[Statgraphics] - Effects of a parameter and Optimization of Response do not correlate

[Luis_Indie]

I perform an analysis on DOE and as it can be seen on the attached pictures, the effect of Pressure on my experiment is supposed to be negative (less oil obtained the higher the pressure)

However, when checking the optimization given by the program to maximize the Oil, the value given for the pressure is almost 450 bar, which is the maximum possible in my experiment (see other picture attached)

How is this possible? Higher pressure is negative for the interests of my experiment but in order to obtain the best results I have to set it almost to the maximum level?

Does anyone in here have an explanation for that?