Wednesday, May 18, 2016

Analytic Support for Why BYU Deserved its 1984 National Championship

Monday, May 2, 2016

Wei-Chung Wang doesn't recognize FIP (Fielding Independent Pitching), or does he?

A little dated, but in this interview with Kyle Lesniewski of Reviewing the Brew, he fielded this question about awareness of advanced pitching statistics:

The advanced stats say you may be suffering from some bad luck this season, given your high BABIP rate (.357 despite low 32.8% ground ball and 13.6% line drive rates) and low FIP. Do you look at the advanced stats at all, or do you focus more on the traditional things like wins and ERA?
This was Wang's reply: 

Wins and losses mean little at this level of the minors, so I don’t really pay attention to my win-loss record and I don’t look at my ERA. I keep my focus on the things I can specifically control during the games I pitch, especially walks and home runs. 
So though he did not assent to knowledge FIP, he appears to know that he could only control things like BB and HR, which along with SO, comprises FIP.

Friday, January 15, 2016

Wei-yin Chen and the Art of Changing Speeds

According to this Fangraphs article, Chen is among the top twenty pitchers in changing speeds, averaging a 18.2 difference between his fastball (92 mph) and his curve (73 mph).

Wei-yin Chen and Exit Speed Suppression has an article on exit speed, the speed of the ball as it comes off the bat. Apparently, Chen leads the MLB in suppressing said speed.

Tuesday, June 3, 2014

Great Sabermetric and Pitch f/x Writeup on Wei-yin Chen

Great sabermetric and pitch f/x writeup on Wei-yin Chen's improvement over prior years, with a higher groundball rate, an only slightly lower K rate, and very few walks. The analysis is much better than this goofy piece from Chen's rookie season. The latter is basically saying that Chen got batters to whiff on strikes high up in the strikezone, pitches that usually go for homeruns. Donut3 in the comments section sarcastically remarked, "Extreme flyball pitcher pitching in Camden with a tiny HR/FB percentage. This will definitely last." Indeed, Chen's HR rate soared.

Thursday, January 30, 2014

How Competitive is Taiwanese Professional Baseball? A Look at CPBL Hitting.

Most fans of professional baseball in Taiwan (CPBL) will tell you that its competitiveness stands at around AA level. Some outsiders assess it as A+. Clay Davenport of Baseball Prospectus, who has rated the Japanese baseball as higher than AAA, and the Korean league as AA, guessed that the CPBL was about A+ after a cursory glance.

To answer this question, I use Clay Davenport's methodology of comparing performances of players who played across different leagues. His work in league comparisons has been the most extensive, being the only sabermetrician to assess foreign leagues. Looking at foreign position players who played in the CPBL from 2005-13, as well has Taiwanese position players who have played in the Americas, with at least 90 common plate appearances in each league, I compare their equivalent averages (EqAs) both playing in Taiwan and in the Americas (that includes the MLB, MiLB, and Dominican, Venezuelan, Puerto Rican, and Mexican winter leagues). I divide their Davenport-Translated EqAs in the other leagues by their EqAs in the CPBL. Then I take the weighted average of that percentage to determine the CPBL's competitiveness level in batting. I find that the CPBL has a rating of 0.74, which is below AA level. I also find that Taiwanese players who return improve their performances by a much wider margin than foreigners who play in Taiwan. If the CPBL were a major league, foreigners increase their performance by 33% on average. Taiwanese players increase their performance by 48%.

This is how Taiwan stacks against other leagues:
League Rating
MLB 1.000
Nippon Professional Baseball 0.92
International League (AAA) 0.89
Pacific Coast League (AAA) 0.85
Eastern League (AA) 0.82
Southern League (AA) 0.82
Texas League (AA) 0.80
Mexican League (AAA) 0.78
Chinese Professional Baseball League [Hitting Only] 0.74

In the future, I will redo the assessment for pitchers, and include the 2014 season, as well as use more park-centric park factors.

Methodology and Sources
I took all position players who played in the CPBL from the 2005-13 seasons, computed their EqAs in the CPBL. Numbers to calculate EqA were culled from the CPBL website, which had individual and league statistics for each year in existence. The EqA formula is available from Baseball Prospectus.

If you look at the formula, you will see that the EqA formula includes park factors (PFs). CPBL "fake" PFs for the years 2005-10 were calculated by Taiwanese sabermetrician Madboy. As explained in a previous post, the tongue-in-cheek name stems from the lack of a fixed stadium for each CPBL team. Since Madboy's "cooked" PFs for 2011 were a two-year average, I should augment it with the 2012 raw PFs. Using the same methodology from that Madboy used, I calculated the 2012 raw figure then computed the 2011 "cooked" figure with the following formula F2011=(2*F2011+RAW2012)/3. I also calculated 2011 independently. My results differed from Madboy, so for consistency I used my own to compute 2012's "cooked" figure.

For each player, I compared their EqA in the CPBL to their Davenport-Translated EqA in the major, minor, or Winter leagues, available and The comparative methodology I am about to describe is basically the same Clay Davenport used to assess the Japanese league. The only difference is that I use 90 minimal common plate appearances (PAs) as opposed to 100. The EqA have already been translated into their major-league equivalents. If a player played in a league that has Davenport-translated states before and after playing in Taiwan, I count them as two observations. Take Manny Ramirez for example. I count both the majors/minors to CPBL and the CPBL-Round Rock (AA). For the American/Japanese/Korean leagues, I use one to three seasons before or after the CPBL year, whatever is necessary to get my 200 PAs. Then I divide the Davenport-Translated EqA by the CPBL EqA and take the weighted average off each observation. had translated EqAs for every player and every year needed, except players who played in the Mexican League pre-2005. That Clay Davenport did not have them is understandable, as the MiLB website has Mexican League statistics only back to 2005. League statistics are required to normalize a player's EqA relative to the league's level. And the universe contained two players that played in the Mexican League in 2004 before playing in the CPBL in 2005: Mario Encarnacion and Wilton Veras. Fortunately, provided the league totals for 2004. I calculated their EqAs, using park factors for their teams available from the Baseball Prospectus 2005 Annual, and then applied Clay Davenport's multiplier for the league of 0.780.

The multipliers for the Mexican and Japanese leagues were obtained from the sources mentioned above. The minor league multipliers were obtained here.