Thursday, January 12, 2006

Estimating and Forecasting Traps

We make estimates everyday. For most of these estimates we don’t normally get feedback on their accuracy therefore as a result our minds never become calibrated for better accuracy. In making big decisions involving estimating under uncertainty we use the basic instinct we have, which is not accurate, which has a distorting effect in uncertain situations. There are 3 of the most common of these uncertainty traps:-

  1. The overconfidence trap – we tend to be overconfident about our accuracy. This may due to the fact that anchor has been set in our mind and we are pulled to that anchor therefore being overconfident in our estimates in order to get closer to our goals. Aggressive people will fall into this trap.

  2. The prudence trap – this takes the form of overcautiousness, i.e. a reverse to the first trap above. This is actually practiced in many industries where worse case scenarios are required to be examined even though the odds of those circumstances actually coming to pass were infinitesimal.

  3. The recallability trap – even if we could avoid the traps described above we may still fall into this one. In our mind we forecast based on past events in our memory. We could be overly influenced by dramatic events that leave a strong impression on our memory. E.g. a plane crash that killed hundreds but daily killing due to car accidents do not forge as much an impression in one’s mind as for a plane crash. Anything that distorts your ability to recall events in a balanced way will distort your probability assessments.



Ways to avoid these traps:

  • To avoid the overconfidence trap in making estimates, always start by considering the extremes, the low and high ends of the possible range of values. This will avoid being anchored by the initial estimate.

  • To avoid the prudence trap, always state the estimates honestly and explain to anyone who will be using them that they have not been adjusted. Test estimates over a reasonable range to assess their impact. Do a sensitivity analysis to reassess your estimates.

  • To minimize distortion due to your memory, carefully examine all your assumptions. Invite someone to offer fresh perspectives that do not have those memories that may have affected you. Get actual statistics if all possible.

No comments:

Post a Comment