Internally what the physics engine does is it solves dozens of equations. Due to machine precision some of these equations are ill-formed, e.g. when you're trying to solve
x-1=0 and x-1.1=0
at the same time you won't be able to come to a satisfying solution which visually means that the physics object will do something weird. The whole point of setting correction factors and adding the +/-2 values above is to allow for a fudge factor so that a satisfying solution can be found. Essentially you are now telling the PH_WHEEL that they should point roughly in the right direction as opposed to forcing them to be 100% aligned which leads to instability.