The next step is to choose the best settings for factors A, B, C, and D to minimize both s ̂ and y ̂. Starting with s ̂, the best setting for A is +1 coded (10 actual). This setting for A will result in a predicted
s ̂ = .4348 - .198(1) = .2368.
Now to minimize y ̂, with A already set to +1 to minimize s ̂, we have
y ̂ = 4.8875 - 1.0938(1) + .4125C + .2188(1)C = 3.7937 +.6313C
Which is minimized when C = -1 coded (170 actual). Thus, when C = -1
y ̂ = 3.7937 -.6313(-1) = 3.1624