HQ NOW: Experts Leagues Winner's Circle, 2023RESEARCH: First-pitch strike rates of starting pitchers
@StephenNickrand
June 4th, 2014 11:04pm
Our research in 2013 on swinging strike rates (SwK%) illustrated the strong correlation between a pitcher’s level of swinging strikes and one of the staple pitching metrics we have used for years—Dominance (K/9) rate. SwK% is a metric often used today to validate strikeout levels and to forecast the potential for a pitcher to experience a surge or decline in strikeouts.
Like Dominance rate, Control (BB/9) rate is another indicator in our toolbox that has driven our pitching roster decisions for a long time. In fact, it is a significant component of our base performance value (BPV) metric for pitchers. However, we haven’t been able to incorporate a more granular measurement to validate a pitcher’s control rate—nor anticipate changes in a pitcher’s future level of walks—using a comparable indicator to SwK% for strikeouts.
Our research here will show that first-pitch strike rate (FpK%)—the percentage of first-pitch strikes a pitcher throws—can serve this purpose.
Hypothesis
It seems intuitive that pitchers with a high FpK% would tend to have low control rates—and therefore lower WHIPs—than those with a higher FpK%. And perhaps pitchers who are allowing a lot of walks—even though they are getting a lot of first-pitch strikes—could be forecasted to expect a reduction in their control rate in the future, and vice-versa.
In fact, our initial research on stats and skills by starting pitcher ball-strike counts confirms the significant positive impact on a pitcher that starts the count 0-1. We found that ball-strike counts that started 0-1 resulted in a walk in just four percent of plate appearances.
Method
Let’s take a closer look at FpK% to see how strongly it is correlated with the common pitching metrics you will find at our site.
To do this, we took a look at starting pitchers that posted 40 IP or more per season from 2010 to 2013.
This threshold was reached a total of 775 times during this period.
As a reminder, correlations can range from +1.0 to -1.0. Let’s segregate them into the following groups to describe the correlation strength or lack thereof:
+0.70 to +1.00 Strong positive relationship +0.40 to +0.69 Moderate positive relationship +0.20 to +0.39 Weak positive relationship ------------------------------------------------- +0.19 to -0.19 No or negligible relationship ------------------------------------------------- -0.20 to -0.39 Weak negative relationship -0.40 to -0.69 Moderate negative relationship -0.70 to -1.00 Strong negative relationship
Results
From 2010 to 2013, the average FpK% of pitchers by type of pitcher and league were as follows:
Pitcher League FpK% ============ ====== ===== Starters AL 59.8% Starters NL 60.4% Starters MLB 60.1% --------------------------- Relievers AL 58.4% Relievers NL 58.8% Relievers MLB 58.6% --------------------------- All pitchers AL 59.3% All pitchers NL 59.9% All pitchers MLB 59.6%
The following indicators had positive correlations with FpK%, meaning that they had a tendency to move in the same direction as FpK%:
Indicator FpK% Correlation ========= ================ BPV +0.49 Age +0.19 Win Pct +0.19 SwK% +0.17 S% +0.13 Dom +0.13
Here is a graphical look at the above table:
Conversely, these indicators had negative correlations with FpK%, meaning they tended to move in the opposite direction of FpK%:
Indicator FpK% Correlation ========= ================ Ctl -0.62 WHIP -0.42 ERA -0.26 GB% -0.18 FAv -0.12 hr/f -0.11 HR/9 -0.05 H% -0.01
So we see that FpK% has the strongest correlations with the following three HQ metrics:
Indicator Type of Correlation Strength of Correlation ========= =================== ======================= Ctl Negative -0.62 BPV Positive +0.49 WHIP Negative -0.42
In summary:
- There is a moderate-to-strong negative correlation between control rate and FpK%, meaning as a SP’s first-pitch strike rate goes up, his walks are likely to go down.
- As control rate is a significant component in our pitching BPV calculation, it should not come as a surprise that there is a moderate positive correlation between BPV and FpK%. This means that as a starting pitcher’s first-pitch strike rate increases, so too will his BPV.
- Given that walks drive up WHIP, it is also logical that there is a moderate negative correlation between WHIP and FpK%, meaning a SP's WHIP will go down as his rate of first-pitch strikes goes up.
FpK% Stability
What about FpK% from season to season? How much is it likely to vary for starting pitchers?
A total of 82 starting pitchers threw at least 40 IP in each season from 2010 to 2013.
The average FpK% variance by starting pitcher from one season to another during this period was only +0.6%. Large increases in FpK% from one season to the next typically were offset by similarly large reductions within the same three-year period.
This confirms that FpK% does not regress towards league norms. It is in control of the pitcher.
We also wondered if FpK% tends to regress to a pitcher’s three-year rolling average, similar to how batters set their own hit rate level.
- 41% of starting pitchers tended to approach their prior season’s FpK% more than their three-year FpK% or career FpK%.
- 42% of starting pitchers tended to approach their three-year FpK% more than their prior season’s FpK% or career FpK%.
- Only 17% of starting pitchers tended to approach their career FpK% more than their prior season’s FpK% or three-year FpK%.
This means that a starting pitcher’s FpK% is much more likely to approach his prior season or three-year FpK% levels than his career FpK%.
Now let’s look at extreme FpK% changes from one season to the next.
There were 19 instances of SP whose FpK% increased by 5 percentage points or more from one season to the next from 2010 to 2013.
