<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Paul Sirena on Medium]]></title>
        <description><![CDATA[Stories by Paul Sirena on Medium]]></description>
        <link>https://medium.com/@paulsirena?source=rss-7b7cf8b49a94------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*aN67Ig8q8acmQBGxiUMCyw.jpeg</url>
            <title>Stories by Paul Sirena on Medium</title>
            <link>https://medium.com/@paulsirena?source=rss-7b7cf8b49a94------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 01:30:19 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@paulsirena/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Identifying Optimal Matchups with Simulated Plate Appearances]]></title>
            <link>https://medium.com/iowabaseballmanagers/identifying-optimal-matchups-with-simulated-plate-appearances-513c7992d5bd?source=rss-7b7cf8b49a94------2</link>
            <guid isPermaLink="false">https://medium.com/p/513c7992d5bd</guid>
            <category><![CDATA[college-baseball]]></category>
            <category><![CDATA[iowa-baseball]]></category>
            <category><![CDATA[iowa-baseball-managers]]></category>
            <category><![CDATA[baseball-analytics]]></category>
            <dc:creator><![CDATA[Paul Sirena]]></dc:creator>
            <pubDate>Tue, 18 May 2021 16:30:24 GMT</pubDate>
            <atom:updated>2021-05-18T16:30:24.356Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Note: This post was co-authored by </em><a href="https://medium.com/@apalensky"><em>Al Palensky</em></a><em>.</em></p><h3><strong>Introduction</strong></h3><p>A team with the ability to know the outcome of a plate appearance before it happens would have the perfect strategy when picking a reliever or pinch hitter to put in the game. This is obviously not possible, but there are ways to optimize matchups.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*khy3KTiSKuHEI02L.jpg" /><figcaption>Adam Dunn had success against Clayton Kershaw.</figcaption></figure><p>To make this easier to think about, we can consider major league players. If Clayton Kershaw must face either Mike Trout or Adam Dunn in a hypothetical situation, looking at stats will show that Mike Trout has struggled against him with three hits in 18 at-bats. Meanwhile, Dunn did very well against him with eight hits in 13 at-bats and four home runs. However, a manager would never play Dunn over Trout even if these stats suggest it is the best idea. Another example is looking at batter handedness to make decisions. Assuming that any right-handed hitter is a better matchup than a given left-handed hitter against any left-handed pitcher is not always true. A team’s scouting process can be greatly improved by quantifying the process of matchup optimization.</p><p>This task began with clustering pitch types to see how the hitter performed against similar pitchers. The idea of this method is to group pitches together based on factors such as velocity, movement, and release point and assign labels to pitches with given qualities. Luckily, this work had already been done last year by former Iowa data analyst and current Driveline employee, <a href="https://twitter.com/aosnacz">Anthony Osnacz</a>. Thanks to him, we can assign a cluster to each pitch in our database. For example, two right-handed pitchers with a 92-mph fastball, similar release points, and equal movement would be grouped together. Using this data, it became possible to model different matchups and simulate what would happen before the games took place.</p><h3><strong>Methodology</strong></h3><p>This project had two attempts. The first was last spring and ended up having room for improvement. The goal was to simulate every pitch of every plate appearance. Each pitcher and hitter matchup would be combined to find the probability of each pitch type being thrown in each count. After the pitch type was predicted, we would then find if the pitch was in the zone. This would then determine if the batter swung, if contact was made, and if the contact was over 95 miles per hour. This was completed and showed some promise. Play-by-play data was available using this method with a pitch type, location, and outcome available. However, the final stats also ended up featuring far too many strikeouts and walks. With this in mind, we decided the best option was to redo the methodology using a more simplified method that did not require every pitch to be documented.</p><p>The second attempt, which was completed with fellow analyst Al Palensky, created a more effective method for simulating plate appearances. The first step was the same as the project’s previous attempt. We found every cluster and group thrown by a given pitcher and combined it with the hitter’s stats when facing that pitch type. These group of stats include the swing, contact, and hard-hit rates for pitches inside and outside of the zone. Next, the percentage of strikeouts among all walks and strikeouts was calculated. The final variable, which would become the target variable in a future model, is the percentage of plate appearances ending with the given pitch type and a ball in play. A success is when this value is over 50% and a failure to put the ball in play is when this is below 50%.