Dr. Barseghyan’s key message is the unreliability of historical data to predict future project outcomes using simple regression analysis. This is similar to the core argument I raised in my paper Scheduling in the Age of Complexity presented to the PMI College of Scheduling conference in Boston earlier this year. Historical data is all we have but cannot be relied on due to the complexity of the relationships between the various project ‘actors’. As a practitioner, I was looking at the problem from an ‘observed’ perspective it’s fascinating to see rigorous statistical analysis obtaining similar outcomes.
A counterpoint to Dr. Barseghyan’s second argument that improved analysis will yield more correct results is the work of N.N. Taleb particularly in his book ‘The Black Swan’. Taleb’s arguments go a long way towards explaining much of the GFC – models based on historical data cannot predict unknown futures. For more on this argument see: http://www.edge.org/3rd_culture/taleb08/taleb08_index.html
Personally I feel both of these lines of reasoning need to be joined in the practice of modern project management. We need the best possible predictors of likely future outcomes based on effective modelling (as argued by Dr. Barseghyan). But we also need to be aware that the best predictions cannot control the future and adopt prudent, effective and simple risk management processes that recognise each project is a unique journey into the unknown.
I would certainly recommend reading Dr. Barseghyan’s paper.