Undervalued Hitter Study

Essentially, we believe that differentiating a hitter’s offensive numbers based on the quality of pitcher faced reveals a hitter’s value that would otherwise remain hidden, as well as be an indicator of future increase or decline in performance. Hitters that perform well against better pitchers yet put up marginal numbers overall become more valuable in the playoffs because a substantial amount of innings pitched in the playoffs are by better pitchers. Identifying young hitters who have performed well against good pitching but have subpar numbers elsewhere might be an indicator of a hitter on the verge of a breakout season. Conversely, identifying hitters whose performance against the top pitchers has turned subpar may be an indicator of a hitter ready to decline.

To give you an idea for how to read and use these statistics, we have provided an example. Here are the both Dustin Pedroia’s and Ty Wigginton’s offensive statistics from 2006 to 2012.

Dustin Pedroia:

PA	AVG	OBP	SLG	OPS
3824	.303	.369	.461	.830

Ty Wiggington:

PA	AVG	OBP	SLG	OPS
3410	.263	.324	.443	.767

Now look at Pedroia’s and Wigginton’s offensive statistics for those same seasons except differentiated by the quality of pitcher faced.

Dustin Pedroia:

Pitcher Rating	PA	AVG	OBP	SLG	OPS
1	321	.260	.307	.368	.675
2	873	.277	.344	.422	.766
3	1410	.303	.364	.453	.817
4	934	.328	.396	.520	.916
5	286	.347	.434	.547	.981

Ty Wiggington:

Pitcher Rating	PA	AVG	OBP	SLG	OPS
1	295	.237	.274	.417	.691
2	759	.254	.314	.394	.708
3	1373	.272	.330	.477	.807
4	777	.260	.321	.432	.752
5	206	.285	.343	.473	.816

When looking only at Pedroia’s and Wigginton’s overall numbers, Pedroia is clearly superior. But, when looking at each player’s differential hitting statistics, both Pedroia and Wigginton have been very similar when facing the top three pitching tiers. Pedroia’s better overall offensive numbers are a product his superior ability to perform well against weaker pitchers. While we are not arguing that Pedroia and Wigginton are equals, we believe the differential hitting statistics show that in certain situations, both players are very similar. This becomes meaningful in the playoffs, for example, where hitters are far more likely to face pitchers from the top three pitching tiers.

Along with an explanation of the methodology used and hitter database, are two studies that aimed at identifying undervalued hitters and examining how these differential statistics affect a team’s playoff performance. The “HITTER SEARCH” tab serves as a platform for the database, which includes every hitter’s statistical performance against each pitcher rating for each of the last six seasons. It will allow you to manually enter a player’s name or select any player from the drop-down menu. All of that hitter’s statistical data over the last six years will be displayed along with their combined performance during that span. The “TEAM SEARCH” tab allows you to see each team’s performance from 2005 to 2013 against each pitcher rating. Each team’s line is a compilation of all the hitters for that team during the given year. For a perspective on how hitters and teams performed compared to the rest of the league, the league averages for each season was compiled and displayed in the “LEAGUE AVERAGES” tab.

Explanation of Methodology

This study differs from others in many regards. Existing stats are available that separate pitchers into tiers according to breakpoints in stats.  For example, one can find a stat that states “Hitter X has a 0.330 average against pitchers in the top 10% of era”.  This is indicating that a hitter hits .330 against a tier of pitching relative to ERA. Our study attempts to get to the root of the question, “What is a good pitcher and how hard is it to get a hit off him?” by creating an algorithm that attempts to blend a subset of statistics and subjective qualities to rank pitchers.

Anyone who has played baseball knows that stats aside, some pitchers are very hard to hit and others are not. Hitters with the ability to succeed against the best pitchers in the game give them a quality that few hitters own. We wanted to develop a metric that would determine which hitters display this ability. What we came up with was a system that would identify every pitcher into subsets based on their ability and then compile each hitter’s performance against these subsets. Before we could begin to evaluate hitters, we needed to determine a strategy for which we would rate each pitcher. The method that we decided upon utilizes a variety of stats and ratings to form a composite end result that takes into account several parameters that define a pitcher’s ability. It combines a series of statistics sorted using both bell-curve and power-law techniques and places a weight variable that is inherent in pitching.  Nothing is exact, however, to steal from Warren Buffet and John Maynard Keynes, “We would rather be vaguely right, than precisely wrong.”

We are not claiming any extraordinary new method to determine the value or ability of a pitcher, we simply derived a method to incorporate statistics that define what makes a pitcher tough to hit. Our system uses similar parameters that are common in many current statistics that evaluate a pitcher’s performance today, such as the FIP method. The methodology in which we rated pitchers is then applied to a hitter’s performance in order to compile a set of what we would call “differential statistics”. We are attempting to identify hitters that are either undervalued or overvalued according to how they perform against different tiers of pitching. Several techniques, constants and proportions had to be juggled in our algorithm; however the end result did not drastically swing the outcome of expected results in most cases. Many of the results indicated what could be reasoned by simple intuition. However, there are some players that tend to stick out in both directions. Some of these players are well known and some are not. Knowledge of both sets of players offers value in the baseball marketplace.

