I was wrong and you were both right about the end of the chord being in compression. In my haste I got stuck on simple support, rather than cantilever and never drew a diagram. I would have seen the torque about the top of the king post then. I'll be looking at the MNDOT site, but not right away.
I see that sheer facture in compression is consistent.
Eh, it happens. I've been playing some with that bridge design page and its sorta fun, though it will occasionally accept bridge designs where part of the lattice is redundantly connected and part is flexible (as opposed to everything being non-redundantly rigid). Such cases can produce really bizarre nonsensical numbers.
It's also interesting, though, to observe the amount of mechanical advantage that's available even when one only puts objects on grid squares. Obviously if one puts objects at finer boundaries one can get really huge levels of mechanical advantage, but even when confined to grid squares the values are still pretty huge.