The 2017 Major League Baseball season kicks off on Sunday and we’ll tackle it with daily articles, much like we’re doing with the NBA totals. While we have four different NBA methods for totals, and then run those through a Points Per Possession routine if they all agree, we’ll use one basic math-based system for baseball. It’s one that I wrote a few years back and is based on one Jim Jasper created and featured in his book, “Sports Betting: A Computer Expert’s Winning Secrets For Baseball and Football.”
The premise of the system is rather simple. You take each team’s scoring average and compare it to their league average to come up with a percentage. If the average number of runs scored by National League teams is 4.2 and the Pirates are scoring 4.0 runs per game, you have 4.0 divided by 4.2, which equals .952. If the Cubs are averaging 4.5 runs per game, you have 4.5 divided by 4.2, which equals 1.071. Those will be the offense numbers.
Defense, obviously, is going to change depending on who the pitcher is. ERA has proven to be a somewhat questionable statistic, so we’ll simply use the average number of runs allowed each time a particular pitcher takes the mound. Using the Mets’ Seth Lugo as an example, he had an ERA of 2.67, but he allowed an average of 4.25 runs per start. That includes the bullpen and unearned runs. Using our numbers above, we’d expect the Pirates to score 4.05 runs against the Mets with Lugo starting, which is .952 times 4.25 equals 4.05. The Cubs would be expected to score 4.55 runs, which makes sense, as they are averaging a half-run per game more than the Pirates.
Then the run difference is turned into a moneyline price by dividing the difference by .5 and then making an addition for home field.
One surprising aspect was that the American League is really no longer a much higher-scoring league than the National League. In 2016, the average number of runs scored by NL teams was 4.43, while the AL wasn’t much more at 4.51 runs per game. In 2013 for example, AL teams scored 4.32 runs per game compared to 4.0 for NL teams.
We’ll begin with last year’s stats, which are always a bit tricky, and work current stats into the equation and also account for the scoring difference teams have with right-handed starters compared to left-handed starters. Once we have some games under our belt, we can tinker around with park adjustments for totals and other things.
It should definitely be fun.