</p><h3><strong>Model</strong></h3><p>Once the data was prepared, we were able to train a random forest model. With the first step, the goal is to see results of all pitches from each specific group. Some pitches may end up being out of the zone more often and some may end up getting swung at more often league wide. The specified pitcher’s stats with each of their pitches were then put into the created model. Here is an example of what this would look like:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*WD9zJfeg5vekZs4OIRmrRA.png" /></figure><p>This gives us an idea of how this pitcher has performed with these five pitch clusters. The next step is to find these stats for a given hitter. The hitter’s stats are calculated very similarly based on the clusters that they have faced. Here is an example of this output, with pitcher handedness specified:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*aVO90uAFS714dgOPQB9UPw.png" /></figure><p>You will notice that these are the exact same stats. There is only one addition to the hitter stats, which is batted ball statistics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*sdLS4gu2lYYCv7nL72_ljQ.png" /></figure><p>If the model predicts that a plate appearance will result in a ball in play, these three variables will be generated under the assumption that they all follow a normal distribution. These two tables give us all the information we need to move onto the next step, which is to combine the two tables.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*vpt-1tbFfOYv9iIZFAwH3w.png" /></figure><p>The idea here is to balance out the strengths and weaknesses in each matchup. A hitter with a good K/BB ratio will end up striking out more than normal against a pitcher who accumulates more strikeouts, while a hitter with a bad K/BB ratio may end up improving this number against pitchers who do not get a lot of strikeouts. Usage is normalized from zero to one, so the most frequently used pitch is given a one and the least frequently used pitch is given a zero with everything else in between.</p><p>Now that this table has been created for the matchup, we are ready to simulate. As mentioned previously, the variable being predicted is whether the plate appearance ends in a ball in play or not. If a ball in play is not predicted, a strikeout or walk is assumed. This will be based on a randomly generated value where the probability of a strikeout is also calculated using the pitcher and hitter stats, while one minus that is the walk probability. If a ball in play is predicted, the exit velocity, launch angle, and hit direction are generated based on a normal distribution and the values are put into another model that predicts the outcome.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*er5N48JLgM_RINibU1SHMQ.png" /><figcaption>HHB is hard-hit balls, HH Rate is HHB / Balls in Play</figcaption></figure><p>Here is an example of the output when plate appearances are simulated. You can see that the model resulted in 260 out of 300 plate appearances ending in a ball in play with 51 strikeouts and 40 walks. The hitter’s projected line is .300/.393/.650 with a 1.043 OPS, 30.6% hard-hit rate, and 17 home runs (17.7 AB/HR). This matchup is unfavorable for the pitcher with the excellent stat line.</p><h3><strong>Application</strong></h3><p>Adding projected matchup results to advance scouting adds an invaluable dimension to our game plan. It is important to not be discouraged by results if the matchup is not predicted correctly in one instance, since the idea is to project stats over a large sample. Before playing against a team, we can simulate plate appearances against every hitter or pitcher with enough data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*T3NBMbHsWZsgLGWzkKcQ9A.png" /></figure><p>This is a sample taken from a pitcher’s simulations against a team’s lineup. Each row represents a hitter on the opposing team and their stats against the given pitcher. The third row immediately stands out with zero hits in 300 plate appearances. When this happens, it means that there was not enough data on the hitter for the model to be useful. These hitters must be removed from the final product, unfortunately. The other four hitters’ projections indicate that they would be tough matchups for this pitcher. The fourth hitter would represent a matchup where one of the three true outcomes can be expected. If this pitcher faced this hitter, we would be risking the possibility of a home run or walk for an increased chance at a strikeout.</p><p>With plate appearances simulated and stats calculated, the final product can now be produced. We would generally view an OPS under .750 as a favorable matchup for the pitcher and an OPS over .850 as a favorable matchup for the hitter. Anything in between that can be viewed as a neutral matchup or adjusted as needed if there are a lot of home runs, strikeouts, or some other trend.