I wonder if the pull test was done directionally to replicate a fall? Or even if it possible to do that?
Direct pull-out test I believe, so not representative of slightly eccentric shear from a fall. In the same ballpark of load though, given that it was only 150kg to failure and the peg didn't pull-out, it failed in the thin bit of metal around the eye-hole. Can't zoom in enough to see the fracture surfaces, but the description was :
"The loop where you would clip was fully corroded when I looked at cross sections and riddled with micro fractures throughout. (Likely held a few falls)"
Contrast this with another description from a very experienced climber on the same thread :
"The peg was chrome moly not mild steel. It wasn’t rusted through and was in pretty good condition when I checked it. I tried to replace but it wouldn’t move and appeared solid throughout"
Clearly appearances can be deceptive.
In answer you your question about testing direction, I guess you could test it by loading it up from below with test weights or in tension against another set of bolts lower down. Much simpler to test in pull-out with a high factor of safety.
(Photo below taken by Stuart Holmes published on facebook, with the new bolt and the peg has now been fully cut off)