Last week the announcers during a Phillies game mentioned that a couple teams that had poor records had also struck out alot. One of them said "there's got to be a correlation."
The correlation between runs per game and strikeouts per game was -0.14 for all 150 team/seasons from 2021-25. So it was negative but not very strong.
Then I ran a regression with runs per game as the dependent variable and OBP, SLG, and strikeouts per game as the independent variables. Here was the equation
R/G = .22*SO + 10.57*SLG + 15.41*OBP - 4.86
R-squared = .901 (90% of the variation across teams in runs scored per game is explained by the model)
Standard error = .147 (about 24 runs per season)
T-values & p-values
SO = 1.18 & 0.24
SLG = 12.24 & < 0.0000001
OBP = 8.29 & < 0.0000001
Strikeouts are not significant. SLG & OBP are significant. There is an extremely low chance that we could have gotten the coefficient values in the equation if the true value is zero. But the result for SO would be likely.
Strikeouts are also positive. Runs increase with strikeouts, but not much. 1 more strikeout per game increases runs by 0.22 per game.
Here is the regression equation without strikeouts. The coefficient values are about the same, so adding in strikeouts does not change things much.
R/G = 10.80*SLG + 14.67*OBP - 4.53
R-squared = .900
Standard error = .147
