipping is relative to the number of the pitches that a pitcher throws and can be shown like this:
.125 / probability(guess pitch)
Where .125 is just the difference between the baseline performance (.248 BA) and successfully selecting the coming pitch (.373). We can show this then in a simple table:
|3||2 x .125||0.062||0.310|
|4||3 x .125||0.042||0.290|
|5||4 x .125||0.031||0.279|
The most important thing here is the column on the right – the expected batting average column. As you can see, the greater the number of pitch types a pitcher employs the less vulnerable we could expect them to be to the effects of the elimination type of pitch tipping.
Of course, as we noted above, batting average is a poor measure of performance so lets recreate this table using a better measure – wOBA. Please note that the wOBA used here is excluding walks and hit by pitches as it would be skewed by the 3-0 count.
You will notice that I skipped pitchers with 2 pitch types in the table above as eliminating 1 pitch type in this case would select the other by default.
As you would expect, a similar pattern emerges to the one above when using batting average. Again, having multiple pitch types helps a pitcher to defend against the elimination type of pitch tipping.
So let’s try and bring this back into a real life example. Assuming that Kimbrel and Price were tipping their pitches AND that the hitters were able to recognise it, to what extent would we expect each pitcher to be effected.
Starting with Kimbrel, we know he throws just a fastball and curveball – 2 pitch types. From this then, we would expect a hitter to see a .219 jump in wOBA. In 2018, Kimbrel posted a wOBA .188. From this then we would expect the pitch tipping version of Craig Kimbrel to post a .407 wOBA. In contrast, David Price utilises 5 pitch types however, I will eliminate two of them here (a fastball and two-seam are very similar and the curveball was used less than 5% of the time in 2018). Working with three pitches then, we would expect a .110 increase in Price’s wOBA were he to tip his pitches. Having posted a .266 wOBA in the regular season, we could expect the pitch tipping Price to post a .376 wOBA allowed.
Pitch Tipping is an old flaw and it’s one that can be hard to avoid. Nonetheless, it can be devastating for a pitcher’s results.