Category Archives: Baseball

Baseball Currentview Part II: Batters

Posted on by
Reply

Part I of this series examined pitcher performances and expectations.  We rejoin our Hoos after they dropped a tough series to #9 NC State over the weekend.  This post, as the title would suggest, will assess the Hoos batters beyond the usual stats of batting average and home runs.

The art of hitting boils down to two key and distinct skills: getting on base and hitting for power.  Lineups are often dictated by which players possess which skill sets.  For instance, in a lineup of 1-9, the 1 batter is often very good at reaching base, while the 4 and 5 batters often sacrifice on base skills[1] for sheer power numbers.[2]  The 2 and 3 batters often have a blend of both, with 3 usually being the best overall hitter on the team.  I wanted to determine the Hoo most talented in each of the important categories.

I used Weighted On Base Average (wOBA) to rank the Hoos skills at reaching base.  wOBA alters the basic on-base percentage stat[3] by giving more weight to a player’s ability to put his team in a position to score; granting higher values to doubles, triples, and home runs.

Of the regular starters[4], those who most often stroll the base-paths are King, Fisher, and Taylor.  Respectively, they’re 4th, 5th and 1st in the usual lineup.  The fourth and fifth spots seem a bit odd given our previous discussion.  To hold down those positions, Fisher and King need to be two of the strongest power hitters on the team …

Which brings us to Isolated Power (ISO).  ISO is simply a measurement of the number of extra bases (i.e. beyond a single) that a player reaches per at bat.[5]

King, Fisher, and Taylor are also the three regulars who display the most consistent power, which easily makes them the best hitters on the team.  At least for this Hoos team, or maybe for college baseball in general, the best hitters tend to be good at both key skills.  This differs greatly from MLB where players are often pigeonholed into specific offensive roles.

The above graphs provide a more accurate idea of the Hoos ability to get on base and hit for power.  But naturally, I couldn’t stop there.  I had to know, especially among the top three hitters, who is the single best offensive player?  How can we combine these two important metrics, arriving at one stat to rule them all?[6]

MLB already has WAR (Wins Above Replacement) to capture all elements of a player’s offensive contribution.  I decided to create my own offensive stat because the concept of “wins over the average MLB replacement player” would be skewed in translation to the college game.  In other words, WAR constants are tailored to MLB, and the various formulas often employ some obscure data not available for college players.

So I devised Hoos Relative Offensive Contribution (hROC).  First, in order to combine wOBA and ISO, I had to standardize them; otherwise their different magnitudes would have granted wOBA undue influence on the final sum.  I used the historical average of Hoos players in both stats to create a standard base.  Thanks to the magical powers of stat databases and cut and paste, this wasn’t nearly as difficult as I anticipated:

I could then calculate each current Hoo’s variation from each historical average (displayed in cwOBA and cISO below).  This put both wOBA and ISO on an even playing field.  After simple addition, voila!, we have a figure representative of both stats (cTotal).  … but this is really small and difficult to differentiate.  So I decided to multiply this total by each player’s number of plate appearances in order to weed out those[7] who have unsustained impressive stats.  However, this would unduly punish Hoos with lots of plate appearances who have a negative cTotal, making them look worse than some of the backups.[8]  To solve this problem, I moved all the cTotal values to positive by adding the minimum score to each before multiplying by PA.

This means that one player will always be the reference point of hROC.  That current “honor” goes to Nate Irving.  The rest of the players have an hROC number indicative of their relative contribution to the Hoos offense.  These magnitudes reveal both the individual differences and the drop off between the best and the rest[9]

I will probably continue to tweak this stat in through future posts.  For now, though, the chart and graph should provide an interesting reference point for the Hoos batters.  The Hoos face another important ACC weekend against Wake Forest.[10]

  1. [1] by striking out a lot, *cough* Ryan Howard *cough* ↩
  2. [2] Just for reference, I’m a huge Phillies fan.  I’ll try not to let that leak into my posts here, but sometimes, as in the previous footnote, I’ll reference Phillies players that are representative of my point. ↩
  3. [3] percentage of plate appearances where a base is reached in any way ↩
  4. [4] see the last column of plate appearances ↩
  5. [5] slugging percentage minus on base percentage ↩
  6. [6] I briefly considered calling the stat PRECIOUS and coming up with some baseball related words for each letter.  But that was a little too nerdy, even for me. …… Actually it was just too hard to come up with the right words ↩
  7. [7] like Brandon Downes in the graphs above ↩
  8. [8] since they would have a more negative number ↩
  9. [9] the drop off is best shown in the curve of the next graph ↩
  10. [10] also, go Phillies ↩

Baseball Currentview Part I: Pitchers

Posted on by
1

This extremely belated pre currentview of the baseball season, and Wahoo Metrics Baseball Opening Post(c), will begin with an examination of the Hoos pitching staff.  I’m excited about working baseball into the mix on this site since advanced statistical analysis is already entrenched in the sport.[1]  I currently plan to post series recaps and other analyses as the mood strikes.  Hopefully you’ll see more frequent posting … we’ll see how it goes.