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*r_I5NozsG1Tu66OGn4DORw.png" /></figure><p>This matrix is an example of what the final product looks like with a team’s pitchers as columns and opposing team’s hitters as rows. It has been color-coded to make favorable, unfavorable, neutral, and insufficient matchups easier to find. Insufficient matchups are the blacked-out rows that did not have enough data to create a meaningful projection. When hitters do not have enough data against any pitcher or pitchers do not have enough data against any hitter, they are removed. Some hitters are obviously far more favorable than others, but we can see who might be favorable against some of the better hitters. Neutral matchups are not the best scenario, but they are better than unfavorable matchups.</p><h3><strong>Limitations and Biases</strong></h3><p>There is one major limitation that the first attempt addressed much better than this one. While we are modeling the outcome of plate appearances, we do not have any idea how many pitches each plate appearance lasts or how the hitter performs in different counts. Finding a way to combine the two methods so that we can predict each pitch and find results based on the count would be a great addition.</p><p>There is also the issue where a first round prospect might throw pitches in the same cluster as a given pitcher against a hitter and cause the projections to be lower. Another addition would be to adjust hitter performance based on the quality of pitching they faced. Right now, all pitchers who throw a pitch in each cluster are assumed to be the same. While this was good enough for creating meaningful output, the model could be improved further with this adjustment.</p><h3><strong>Conclusion</strong></h3><p>Even with its limitations, we have created an excellent tool for projecting matchups before they happen. If there is indecisiveness over who to bring into a game at any given point, this matrix can be analyzed to find the best matchup. This approach to simulations has proven to be useful for predicting the outcome of matchups.</p><p>We are very excited to have simulations added to our growing list of tools created by our analytics department. It will be exciting to see how the simulations can be improved in the future and how they perform the rest of the season.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=513c7992d5bd" width="1" height="1" alt=""><hr><p><a href="https://medium.com/iowabaseballmanagers/identifying-optimal-matchups-with-simulated-plate-appearances-513c7992d5bd">Identifying Optimal Matchups with Simulated Plate Appearances</a> was originally published in <a href="https://medium.com/iowabaseballmanagers">Iowa Baseball Managers</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quantifying Swing Decisions: SDS]]></title>
            <link>https://medium.com/iowabaseballmanagers/quantifying-swing-decisions-sds-d59ccb84d820?source=rss-7b7cf8b49a94------2</link>
            <guid isPermaLink="false">https://medium.com/p/d59ccb84d820</guid>
            <category><![CDATA[batter-swing-decisions]]></category>
            <category><![CDATA[baseball-analytics]]></category>
            <category><![CDATA[iowa-baseball]]></category>
            <dc:creator><![CDATA[Paul Sirena]]></dc:creator>
            <pubDate>Mon, 12 Oct 2020 19:00:28 GMT</pubDate>
            <atom:updated>2020-10-12T19:00:28.918Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Note: This post was co-authored by </em><a href="https://medium.com/@luke.statler"><em>Luke Statler</em></a><em>, </em><a href="https://medium.com/@brookearnold00"><em>Brooke Arnold</em></a><em>, and </em><a href="https://medium.com/@sumner2799"><em>Andrew Sumner</em></a><em>.</em></p><h3><strong>Introduction</strong></h3><p>Making desirable swing decisions can separate a good hitter from a great hitter. A simple way to accomplish this is swinging at pitches that generally result in positive outcomes and laying off pitches that generally result in negative outcomes. The basis of a hitter’s decision to swing or not to swing focuses solely on if the decision typically leads to a successful outcome. That is, the result of a given swing decision does not influence the score that the hitter receives for that specific instance. Therefore, even if a hitter swings through a pitch in the middle of the zone, they will be awarded with a positive swing decision score. Although some hitters may not have the best results, an exceptional swing decision score may show that the hitter has untapped potential to unlock. Additionally, this metric can be used alongside expected outcome statistics such as xBA or xSLG to show that hitters may be benefiting from good luck or experiencing a case of bad luck.</p><h3><strong>Methodology</strong></h3><p>There are already simple ways to measure plate discipline, such as in-zone or out-of-zone swing percentage, but these metrics only cover plate discipline from a basic perspective. When we first set out to quantify swing decisions, our goal was to create a system that eliminates the multi-faceted aspect of these metrics by developing an all-encompassing score for each pitch.