I agree national polling should have an effect on state races, but a perceived bump is going to be larger in a state like Massachusetts than it is in West Virginia.  I think recent polling results from Arkansas and Nevada show this.  Obama got a larger bump in Arkansas than Poblano's method predicted and a smaller bump in Nevada.  Hillary was very strong in Arkansas and wasn't very strong in Nevada.  I agree with the usage of the model in Ohio for instance but not in Virginia.

(As for Sanskrit, I have no idea how to type that let alone make coherent words or sentences)