The (17-8-1) Hoos had a rough start to the season, losing to Boston College in the opener, before riding a current six game winning streak that includes a sweep of #23 Clemson.  In this post we’ll examine: what can we expect from the pitching staff over the rest of the year, and are their current statistics reflective of their actual quality?

We’ll use FIP to predict future trends in pitcher ERA.  The acronym stands for Fielding Independent Pitching, and as the name would suggest, attempts to measure a pitcher’s performance by eliminating any runs contributed to his ERA that were the product of sub-par fielding.[2]  FIP = (13*HR + 3*BB – 2*K)/IP + 3.10.

Just some of the[3] philosophy behind FIP:  The best possible outcome for a pitcher in a given at-bat is a strikeout.   An out is recorded, and runners cannot advance as they might be able to on a ground ball or a fly ball, when they could either score a run or move into better scoring position.  A strikeout is also entirely a product of pitch quality.  Fly balls and ground balls depend on the position of the defense and hand-eye coordination of the defenders.  Failing an out, batters can reach base via a single, double, triple, homerun, or walk.  Of the ways for a batter to reach base, homeruns and walks are entirely within the pitcher’s control.  Over a number of innings, these pitcher-reliant statistics can be related to an anticipated ERA according to the previous equation. It’s simple, but remarkably effective in eliminating defensive effects.  Furthermore, since FIP tends to measure an average defense, an FIP that is significantly lower than ERA could suggest that a pitcher is “due” for an ERA improvement.[4]

In addition to calculating the current FIP for the Hoos pitchers, which you’ll see later, I wanted to come up with a method of predicting what their FIP should be for this year given the average progression of a college pitcher.  So I took six Hoos pitchers who have contributed over the last few years and plotted their FIPs vs. years of college pitching experience:

On average, the Hoos progressed,[5] which is also reflected in the image below.  The first chart shows the data points that created the above graph.  The second chart shows the change in FIP as compared to the previous year, which is averaged at the bottom.  So the average Hoos pitcher sees his FIP drop .61 between his third and fourth years, and exactly 1 whole FIP point (interestingly enough) over the course of his college career.[6]

I used these anticipated FIP changes to create the final chart below.  This chart compares the expected 2012 FIP (“eFIP”) to the current FIP and ERA of each Hoos pitcher.[7]  As previously mentioned we can anticipate future changes in ERA, via changes in defensive performance, based on a comparison to FIP.  Here eFIP acts as a comparison tool for change in FIP itself.  So with FIP, eFIP, and ERA’s powers combined, we’ll get [8] an idea of expected change in both defensive support and quality over the rest of the season.  The pitcher expectations are rated on a[9] range of three negative signs, meaning we expect their ERA to significantly drop to match their performance,[10] to three positive signs,[11] which would indicate an expected dramatic increase in ERA.

According to my analysis, Branden Kline could see a dramatic improvement in ERA.  He has received relatively poor defense (-.37 FIP vs. ERA), but most of his under-performance is due to an inflation of nearly +1 FIP over expected.  So his own pitching quality is the primary culprit.  If he can fix whatever has caused him to regress this year, and come more in line with the average progress of Hoos pitchers, he could see a huge drop in ERA over the rest of the season.

Kyle Crockett probably wonders why all the fielders suddenly stop caring when he takes the mound (-1.68 FIP vs. ERA).  His ERA at this point is largely the product of flukey defense and will certainly drop before season’s end.  Scott Silverstein, on the other hand,  could see an ugly ERA increase; I wouldn’t be surprised to see his 2.10 ERA end up between 3.00 and 3.25.  But I won’t bore you with more sentences of just names and numbers, the chart with trend predictions is there for your perusal.

The pitching staff as a whole has received poor defensive support from the Hoos fielders.  Unless the Hoos are just a generally bad defensive team, we should see an average ERA rebound buoyed by some strong individual improvements.[12]

  1. [1] See Moneyball.  If you enjoy this post at all, you should read the book, it’s great. I just (sort of) restrained myself from comparing the (also good) movie to the book.  Go back to the post, escape while you still can. ↩
  2. [2]  Caution: Formula Approaching! ↩
  3. [3] brief …. I promise ↩
  4. [4] Although making this type of claim probability-wise is very sticky, as a low FIP with bad defense does not give a higher chance of receiving good defensive support over the rest of the season.  Nevertheless, most pitchers on a given team end up with similar defensive support at the end of the season. ↩
  5. [5] as one would expect ↩
  6. [6]   This is a significant 1 point.  Since the magnitude of FIP is designed to match that of ERA, we can imagine the 1 point drop as the difference between a 4.50 ERA and a 3.50 ERA.  In MLB terms, the former would have trouble holding down a spot in the rotation, while the latter would probably be the #2 or #1 starter on his team. ↩
  7. [7] My cutoff for inclusion was 10 innings pitched. ↩
  8. [8]  Captain Planet! ↩
  9. [9] extremely scientific ↩
  10. [10] Green, because it’s good … even though it’s a negative sign. ↩
  11. [11] again, though, positive is bad ↩
  12. [12] I’ll try to figure out some way to assess my predictions at the end of the regular season ↩