Before any hitters could be evaluated, the first step that needed to be completed was to determine a mechanism to rate every single pitcher who threw a pitch during the past five seasons. This was accomplished by rating them on a scale from one to five (1 being the best, while 5 being the worst). Every season was treated independently in order to secure an accurate representation of a pitcher’s ability at the current time. As stated earlier, the methodology for our rating system uses similar metrics to common pitching statistics. What we aimed to do was isolate the statistics that best define a pitchers ability to get hitters out. This is where we were forced to make a few assumptions. We ultimately decided upon four different statistics: strikeouts, walks, home runs, and hits allowed. While walks and strikeouts are universally considered to be measures in which to evaluate a pitcher’s ability, home runs and hits allowed not always are. Some believe that a pitchers ability to limit home runs is simply a factor of their fly ball percentage of around 11 percent. We believe to the contrary. Several pitchers perform well above or below that mark on a consistent basis. For this reason, home runs allowed were incorporated in our rating system.  Hits allowed by a pitcher are generally regarded as a component of luck but we felt that it also provided an indication of how effective a pitcher was during the course of a season. The level of importance of each of these statistics was controlled by an importance factor within our computation. Every statistic we used to was regulated on a per nine innings basis so a pitchers can be compared evenly. Baselines were set up for four the different statistics, K/9, BB/9, HR/9, and H/9. These statistics were chosen based on several factors, but were ultimately chosen based on their correlation with our definition of a good pitcher. These baselines were set at 20% thresholds for each statistic; meaning the top 20% of pitchers with the best K/9 would receive a rating of 1. Rating tables were used to compile a set of parameters for each applicable statistic. Adjustments were made for starting pitchers and which league they played in. Then each of these statistics was multiplied by the previously mentioned importance factor that was determined subjectively. Strikeouts and walks allowed were considered to be of greatest value and were assigned the largest importance factor. Again the importance factors are human generated but attempt to take into which stats need to be corrected and weighted according to the definition of pitching. The numbers were then summed together to give each pitcher a rating between 1 and 5.

After the rating system was applied to each pitcher during the last seven years, we began to compile hitters’ statistics against pitchers of each rating. We have done this for seasons 2005 through 2013. As we continue to analyze hitters, we are hoping to find correlations or trends. There are two aspects of this study that we would like to look at. The first being, individually analyzing hitters; by identifying hitters that perform well against the top pitchers and those who perform poorly against these same pitchers. We would also like to look at teams as a whole, particularly playoffs teams, and look to see if there is any correlation between a team’s performance in the playoffs and how they did against the top pitchers in the regular season. Generally, teams see better pitching in the playoffs; which leads us to believe that teams with hitters that perform well against “good pitching” will be more successful in the playoffs.   

Abstract of Analysis

What we were attempting to do with this project is to show that there are in fact hitters that perform at a superior level against better pitching than other hitters. We wanted to expand on the notion that certain particular players hit quality pitching better than most other hitters. This certain quality that these types of hitters have may not always show up in their overall stat line. For example, a hitter may produce an OPS of 0.800 overall, but that hitter has an OPS of 0.800 versus all types of pitchers. Whereas most hitters follow the standard league average curve and produce an OPS of 0.600 against the best pitchers in the league and an OPS of 1.000 against inferior pitching; causing their overall stat line to have an OPS of 0.800. To show that this case does in fact exist, we had to come up with a way to compare every player to the rest of the league. We decided that looking at a hitter’s OPS compared to the league average OPS versus each level of pitching quality would be the best way to do this.

To verify the consistency of our method of determining pitcher ratings, we looked at the league averages of a hitter’s stat line for each year of compiled data. This was done to ensure that a hitter’s performance, on average, would differ based on the type of pitcher they were facing. It would also verify whether there was a correlation between the league’s performance and different tiers of pitchers from year to year. This was done compiling every hitter’s stat line against each different pitcher type for each year. The following are the league averages versus each pitching tier:

2006:

Pitcher Rating	OPS
1	.661
2	.710
3	.785
4	.854
5	.946

2007:

Pitcher Rating	OPS
1	.627
2	.695
3	.773
4	.843
5	.928

2008:

Pitcher Rating	OPS
1	.631
2	.695
3	.752
4	.840
5	.922

2009:

Pitcher Rating	OPS
1	.618
2	.696
3	.749
4	.824
5	.942

2010:

Pitcher Rating	OPS
1	.602
2	.679
3	.760
4	.835
5	.928