So I was thinking about how dumb it is that you often get shitty wallbugs when you least want them, particularly for verts etc. specially after how it happened on the last jump of free with straight walls. Myself like many others probably started with afterglows tutorial where he recommends the func_brush to get non-stick walls. In short func_brush in most cases seems to be absolute turd, causing wallbugs even on orthogonal walls on the grid.
So I tested some alternatives that people may or may not use. Func_clip_vphys simply placed 1 unit thick off the walls, is non-solid to players and doesn't stop rockets; physbox_multiplayer, set up described here:
http://jump.tf/forum/index.php/topic,1517.0.html; and finally against walls entirely made up of func_brush and 1 unit nodraw func_brush off the wall.
I made a quick test map here:
https://www.dropbox.com/s/ti4xqnljq7mjzaq/teflontest.bsp?dl=0.
In short the 1 unit thick func_brush felt the worst with bugs, even though overall its fairly clean. I didn't get any easy wallbugs on the walls made entirely of func_brush, but in general I think this is bad practice to do, as you can easily get leaks if any of these poke into the void.
The physbox and clip_vphys both didn't produce an errors for me and in general seem like better alternatives, with clip_vphys probably the best overall as it wont affect soldier wallpogo and is much quicker to set up than the phys_box; just drop and drag a wallcopy, make it 1 unit thick and tie to the entity and you are done.
IN SHORT:Unless you actually want to produce wallbugs stop using func_brushes, otherwise I see no reason why using clip_vphys isn't intrinsically better in every case.