This was a really fun exercise on Detroit Tigers Run Values. This gist of this post: who has seen discrepancies between their run values based on bad calls? It’s one that came to me after the discussion about a month ago on the podcast about the gradual movement to automatic balls and strikes instead of umpires.
What is Run Value?
If you are familiar with run values or run expectancy charts, please feel free to skip this section. If you aren’t familiar, he’s a quick rundown of what this is.
Run values show how good or bad a pitcher or hitter is when facing a particular pitch. It uses a chart very similar to RE24, but instead of just outs and on-base situations, it also includes the count. So it goes from RE24 to RE288. Each time a pitch is thrown, the chances of scoring runs changes. And with that, it can be calculated out to give a value that a pitcher gets for each pitch thrown and again for hitters with each pitch seen. On top of that, these values can be even more granular by including the specific pitch that was thrown.
For example. A pitcher and batter are in a 2-1 count with no one on and one out. That specific situation has a run expectancy of 0.30. So that’s saying that given all the 2-1 counts, no one on, and one out, there is an average of 0.30 runs scored the rest of that half inning. Now the pitcher throws a fastball and it generates a swinging strike. Now we have a 2-2 count, still no one on, still one out. The run expectancy is now at 0.25
For run values, we can say that both have earned a -0.05 run value by subtracting the ending RE from the initial RE for fastballs.
It’s important to note that negative values here are good for pitchers (they prevented these “runs”) and bad for hitters (they have lowered the expect runs to be scored).
The Purpose
I was very curious about what run values would look like if we saw an automatic balls and strikes put in place. To obtain the required data, I documented pitches in the strike zone that were called balls and those outside the strike zone that were called strikes without swinging. There is data for both sets of data for Tigers hitters and Tigers pitchers. I wanted to get the actual run value based on the call and then again the run value if that pitch was called correctly and calculated the difference. And sometimes there were real runs scored that shouldn’t have been as well as runs that a should have scored but didn’t for these particular pitches, so I had to include them in as well!
The Data
I used the baseball savant search to get the datasets that I mentioned above.
Four separate searches to ensure I wouldn’t get a mixed data, I then combined them into one spreadsheet, did the calculations, combined the results together.
The RE288 was a little more difficult. I was able to find and used this matrix here (http://tangotiger.com/images/uploads/re288_recursive.png). So as a note or warning: these values could be slightly different for the 2023 season only. But for this, I was fine using this matrix.
Another note I’d like to make here — I did not take into account any sort of catcher framing with this, and that is likely to take part in some here. This was an exercise where I said if there was a robo umpire and none of framing or human error were to be considered, what would the call and results look like.
The Results
So without further ado, here’s the combined total run values for hitters:
| Player | # Strikes outside | Strikes outside RV | # Balls inside | Balls inside RV | Combined | Runs | Missed Runs |
