Thanks to all. I got it solved because of this forum. It was a friction issue in the x-drive, like gr5's problem but with a different reason. My x rod drive pulley was mounted 1/2 mm to far down the shaft. Since the pulley rides on the race of the rod bearing, this allowed more of the rod to protrude from the bearing and rub on the end plate. Moving the pulley slightly took care of it.
Thinking the slant over, it's surprising how infrequent I was missing steps. An 11 degree angle is 0.2 radians meaning I was missing a single additional step in the +x direction every 5 layers to make the pattern skew to the negative x.