Yes that is an interesting point. While on some level I agree with you (in the same way all people die from lack of oxygen to the brain stem), the path that failed startups take to get to the point of failure is very diverse. Every time I think I have seen every possible way for a startup to fail I am surprised with something new.
If I was going to go with your thesis, the common thread of failure is to be found very early - people are working on startups that just will never be successful because the underlying idea is bad. It is really hard to come up with an idea that can allow the unavoidable mistakes to be made and still succeed.
More than anything, it proves that you can use quotes and statistics in a manner to confirm your own biases. :) The decision tree has far too many branches to predict where the starting point will lead you but my point was, even when the recipe of successful startup (build something people want, talk to people, start small ... etc) sounds so obvious, the process of accomplishing it is anything but obvious.
If I was going to go with your thesis, the common thread of failure is to be found very early - people are working on startups that just will never be successful because the underlying idea is bad. It is really hard to come up with an idea that can allow the unavoidable mistakes to be made and still succeed.