Okay, this has been really eating me up... #11 is clearly a compression member in the configuration at failure. So, why was it not detensioned immediately after/during placement? Or if it was, why was it (presumably) being tensioned after the fact?
This is the single biggest mystery to me.
