One problem with these postmortems is that they are biased: successful products are more likely to get postmortems than unsuccessful ones, partly because the pu

One problem with these postmortems is that they are biased: successful products are more likely to get postmortems than unsuccessful ones, partly because the public is more interested in products they have heard of and partly because the creators would rather not publicize failure. But a postmortem of an unsuccessful product made by otherwise-competent developers could be more instructive than a postmortem of a successful product, especially in the "what went wrong" section. The book "Dreaming in Code" feels like such a postmortem, and I think I learned more from it than from the various Game Developer postmortems I have read.