It is not possible to estimate what?

There are some uncertainties that cannot be avoided or reduced, while instead they can be recognized and managed 

Putting more time and effort to improve estimations of the uncertainties that cannot be avoided or reduced, will not improve the accuracy and reliability of those estimations. It is a waste. And not recognizing those uncertainties increase the possible consequences.

Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them - Laurence J. Peter

It is not possible to estimate what
  • you don't know
  • has never been done before
  • has more uncertainty that certainty
  • you do not know that you don't know
  • is not yet invented, discovered or explored
  • you can learn how to do it only doing it for real
  • has no definitive or agreed formulation of the problem
  • is changing more frequently than the frequency you re-estimate it
  • has neither an immediate nor ultimate test/validation of a solution
  • has no "stopping rule" (i.e. we never know when or if the problem has been fully solved).
  • has an unbound set of possible solution and you can know which work only after trying them
  • has no solutions that match world views and frames for understanding for all the stakeholders
  • change substantially in presence of minimum changes of the initial conditions or external perturbations
  • ...
What else ?

And finally a Quiz to check if everything is crystal clear.
Given that:
- It is possible to reduce the consequence of uncertainty and risks, i.e. buy an insurance
- It is possible to have what you want at an arbitrary deadline, i.e. get a loan
- It is possible to make good profits from diversity in opinions/forecast, i.e. betting or investing in the stock market
What if someone want an high ROI without uncertainty/risks, at a pre-fixed deadline agreed by everyone?
Solution by Willem van den Ende :  that is called a software project ;-)

See also:
- Requirement's principles
- The dimensions of Agile

Print | posted @ Wednesday, February 8, 2012 4:10 PM

Comments on this entry:

Gravatar # re: It is not possible to estimate what...
by LudovicoVan at 2/8/2012 9:35 PM

It is usually necessary and *always possible* to provide an estimate.

An estimate can be more or less strict, but the error range must be explicit. An estimate always implies a feasibility and some cost/benefit study, hence it can result in a "don't go".

Murphy's Law: There is no uncertainty in engineering.

Comments have been closed on this topic.