</p><p>To do this, we used various pitch metrics to create models that gave probabilities for called strikes, swings, contact, and hard-hit balls. These probabilities are then used to calculate the Swing Decision Score, which is also known as the SDS. The score is converted to a scale between –100 and 100. We can then compare the value to the average of every pitch, which gives SDS+. Along with looking at every pitch, we can look at different counts and determine how a hitter performs compared to the average in that count type.</p><h3><strong>Model</strong></h3><p>For this model, we decided that the random forest method was the best way to get a useful probability. The first step was randomly putting half of our data in a training set and half into a test set. From the training set, non-swings went into the called strike probability model, all pitches went into the swing probability model, swings went into the contact model, and all balls put in play went into the hard-hit ball model.</p><p>A variable to determine if the pitch is in a given zone determined by the number of strikes was the first aspect to go into the model. The zone that determines this variable gets larger as the number of strikes increases. It is important to note that this is not the actual strike zone. With zero strikes, we use a zone slightly smaller than the strike zone, with one strike we use the strike zone, and with two strikes, the zone is slightly larger. These are the other variables that went into each model:</p><p><strong>Called Strike: </strong>Pitch Location</p><p><strong>Swing, Contact, Hard-Hit:</strong> Pitch Speed, Pitch Movement (induced vertical and horizontal), Pitch Location</p><p>In the model, the contact probability is calculated on the condition that a swing occurs, and the hard-hit ball probability is calculated on the condition that contact is made. This results in additional necessary work to produce the true probability for each pitch. Consider the conditional probability formula:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/261/1*SoemTcXEtEf4D1bmXIKqnQ.png" /></figure><p>The probability we want to use to calculate SDS is P(Swing &amp; Contact), so we take the contact probability from the model and multiply it by the swing probability. The same can be done for hard-hit ball probability:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/353/1*Oj71k71vnEcgoQk2TiJc3w.png" /></figure><p>For SDS, we will be using P(Contact &amp; Hard-Hit Ball). This means we can simply take the hard-hit ball probability from the model and multiply it by the contact probability calculated in the first example. Contact probability is not actually used to calculate the score, but it is a necessary step to get the hard-hit probability that does go into the formula.</p><h3><strong>Calculation</strong></h3><p>The final called strike, contact, and swing probabilities are then implemented into our formula for SDS based on whether the hitter swung or not. On swings, we chose to reward hitters for pitches likely to be called strikes, likely to be swung at, and likely to be hit hard. On takes, we chose to reward hitters for pitches likely to be called balls and not likely to be swung at. The equations for calculating this score are as follows.</p><p>Given a swing:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/309/1*zZ-943Ny23ddktiUYyyskA.png" /></figure><p>Given a take:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/423/1*JWorK1Q_cNKe0BrzzP_QFQ.png" /></figure><p>Furthermore, these formulas are weighted heavier on two-strike counts, because of the significant difference between a good and bad swing decision in these counts. A poor decision can lead to a strikeout whereas a better decision leads to a ball in play or continuing the plate appearance by putting more pressure on the pitcher to throw a strike. For example, the hitter will receive a worse score if they take a pitch down the middle for a strikeout than if they get called for a strike in the same spot with the first pitch of the plate appearance. The formulas above are adjusted as follows.</p><p>Given a swing:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/332/1*01e_nkGBcLQOT-3vlmybAw.png" /></figure><p>Given a take:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/449/1*ndeN5U7MW_M6ukL_S5_FOw.png" /></figure><p><em>Note: In our initial research, we were inspired by Matt Boyd’s work on quantifying plate discipline. Check out his article </em><a href="https://mattboyd1.wixsite.com/website/post/plate-discipline-score-can-plate-discipline-be-quantified"><em>here</em></a><em>.</em></p><h3><strong>Limitations and Biases</strong></h3><p>Understanding that common approaches are often taken in certain situations and counts by hitters required us to make some adjustments to the model as well. In order to avoid additional bias, we removed 3–0 counts due to the large volume of high strike and high swing probability pitches taken in these situations. It would not make sense to give a hitter a poor SDS for such a routine and frequent strategy in baseball.