The problems may be fixed now. Please try again.
As for the option, no, it does not work this way.
The algorithm maintains the "real" mouse position (where the mouse pointer would be if the patch wasn't installed). Now the actual mouse pointer is always accelerated towards this "real" position. Acceleration value is the distance divided by the "divisor" parameter.
This way it works as if the mouse pointer was connected to the "real" (invisible) position by a rubber band. The only changeable parameter is the divisor.
Well, we are on a programmer's forum here. You could ask for the source code and adjust it yourself. I really don't wish to put so much brain work into this.
Bye,
Thomas