The mid-summer classic is here, and I thought it would be fun to break out my series projection model to handicap each leagues odds.
I’ve had to make some minor adjustments to the way that I normally do things because the All-Star game is pretty different from a regular season game.
Here is a quick description of how the model works and what changes I made to take into account the differences from the regular season for the All-Star game:
- I started by using the Lineup Analysis tool from Baseball Musings to get the run scored projections for each team based on their starting lineups. Obviously these guys don’t play together everyday, and this should give a rough estimate of what to expect. I wasn’t sure how to account for the reserves, so just know that I made no adjustment and that could be a factor in accuracy of the projection.
- For the pitching, I assumed that both starting pitchers will go two innings, and then it will go to the bullpen. Again, because I can not forecast with any real accuracy who will pitch and how many innings, I took the total season stat line for each pitcher and created a bullpen FIP for each league to complete the game.
- The data used in the projection model is based on the current season’s statistics to date.
- The estimated run differential is then converted into a projected winning percentage using the pythagorean expectation.
- Then, it’s converted into an odds of winning the game using the log5 method developed by Bill James
- I have removed the home field advantage factor because with just one Royal represented, I am not comfortable saying that any one team would have a measurable advantage at what is essentially a neutral ball park
Here is how the model breaks everything breaks down when each is up against an “average” team:
| Runs Scored | |
| American | 6.353 |
| National | 6.530 |
The National league has a slight advantage in projected runs scored. Most of that comes from the advantage that they get from their first five batters all having an OBP above .380 compared to just two of the first five for the AL team. Overall, the NL has the better starting player at six of the nine spots when you look at wRC+. Both are very good offensive teams but with a slight edge to the NL.
| Runs Allowed | ||||
| Starter | Bullpen | Defense | Total | |
| American | 0.653 | 2.450 | 0.326 | 3.429 |
| National | 0.693 | 2.318 | 0.054 | 3.065 |
In the battle of the starting pitchers the AL has the slight advantage. Justin Verlander edges Matt Cain in expected runs given up in the two innings I expected each to go. When the game goes to the bullpen, the edge moves to the NL. It will be interesting to see how each manager utilizes the reserve pitchers.
The NL has a pretty big edge in the defense of the starters with a .25 run advantage, but this is something that should really be taken with a grain of salt. The defensive metrics have a high level of uncertainty, and it also doesn’t take into account that these players don’t play together regularly, or what the reserves might be able to contribute defensively. I included them here for illustrative purposes, but for the odds in the game, I have removed them.
| Winning Percentage | |
| American | .753 |
| National | .797 |
Overall, the NL is the slightly better team. If they played a full 162 game schedule, they would win nearly 130 games. For comparison, the AL would be expected to be around a 122 win team.
When all is said and done the NL is the favorite, they should be expected to win around 53 percent of the time.