</p><p>A similar bias came into play when discussing the number of hitters that routinely take the first pitch of every at bat. Our metric seems to punish hitters with this patience-oriented approach, where we noticed a significant number of hitters that received a low SDS score. However, we decided to leave 0–0 counts unadjusted, figuring that many of these patient hitters would be compensated for moving further into counts when weighting two-strike decisions more heavily than other counts, and thus making up for the punishment received from habitually taking a first pitch strike.</p><p>One issue that consequently comes up is lower scores for patient hitters both on the first pitch and in a hitter’s count. Some hitters may not swing at first pitch strikes that could be hit hard, and so their score would be very low. However, we found that these hitters usually tend to do better with two strikes and when behind in the count, which raised their score. Aggressive hitters who also have a good two-strike approach ended up with the highest SDS+ ratings.</p><p>Additionally, hard-hit probability was removed from the model during any two-strike count to account for the inability of hitters to be picky in such situations. With two-strikes, the approach of the hitter changes to one where their main goal is to protect the plate and avoid striking out, as opposed to looking for a good pitch to hit hard.</p><p>Some pitches were deemed to be “without-a-doubt” take situations, due to not requiring a swing decision at all, in which we labeled them “waste pitches”, and removed them completely from the model.</p><h3><strong>Example</strong></h3><p>We will now take a look at a hitter, who will be referred to as Hitter A. This is an example of a player who is regarded as a good hitter but punished by our metric for not swinging at enough pitches that can be hit hard. We start by looking at their takes with zero strikes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/694/1*gDtK4vac4ocqF5wjbwFJWw.png" /></figure><p>The waste pitches can be seen with scores of zero. This hitter took a lot of pitches in the zone, so their score ended up being low.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/697/1*T7sTvChrpf2NSurUld1a6w.png" /></figure><p>With one strike, there are still a lot of takes that could have been hit hard. The score still ends up being below average.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*oy4jD1gatKymeMFNJjLR2A.png" /></figure><p>The increasing strike zone is apparent when all three plots are looked at. With two strikes, Hitter A managed to not take too many pitches that could have been hit hard. This results in an above average score with two strikes. The table below shows the hitter’s scores.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/562/1*aChzN4re8pBj-lUUDWm93w.png" /></figure><p>Hitter A did not swing at the majority of first pitch strikes and was punished for it. With two strikes and when behind in the count, his approach was very good. These raise his score but not enough to grade him as an above average swing decision maker.</p><h3><strong>Conclusion</strong></h3><p>Despite the noted limitations, SDS and SDS+ effectively evaluate a hitter’s swing decisions and can offer further insight into the underlying traits of a hitter. This metric will be useful during the season for evaluation purposes and in the offseason for development purposes. At Iowa, we will continue to further the creation of this metric to support our coaching staff and the development of our athletes.</p><p>SDS allows our hitters to precisely evaluate the decisions they make at the plate and can bring light to otherwise unnoticed necessary approach changes. Some hitters might benefit from being slightly more aggressive early in counts based on the results of SDS+, whereas some hitters might just have a slightly aggressive approach at the plate. Looking at swings and takes separately for each count type can offer a more accurate evaluation of these differently styled hitters. Some hitters might take the first pitch more often than others, so the penalty will just have to be accepted. For these hitters, looking at when they are behind in the count or with two strikes will be more useful. Quantifying swing decisions is difficult but finding out which hitters take advantage of good pitches to hit is a great start. Using this metric adds another dimension to our program’s approach to data-driven decision-making and we are excited to be able to implement this to assist in the development of each hitter.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d59ccb84d820" width="1" height="1" alt=""><hr><p><a href="https://medium.com/iowabaseballmanagers/quantifying-swing-decisions-sds-d59ccb84d820">Quantifying Swing Decisions: SDS</a> was originally published in <a href="https://medium.com/iowabaseballmanagers">Iowa Baseball Managers</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>