The Logic of Failure - Dietrich Doerner

It seems to me that the best way to succeed is to avoid failure. There are many books on succeeding, but few on failure, and avoiding pitfalls seems pretty vital. After all, learning by experience is really learning by failure, and it seems far more convenient to learn from others' failures than your own. So, I had high hopes for this book (translated from the German original - why am I not surprised that this book comes from the land of schadenfreude?).

Sadly, this is not the book I was expecting. The examples it focusses on are from psychology experiments, as people basically fail to manage systems based off differential equations they don't see the internals of. And while it says there is no magic bullet to not screwing up, it then seems to try to provide one. The advice it provides is pretty prosaic, and spelt out simply, so it really does seem like a management book. So much less than it could have been.

Much of the emphasis of the book is on how, when confronted with a complex system, people tend to disasterously oversimplify. It therefore seems odd that the 'complex systems' in question appear relatively simple, if opaque. While it may give us a simplified understanding of complex systems, it ssems unlikely that a simplified understanding of failure due to complexity will helpe us deal properly with real, complex complexity. Ironic, really.

Otherwise, I get the impression that the author isn't necessarily the world's most mathematically astute guy. Ok, he's a psychologist, and being a polymath is probably a bit too much to ask, but 'statistically significant' is used like voodoo, and the maths bits feel woolly. I don't think it helps that the quality of the graphs appears to have suffered massively in translation - I don't think Tufte would approve.

The most interesting thing, I thought, was in the final chapter, where the author reveals that in the abstract management scenarios, the student participants did far worse than people taken from senior posts in industry. The book's reading is that experience is important, and that being able to manage complex systems is something you can pick up. On the other hand, perhaps there might be a correlation between being intrinsically good at managing complex systems, and being promoted into jobs that require the ability manage complex systems? The book doesn't even mention this possibility!

Posted 2007-07-15.