« Victory 2025, Pt. 2 | Main | The Droids You're Looking For »

August 28, 2010


Joseph Fouche

Coming from IT, nothing says "jumped the shark a decade ago" more than Gartner, Forrester, and their ilk using something in one of their "forecasts". They're usually five years behind the cool IT kids' consensus.

I'm not 100% sure who their intended audience is. IT forecasts seemed to be calculated to anger 1337 hax04r wannabes on slashdot and that's not that hard. I would be frightened to meet an IT professional who was shocked by a IT consulting group's forecast.

Of course such people exist. They're the ones still forcing people to use IE6 at work.


IE6? Dang!


...and one day, long after your mum starts using the word "chaoplexic", some obscure thinker is going to define it.

Regarding IT forecasting, it's all utterly worthless, and there is a wonderful Harvard Business Review piece out there about the impressively consistent fact that business forecasts are wrong: http://web.hbr.org/email/archive/dailystat.php?date=051710

Specifically, over the years 1985-2010, consensus forecasts for S&P500 corporates were always wrong. They were wrong in being overoptimistic - in fact, on average they were only hit twice, at the peak of the housing bubble and in the mid 90s. In fact, the average forecast foresaw growth of 13%, about 4 times as fast as GDP. Further, they hugely underestimated volatility - the average forecast is nice and smooth, the out-turns have enormous variance.

If you see anyone with a projected CAGR, run!


I think I was being a bit too kind with Gartner's report, to be honest, which basically seemed to be a description of the last ten years than anything else.


The study of chaoplexity despite all its mathematical rigor is like studying an animal in a zoo. It's not really a science in that we lack a way of testing the theories. We can only make observations that are not real-time because of our analytical abilities.

That is the next step I think, to quit theorizing and observing, and set to coming up with ways to experiment with these systems. Additionally, a more significant bridge between the agent and their system is a good way forward. I know the term has already been used for another field of study. But, calling it metapsychology resonates with me. Especially in the sense of a S.A.C.

The comments to this entry are closed.