I suspect that airfoil behavior at full-scale Reynolds numbers is a lot different than at ours. Gary James or Frank Williams or Igor could elaborate as to why. My airfoil notions are based mostly on experience with models at Reynolds numbers between about 400,000 and 700,000, particularly single-element (no flap) airfoils used on combat planes. It seems to me for this application that airfoils with the highest max lift have the most abrupt stall. For several years I used an NACA 0016.5 with the max thickness squished forward to 25% chord. It worked great, but I had to use bellcrank stops, and the slightest warp made an airplane useless. Any anomaly around 25% chord significantly killed off lift capability, but I don't remember it causing an abrupt stall. When I used a foam wing with a spar at 25%, I carefully planed the spar, then filled any remaining little dips with spackle. Likewise, a pointy leading edge caused reduced max lift, but didn't cause an abrupt stall. If anything, a pointy LE made the stall gentler.
I fly an Impact for stunt. I plotted its airfoil using the NACA four-digit formula, fiddling with the parameters until I matched the root and tip airfoils. As Brent observed, the trailing edge thickness and slope are the same from root to tip. The Impact airfoil is very nearly a straight line for the last 20% or so before the flap, causing it to be hard to sand.
My last two Impacts have a wing structure with slanty ribs that make for pretty bad bumps in the wing cross section aft of the fattest part of the wing. Although the airplane has significant, audible separation, it doesn't do anything perverted. It's a pretty good stunt plane. Vortex generators may have ameliorated the effect of the bumps.
The aft part of Igor's airfoil has a lot of curvature and fairs nicely into the flap. It wins a lot of stunt contests. The only comparison I've done between it and the Impact is with XFOIL via Profili, which favored the Impact airfoil.
Aerodynamically I would reckon--based on no theory--that a ice-cream-cone-looking airfoil at our Reynolds numbers could be OK. I'd use a blunt leading edge and keep the curvature low at the transition to the straight part. Looking at the Endgame II tip template above causes me to reach for a sanding block--just a wee swipe in front of the straight part. Structurally, I'd think that a curve in the wing surface would be better than a straight line.
(edited to remove some stupidity)