Sorry, I thought it would be obvious, both circumstances are identical in terms of force and distance - apart from the addition of the spacer (red bit) on the axle. The crank stop (alloy sleeve, grey on my image) on the axle is 15mm long. Adding a 2mm spacer (red bit) means the force is applied at a point which is 2mm further away from the bearing, right? Therefore a higher bending moment on the axle, as the force (which remains identical for the sake of argument) is being applied further from the pivot (bearing). Increasing the width of the pedal is pretty much irrelevant as this will only increase the (vertical/horizontal, whichever you want to use) component of the force a tiny amount? The maths could be done but nobody knows the exact foot position Dave uses so... Edit: Had yet another look at the BB, looks like it's broken at the end of the crank bolt, which ends at almost exactly the same distance from the crank stop as the thickness of the spacer used. There is of course a stress riser from the bolt anyway (solid to hollow section change), but combined with the crank being in the 'wrong place' and not covering over this stress riser... you get the idea. Looks like more of a torsional failure, doubt it would have happened if the crank was up to the stop as designed... Hence Reset not replacing it.