Our research found an extremely strong tendency for big FpK% surgers from one year to the next to keep most of those gains in year three, rather than regressing to the SP’s prior career FpK% norm:
-
FpK% Season-to-Season Increase > 5 percentage pts: 2010 to 2013 Name FpK% Year One FpK% Year Two FpK% Career FpK% Year Three ================ ============== ============= =========== =============== Lohse, Kyle 56% 68% 57% 69% Morrow, Brandon 53% 61% 53% 60% Hudson, Tim 54% 62% 44% 62% Capuano, Chris 57% 64% 60% 67% Guthrie, Jeremy 53% 59% 57% 60% Garza, Matt 58% 64% 57% 63% Hughes, Phil 63% 68% 61% 66% Sanchez, Anibal 58% 63% 56% 66% Sabathia, C.C. 58% 62% 55% 63% Westbrook, Jake 56% 61% 57% 61% Cueto, Johnny 56% 63% 57% 62% Happ, J.A. 56% 63% 57% 60% Buchholz, Clay 56% 63% 59% 60% Morton, Charlie 55% 61% 55% 59% Garcia, Jaime 58% 64% 56% 68% Arroyo, Bronson 62% 68% 57% 66% Gonzalez, Gio 53% 59% 53% 61% Cahill, Trevor 57% 63% 56% 60% Leake, Mike 58% 63% 58% 59%
If we expand the FpK% increase threshold to +3 points or greater, we find that 70 starting pitchers saw FpK% increases of 3.0% or higher from one season to the next between 2010 and 2013. Fifty of them (70%) experienced a reduction in their control rate during the same season with an average reduction of 0.7.
For guys whose FpK% fell by more than 5 points from one season to the next, all but one saw their FpK% rebound the following season, although it was slightly more common for their FpK% to revert to their prior career FpK% norms:
- FpK% Season-to-Season Decline > 5%: Period 2010 to 2013
Name FpK% Year One FpK% Year Two FpK% Career FpK% Year Three ================== ============= ============= =========== =============== Liriano, Francisco 62% 49% 57% 53% Johnson, Josh 65% 55% 61% 58% Buehrle, Mark 62% 56% 54% 61% Lee, Cliff 70% 65% 64% 72% Wood, Travis 59% 54% 59% 58% Danks, John 63% 56% 60% 62% Marquis, Jason 61% 54% 54% 54% Gallardo, Yovani 63% 57% 59% 56% Harang, Aaron 62% 56% 62% 59%
Expanding this FpK% decline threshold to -3 points or greater, we found that 40 starting pitchers saw such an erosion from one season to the next between 2010 and 2013. Twenty-four (60%) experienced an increase in their control rate during the same season with an average control rate increase of 0.8.
Predicting Control Rate from FpK%
We can forecast future changes of control rate for pitchers whose FpK% is out of line with a control rate normally associated with that level of FpK%.
This table shows the range of control rates (Ctl) over the last four seasons for different levels of FpK%.
Expected Ctl by Percentile
FpK% 10th 25th 50th 75th 90th
====== ==== ==== ==== ==== ====
>68% 1.1 1.3 1.7 2.3 2.7
66-67% 1.3 1.7 2.2 2.5 2.5
65% 1.4 1.6 2.1 2.4 2.9
64% 1.9 2.1 2.4 2.7 3.2
63% 2.0 2.2 2.5 2.9 3.3
62% 1.9 2.1 2.6 3.0 3.3
61% 2.0 2.3 2.7 3.2 3.5
-----------------------------------------
60% 2.1 2.4 2.9 3.4 4.0
-----------------------------------------
59% 2.1 2.6 3.0 3.4 3.7
58% 2.3 2.7 3.2 3.5 4.0
57% 2.1 2.8 3.2 3.8 4.5
56% 2.2 2.8 3.3 3.9 4.3
55% 2.8 3.1 3.8 4.3 4.8
53-54% 3.0 3.3 3.6 4.9 5.3
<52% 3.3 3.5 4.5 5.0 5.9
The table shows a steady erosion in control as a SP’s FpK% declines.
The 50th percentile data means that 50% of pitchers will have control rates below the value listed, and 50% of pitchers will have control rates above the value listed.
For example, a pitcher with a FpK% of 60% (average level for a starting pitcher) is expected to have a 2.9 Ctl. Only 10% of pitchers with a FpK% of 60% will have a 2.1 Ctl or lower, and only 10% will have a 4.0 Ctl or higher. Conversely, even the worst Ctl pitchers among those with elite FpK% of 66% or higher are still better than that 2.9 Ctl.
Let’s wrap up our findings by highlighting the takeaways of this research.
Conclusions
- There is a moderate-to-strong negative correlation between Control rate and FpK%.
- There is a moderate positive correlation between BPV and FpK%.
- There is a moderate negative correlation between WHIP and FpK%.
- Big FpK% surgers from one year to the next tend to hold on to those gains in the third year or revert to their three-year FpK% average rather than regress to their prior career FpK% norm.
- Big FpK% decliners from one year to the next tend to recoup those losses in the third year, but there is a slightly greater tendency for the decliners to revert back to their prior career FpK% norms.
- Levels of Control rate can be predicted based on levels of FpK%. For example, only 10 percent of pitchers with a FpK% of 65% will have a Control rate of greater than 2.9.
As we do with the SwK% metric when validating a pitcher’s Dominance rate, we can use FpK% to validate a pitcher’s Control rate. When a SP's first-pitch strike rate increases, his walks and WHIP are very likely to go down. As it goes down, walks are likely to increase, as will WHIP. So if you’re wondering if a pitcher’s newfound good control is likely to hold, check out his FpK%.
Case-in-point: Jason Hammel (RHP, CHC) is posting the best control of his career. He owns a 2.1 Ctl after 10 starts. Likely to stick? The chances of that happening are tiny. His current 54% FpK% actually is the lowest he has posted since his rookie season, and it’s a level strongly correlated with a Control rate nearly double his current mark.
Expect more studies and applications of FpK% in the coming months.
For more information about the terms used in this article, see our Glossary Primer.
