But GMs turn over and memory flags, and since I do a report for the GBC, that makes it even more obtuse. Just what, after all, is “Plays Above Average” and how do I calculate it?
I’ll document it here with the idea that I’ll simply link to the description in the annual threads.
So, let’s get going, shall we?
How the Data Was Collected:
This data comes from a script I run against every game log that comes from the game itself. To do that, I open every box score (to get starting defenders, among other things), and every log (to get plate appearance results) over the course of the season, and print a little report to make sure the numbers add up to however many games the total should be at the time I run the report. If everyone’s number of games is right, I assume I’m not missing any.
Plays Above Average:
After reading in all play-by-play data, I gather granular data by every plate appearance. This includes whether the plate appearance resulted in a batted ball ( of type ground ball, fly ball, line drive, or pop-up), the zones those balls were hit in (which you can see in the game log), and the final result of the plate appearance.
Since we don’t have the run scoring environment matrix for the BBA, and since I don’t have a way to fully quantify the out/runners/score situations each hit/out was made in, I create the parameter I call Plays Above Average (PAA) for each type of batted ball (ground ball, fly ball, line drive) that is assigned to each zone.
After counting every ball in play (BIP) of every type in every zone, and every error and out made on those BIP, I then calculate the PAA of each team in each zone (as well as the PAA of the entire league as a whole) in the following way:
PAA = (Team Out%)/(League Average Out%)*(Team Batted Ball Type)
For example: in 2064, Bikini pitchers allowed 253 ground balls right at the shortstop (Zone 6). Bikini shortstops converted 243 of those into outs, good for a 96.047% rate. That was just a touch better than the league’s average of 93.996%. To determine Bikini shortstop’s PAA for zone 6, we simply take the number of outs that Bikini shortstops made (293) and subtract the number that a league average shortstop could have been expected to make (253 * .93996 = 237.8099).
So, for Zone 6: Bikini’s PAA was 243-237.8066 = +5.19
Note—at this level, I can’t differentiate Keith Williams from Agatone Louganis. It’s just one big pile of shortstops.
The Final Calculation: Once I have this data for every zone, I then do a simple addition of all performances by all zones. This gives me the final PAA number for that kind of BIP.
Ultimately, for ground balls, it looks like this:
| PLAYS ABOVE AVG | 3L | 3 | 34 | 4 | 4M | 6M | 6 | 56 | 5 | 5L |
|---|---|---|---|---|---|---|---|---|---|---|
| SHALLOW | 0.000 | 1.760 | 1.442 | 1.042 | -0.049 | 1.420 | 1.191 | 2.351 | -4.367 | 0.000 |
| MID | 1.700 | -2.177 | 1.122 | -8.550 | -4.075 | 1.809 | 5.191 | -2.944 | -0.349 | -1.597 |
| DEEP | 0.000 | 1.628 | 2.728 | 0.259 | 1.212 | 7.971 | 0.124 | 0.265 | -0.639 | 0.000 |
In other words, the Bikini infield as a whole made 8.47 more plays (converted BIP into outs) than the average BBA would have. So, despite second base being more than a bit ugly, the Bikini infield was quite good last year. The outfield, however…yeesh.
What’s In, What’s Out:
Easy enough, right? But there’s still a question or two to answer. What do I miss?
Pitcher/Catcher: Note that for this analysis, I’ve ignored pop-ups and I’ve ignored balls hit to the pitcher/catcher. Mostly this is because I’m trying to look at classic infield and outfield metrics, but I also note that I’m not certain how catcher defense is really handled, and I didn’t want to murky the water with pitchers.
Uncertain Results: Or maybe I should say plays that don’t get reported properly on the game logs. There are very few of these, but they exist. Things like some rundowns (they get noted in the game log, but OOTP does not include anything to note the chain of the putout. Ultimately, they are probably calculated correctly here, but I’m not certain.
Bottom line: there’s a little noise in there.
Double Plays: “Outs" do not count double plays as two outs. All we’re doing here is counting the number of balls a fielder converted to an out vs. a hit or error.
Infield pop-ups: I chose not to include infield pop-ups here.
Balls Hit off the Top of Walls: I learned something new recently—that being the “S” zones collect data on fly balls and line drive that hit the very top of the wall and ricochet back into the field of play. Maybe you all already knew that, but it was news to me. That’s cool, but my script will miss them for now. It’s maybe 10-15 plays a season, which will make a difference, but not a particularly big one.