| Baddoo, Akil | 9 | -1.1 | 19 | 3.9 | 2.8 | 0 | 0 |
| Báez, Javier | 10 | -1.1 | 14 | 1.4 | 0.3 | 0 | 0 |
| Cabrera, Miguel | 8 | -1.4 | 9 | 1.8 | 0.4 | 1 | 2 |
| Carpenter, Kerry | 7 | -0.7 | 6 | 0.5 | -0.2 | 0 | 0 |
| Greene, Riley | 28 | -3.7 | 23 | 3.7 | 0.0 | 0 | 0 |
| Haase, Eric | 19 | -3.1 | 14 | 2.5 | -0.6 | 0 | 0 |
| Ibáñez, Andy | 14 | -1.7 | 6 | 0.6 | -1.1 | 0 | 0 |
| Kreidler, Ryan | 2 | -0.3 | 1 | 0.1 | -0.1 | 0 | 0 |
| Marisnick, Jake | 3 | -0.5 | 3 | 0.3 | -0.2 | 0 | 0 |
| Maton, Nick | 17 | -2.0 | 18 | 2.9 | 0.9 | 0 | 0 |
| McKinstry, Zach | 25 | -3.3 | 16 | 2.7 | -0.5 | 0 | 1 |
| Meadows, Austin | 1 | -0.1 | 0 | 0.0 | -0.1 | 0 | 0 |
| Nevin, Tyler | 7 | -1.1 | 3 | 0.3 | -0.8 | 0 | 0 |
| Rogers, Jake | 9 | -1.0 | 10 | 1.3 | 0.3 | 0 | 0 |
| Schoop, Jonathan | 5 | -0.9 | 7 | 0.9 | 0.0 | 0 | 0 |
| Short, Zack | 7 | -0.6 | 7 | 1.6 | 1.0 | 0 | 0 |
| Torkelson, Spencer | 29 | -3.4 | 23 | 2.5 | -0.9 | 0 | 0 |
| Vierling, Matt | 22 | -3.8 | 14 | 1.9 | -1.9 | 0 | 0 |
And for pitchers:
| Player | # Strikes outside | Strikes outside RV | # Balls inside | Balls inside RV | Combined | Runs | Missed Runs |
| Alexander, Tyler | 16 | -1.7 | 8 | 1.3 | -0.5 | 0 | 0 |
| Boyd, Matthew | 22 | -2.3 | 27 | 3.8 | 1.4 | 0 | 0 |
| Bristo, Braden | 2 | -0.2 | 1 | 0.1 | -0.2 | 0 | 0 |
| Cisnero, José | 10 | -1.7 | 5 | 1.0 | -0.7 | 0 | 0 |
| Englert, Mason | 12 | -2.1 | 12 | 1.2 | -0.9 | 0 | 0 |
| Faedo, Alex | 3 | -0.3 | 5 | 0.4 | 0.2 | 0 | 0 |
| Foley, Jason | 4 | -1.2 | 6 | 1.2 | 0.0 | 0 | 0 |
| Hill, Garrett | 8 | -1.5 | 5 | 0.5 | -1.0 | 0 | 0 |
| Holton, Tyler | 9 | -1.2 | 9 | 1.0 | -0.2 | 0 | 0 |
| Lange, Alex | 15 | -1.5 | 6 | 1.5 | -0.0 | 0 | 0 |
| Lorenzen, Michael | 21 | -2.2 | 14 | 1.6 | -0.6 | 0 | 0 |
| Manning, Matt | 3 | -0.2 | 5 | 0.4 | 0.2 | 0 | 0 |
| Olson, Reese | 4 | -0.3 | 5 | 0.7 | 0.3 | 0 | 0 |
| Rodriguez, Eduardo | 30 | -3.6 | 12 | 2.0 | -1.6 | 0 | 0 |
| Shreve, Chasen | 11 | -1.0 | 6 | 0.7 | -0.3 | 0 | 0 |
| Turnbull, Spencer | 9 | -1.6 | 17 | 2.6 | 1.1 | 0 | 0 |
| Vest, Will | 12 | -2.0 | 5 | 0.5 | -1.5 | 0 | 0 |
| Wentz, Joey | 26 | -3.8 | 22 | 5.8 | 2.0 | 0 | 0 |
| White, Brendan | 1 | -0.1 | 0 | 0.0 | -0.1 | 0 | 0 |
| Wingenter, Trey | 0 | 0.0 | 1 | 0.1 | 0.1 | 0 | 0 |
I mentioned before that this could be broken down by pitch — which I have also done. However, that table would end up being over 400 rows long… so instead, I’ve got a heatplot for you to check out. Red means they’ve received more run value than they should, and blue the opposite.
Conclusion
I find it interesting to see this and make me wonder how much it could change overall numbers in the end. I would assume that a majority of these have no real barring on the game overall, possibly not even the plate appearance. There are some definitely cases where a run should have been scored and was not with bases loaded walks. And in one case, it was to end the inning.
It does also make me wonder if this could be a strategy going forward. I think my next task would be to see how the goes for players year-by-year, how consistent is it for players. And is it consistent more so for hitters or pitchers. Consistency could provide strategy for targeting players leading up to the implementation of the automatic balls and strikes.
Please feel free to reach out to me on Twitter (@JerryMackinem) to discuss more or see what other players look like! I will likely be keeping this data up to date and trying other years to see how it goes!
Emily Waldon 