Every year many more fail to hit the target date and target budget and exceed both by significant multipliers.
After a few Google searches I have not found a trend to indicate that the inclusion of frequent unit testing has resulted in less catasrophies.
Granted there are a lot of different factors that cause a software project to fail, and every now and again a new cargo cult produces a term that will fix it. (Extreme programming, Scrum, Lean etc).
Again, I do not find a pattern that it has reduced the catastrophes.
Does anyone know of any proper studies on this?
It acknowledges all the assumptions it makes, but we can all find problems with each and every approach to this.
Generally, looking for incident rates measures may give out some more papers on the same note.