Astros Minor League Game Score Database

One of the projects I started this year was keeping track of Minor League Game Scores. If you don't know what Game Score is check out this article I wrote earlier in the year. The too long, didn't read (TL;DR) version is that Game Score ,developed in the 1980s by Bill James, took a pitchers pitching line (IP, H, ER, R, BB, SO) and gave each number a negative or positive number to multiply and add up. If a pitcher posted a Game Score of 50 or higher it meant they gave their team a greater than 50% chance to win. Anything above 60 is considered a gem and pitchers can go above the 100 mark. Chris Devenski's no-hitter on September 1, for the Lexington Legends, was a 102 Game Score. Devenski's average Game Score with Lexington was 58, take away his no-hitter and it's a 47 Game Score.

Game Score has it's flaws, but it has it's uses and it's another simplistic way in which we can evaluate a pitcher. What I really like about it is that it gives a pitchers outing a singular number to compare to other starts and even other pitchers. On top of that it's a very do it yourself type of statistic which makes it relatively easy to put together a database. I'll be sharing a link to the database at the end of this article. But before you go rushing down to that link I want to discuss a couple things about the sheet.

It's extensive but not exhaustive. We have a tremendous amount of information in there and the initial idea was to get every pitchers outing. After a while of attempting to do this our full-time responsibilities got in the way so unfortunately we didn't get every outing, but I think we did a pretty good job. Accuracy is another small issue. I've double checked a lot of these scores, but I haven't gone through all of them. Next year we plan to only stick with pitchers of interest, this should help with accuracy.

Another note I want to make is that we didn't grab every single start. One of the failings in Game Score is that it doesn't account for short innings, so when pitch counts or injuries are involved in a start a two inning start with no runs allowed can really skew results, so we threw those out. Unless they had a two inning start and got absolutely bombed in which case we probably wanted to keep that in. We didn't run into a lot of cases with two inning starts, but they were there.

One of the things I wanted to do with Game Score was make it more appealing to the masses and so I came up with Pitcher Rating. Like Game Score I kept the math simple, but got it pretty close to the more well known Quarterback Rating. I thought this might help people more familiar with the football statistic get acquainted with Game Score. Both have their flaws, but sometimes those are the best stats to draw in fans. That said I think I've backed off that idea, and that's probably due to the increased usage in unaltered Game Score on this site this year, but it's still on the spreadsheet so that's there to look at as well. PR*1.6 is the one that's supposed to be similar to Quarterback Rating. If there's a strong desire to keep it I'll do so next year, if not it's probably gone next year.

Notable Game Scores

Game Score = GSC

Average Game Score is all a pitchers Game Scores added up and divided by the number of starts.

These are just some of the take a ways I have of the Game Score Database:

• The highest GSC that wasn't a no-hitter was 83 by Ross Seaton in Corpus Christi
• The lowest GSC belonged to Dallas Keuchel who posted a -15. Yes, apparently you can go negative as well. Take out that one bad start and Keuchel's average GSC goes from 49.79 to 54.77.
• Jarred Cosar's Oklahoma City average GSC of 49.67 is very similiar to his Corpus Christi average GSC of 49.31 despite about a run difference in his ERA.
• Wes Musick never posted an average GSC below 50 at three different levels of the Minor Leagues. He started and posted his highest average GSC at Lancaster, 59.60. That's also the highest GSC of any full-season ball affiliate. Of course that also came in only four starts.
  
• The highest average GSC by pitcher in full-season ball with at least 10 starts belonged to Nicholas Tropeano in Lexington. He posted an average GSC of 57 in 13 Lexington starts.
• Michael Fotlynewicz was the second highest with a 55.12 average GSC in 25 Lexington starts.
• Jose Cisnero had the highest average GSC with at least 10 starts between Oklahoma City and Corpus Christi, with a 54.52 average GSC.
• Asher Wojciechowski made a pretty good impression with an average GSC of 57.14 in seven Corpus Christi starts.

What else do you see? TCB Game Score Database