Sorry black'n'yellow I am really struggling to understand why you continue to insist and believe your "hanger/rear mech should fail" for chain too short thesis must be correct, from originally:
If the loads were that enormous, and if the chain was so tight that it caused the frame to fail, then the gear hanger or mech would have failed long before the chain stay or BB gave out. Hangers are specifically designed to be one of the first failure points in order to prevent stuff like this from happening.
and after Mickle said bent hangers is not a necessary outcome, to:
Well, I'm not 'guessing' either and how do you know, exactly..? And why is my own experience of replacing bent hangers less relevant than yours? Clearly you also have some other kind of unique insight into the OP's frame failure, because the OP simply says that the 'chain stay snapped' while giving no indication of where, or in what pattern the damage occurred. That's hardly a clear picture....
That's a remarkable piece of conjecture - a good entry for the Turner prize on bicycle-related creative writing..
So, all that happened as you say - and the hanger didn't fail and the mech didn't snap first..?? I'm particularly interested in why the wheel bearings would 'seize' simply because of some rapid downshifting..??
yes, but my point is that the weakest part of the ensemble should have been the RD or the hanger - not the chainstay. Evidently, the chainstay has failed first for some reason, but my original observation was that under normal circumstances, one of the other two bits should have been the first to go...
Even at full extension, the chain does not follow a straight line through the mech. The guy has taken 2" out of his chain - that's 1" more than most RDs can typically tolerate, so normal convention goes out the window. That's why I'm a bit puzzled by your somewhat 'infallible' attitude to all this...
etc.
Since as you admitted:
the thread has developed into a dispute over which part should have failed first
Why don't you look at a couple of photos demonstrating inadequate chain length from
Park below
and ask yourself the question why would either the hanger or the rear mech fail if the chain is even shorter and tauter so that it becomes a straight line?
What seems obvious to me from those photos is that all that would look different is that the two jockey wheels would rise and fall a little respectively, meaning the mech cage would rotate up another few degrees.
So a question is if there is anything stopping that. I decided to examine a number of rear mechs I have and see whether anything exists to stop the mech cage rotating up for a short chain. Having looked at a fair range of Shimano MTB rear mechs from Tourney to XTR, I have been able to find only one, a C030 rapid rise, that has limited freedom, due to a seemingly cosmetic (!) feature, which limits the cage rotating a long way up. The vast majority have only one stop, meaning they can effectively rotate 330 degrees out of a possible 360 - their cages are free to rotate so far up, that their outer jockey wheels can easily swing up way higher than the chainstay (if it wasn't there - and unless one has a very small large front ring that eventuality is most unlikely to be relevant for the present issue)!
Further since the chain, taut or otherwise, only rides on the plastic jockey wheels, even if the mech cage does not line up perfectly in a straight line between large large looking down from above, which incidentally it does pretty well as part of the cage and parallelograms' design, it is most unlikely to impose any significant load on the mech either.
It seems to me this is what Mickle said all along, that the mech can handle short chain by design, and your suggestion that the mech and/or the hanger should fail before the chainstay when the chain is too short is unjustified.
Since the original equipment of a SL2 is a Dura Ace rear mech and not a MTB rear mech, and since I don't run many Shimano road components, we can look at its design schematic
here. As you can see there is only one stop pin (item 7) against cage rotation, meaning if the outer casing containing the pivot bolt that holds the cage is round, then it will have 330 degrees or so of rotation freedom.
This photo shows that to be the case.
If the above is too much theory and still too speculative for you, why don't you see for yourself by finding a bike, getting someone to hold the pedal still, pull the lower section of the chain tight towards the large front ring with one hand, and examine the vast degree of freedom the mech cage and pulleys can still give you with the other.
Imho a completely straight, taut chain between large large will not necessarily, indeed not generally, impose any unusual load on the rear mech, and therefore not the hanger.