Today is a good day to sleep early for once.
Yesterday, I checked the weather prediction for today: rain for the whole day. Today morning, I opened the window and I saw no rain. I took the risk and rode my bicycle. Sometimes, our theories fail to predict well enough our reality.
Today, Dr. Steve Hinske gave a guest lecture on IT Governance for the course of Management Information Systems at ETH. I was under the illusion that not given enough time to test theories in the ever-changing environment of IT renders useless any predictive of its models (since we try to maximize an expected future profit by allocating scarce resources in an alternative way). Now I think it is more like the weather. An example of an unsuccessful prediction does not mean that the predictions are not better than random.
You will be able to find the slides on the course’s homepage.
To make it clear, the point of this parallelism was to underline that sometimes small samples ( limited or biased experiences) can lead to wrong conclusions.