Baseball Batting Hall of Fame Predictor
With the start of May and unfortunately no baseball, I decided to try and pass the time by seeing how well an ML model could predict who is going to enter the Hall of Fame.
Important Notes:
- Criteria is offense based only (no defense or pitching yet)
- The training cut off is 1995
- The model takes into account the number of years played so someone like Mike Trout who only has 9 years of service might be penalized as 9 years would be unlikely to get you in the HoF
Predictions
| First | Last | Final Season | HoF |
|---|---|---|---|
| Jeff | Bagwell | 2005 | Y |
| Albert | Belle | 2000 | N |
| Barry | Bonds* | 2007 | N |
| Ken | Griffey | 2010 | Y |
| Vladimir | Guerrero | 2011 | Y |
| Derek | Jeter | 2014 | Y |
| Chipper | Jones | 2012 | Y |
| Edgar | Martinez | 2004 | Y |
| Manny | Ramirez* | 2011 | N |
| Alex | Rodriguez* | 2016 | N |
| Gary | Sheffield | 2009 | N |
| Sammy | Sosa* | 2007 | N |
| Frank | Thomas | 2008 | Y |
| Jim | Thome | 2012 | Y |
| Mo | Vaughn | 2003 | N |
| Todd | Helton | 2013 | N |
| David | Ortiz* | 2016 | N |
| Adrian | Beltre | 2018 | N |
| Carlos | Beltran | 2017 | N |
| Albert | Pujols | Current | N |
| Ichiro | Suzuki | 2019 | N |
| Miguel | Cabrera | Current | N |
| Juan | Soto | Current | N |
* Named in the Mitchell Report, suspected of steroid use, or confirmed steroid use
As a baseball fan growing up in the mid 2000s, not many of these names are too surprising to see here. Some of these players are “tained” accofding to the BBWAA and will not actually be inducted (maybe veteran vote eventually). David Ortiz will be the player to watch to see what comes of the other “tainted” players as Ortiz has been rumored to be on better terms with the BBWAA.
It was surprising to see that the model did not predict Mike Trout making it until I extending the data to give him a longer career when Juan Soto was predicted as making it having only played 2 years (not trying to discount how good Juan Soto is). I wonder how extending the career for some other younger players would play out in this prediction.
The Jupyter Notebook can be found on Google Colab if you would like to check the source code or edit the notebook. I might be uploading a pitchers prediction soon as well as working in defensive statistics.
