The AI set-up for Poland is important and would be worth monitoring in the playtesting.
Try this one ... watch the Germans go heavy in East Prussia, medium around Breslau, set up for another campaign like Netherlands or Belgium or France or Yugoslavia, and light in Czechoslovakia. Then set up the extra Poles in Katowice. Poland is going to fall, always. The question is, how can you cost the Germans a few points of something, and maybe gain the Allies some goodies for the future? Remember that the production phase is before conquest, and you can't transport resources through a ZoC. Putting single stacked blocking units in clear terrain around the factories I don't think gains very much...even in bad weather the powerful German units can auto-kill any single Polish unit, losing only the time required. Worrying about trying to use the Vistula to Poland's advantage only works when the Germans are heavy in East Prussia; otherwise they just bulldoze straight across the center. When the Poles are double-stacked, the chances of actually hurting the Germans go up, and in a best-case scenario they get flipped/disorganized attacking a double stack, then the bad weather sets in. Also keep in mind that the CW might want to take a Naval impulse on their first go, to move some units and re-org some transports.
With a light East Prussian set-up, the same trick could work in Danzig...maybe the Polish navy might even survive the surprise impulse that way. But a key to the set-up is whether Germany is attempting any other campaign on the first turn. One way to analyze it is to simply count the factors adjacent to Poland, which is how some German players plan the campaign to begin with, figuring on a minimum of around 65-70 combat factors to subdue the Poles, leaving the rest for elsewhere. If Germany is up around 90-100, it doesn't really matter what the Poles do.