The latest NAM model is showing an impressive front end snow for DCA into the PHL area.  However, it dies down a lot as it comes into the Philly area with accumulations of 1-2".  The NAM model puts a band of 2-3" over the DCA area!  Other models show very little if anything with this front end.  

1) The models showing nothing are speeding up the convection across the southeast.  This results in a faster moving boundary that cuts off the moisture supply to the front end/enhanced convergence.

2) The NAM/WRF are great models for convection and with it being slower with the boundary it allows an enhancement of moisture advection and convergence for the front end.  One possible reason for this is the gulf stream convection helping to back winds aloft slowing down the boundary.

3) The hole in precip...you will see this band that approaches the area and splits into two different sections.  One will go towards the main low in the Upper Midwest and the second will hit DCA on northward.  Why? Well one is the best dynamics hopefully setting up as NAM portrays. Second:

The strong gradient in the wind over this area is producing convergence above the the level of non-divergence.  AKA the wind is forced down and is a sinking motion.  Notice how more jet diffluence and lesser gradient exist over DCA and Northeast.  

Now this forecast is confusing for so many reasons.  But as I showed above we need to focus on the convection out west.  This will give us an idea of what to expect in terms of precipitation Friday into Saturday morning.  The second thing is the temperatures after the front end (front end is snow).  The third thing is the coastal low.

Given the isentropic lifting showing up on the NAM I am concerned with this model not being correct on keeping coastal low precip too far east.  I think we will see the model come more west with the coastal precip and allow for a good deal of rain for Philly but more snow in portions of E PA into NW NJ/S NY/New England.

Conclusion:

1) Watching front end snow.  NAM is the most impressive and could be on to something with a 2-4" surprise before the wintry mix/rain for many.

2) Focusing on convection: If it moves fast like the GFS do not expect much front end precip if any at all.  If it is slow like the NAM we will see some snow maps!

3) For New England NE PA/NW NJ-  I think the models will converge on a much snowier solution.  Boston looking good for another nice snowfall (6-12")

4) Depending on how strong the coastal gets will determine how far south the snow goes, right now it is going to take a lot to extend the back end snows to PHL.  NYC is still in the game (10-20%) for a back end snow with minor accumulations.