Well I did not consider the Soviet really crippled until mid 42, was semi-crippled until then (still a huge army holding itself until then) and the Allies where super rich. I knew I was in deadly trouble by early 42 of course, but I thought it was still worth a little try. So that is why I extended it a little while.
For the mechs, you could very well be right of course, but a single game can not show much as many variables are at play, such as the remainder of the strategy, mistakes, luck, level of play (of course)... For example, the two players I played to most (they will confirm it if they want), have never scrapped the mechs yet managed to easily keep the Soviet alive and strong every game even before the patch (of course, I also believe they might be two of the best players here). And with all the mistakes I made, not sure how the mech variable can be isolated from the rest? As I said, maybe you are right about the mechs, but not sure how than can be rigorously tested. It would be literally be needed to play the exact same game with only that thing changing. And even then, it will only show it for that exact game (or type of?) I think.
Also, my educated guess would be that in a game with so many variables, there are probably very little absolute strategies. Scrapping mechs might be good for some strategies but bad for others; hard for me to imagine a single answer always right to a problem with multiple moving variables.
If there is one think I have seen from all the AAR is how with so many different approaches and strategies, players can still potentially achieve the same results, more or less...