I have an equation: Y = c + b1X1 + b2X2 + b3X3

Y = Cryptographic ability (scaled 0-200)
X1 = Pattern recognition skills (scaled 0-20)
X2 = Mathematical ability (scaled 0-20)
X3 = Logical reasoning skills (scaled 0-20)

It may be assumed there is a linear relationship between Y and X variables.

I have data for 200 variables which are collectively known to describe 100% of Y (this is a simulation).

What is the max # of X variables I can include and how would it effect the robustness of results and individual significance if I jump from 3 explanatory variables to 5 or 7?

Q2 (priority question):
Which is the best model for me to use in Oxmetrics?

Module = ?
Category = ?
Model Class = ?

Any further considerations in the subsequent specification windows?

Thanks for taking the time to read and thanks in advance for any replies