Take SOrate, BBrate, HBPrate, MPrate for all pitchers
SOrate BBrate HBrate MPrate
19.69 5.41 0.67 1.73
13.46 7.65 0.39 1.25
11.37 9.51 0.93 2.54
... ... ... ...
26.03 10.14 0.09 1.84
14.05 7.36 0.78 1.16
9.53 8.43 0.67 2.27
Find the (linear) equation that best matches the data
C1 * SOrate + C2 * BBrate + C3 * HBPrate + K = avg MPrate
Using linear regression, find a formula
Used all three-year spans with 300+ BF (i.e. 2000-2, 2001-3, etc. for K.Wood)
Divided knuckleballers (206) and non-knuckleballers (14,828)