Skip to main content

Your Optimistic Bias won't Allow Your Estimate to Improve

When asked, many people wish to get better at estimation by tracking actuals and then assuming that some form of reflection upon those actuals will lead to better estimation.  I've long held a belief that it doesn't happen quite so simply in reality.  There are many assumptions in that proposal that are just glossed over.   Let's list a few (just for fun):  Actuals are accurate and precise representations, time will be allotted to reflections and learning,  people (teams of people) will all arrive at similar conclusions and learn from the reflection of estimate not equal to actual, future behavior will change significantly and quickly, among other assumptions.



Well finally science has something to say about this.  A study: "How unrealistic optimism is maintained in the face of reality" by Tali Sharot, Christoph W Korn & Raymond J Dolan published in Nature Neuroscience (2011) has some fMRI proof that these behaviors are hard to change.

In the study the authors find:
"Unrealistic optimism is a pervasive human trait that influences domains ranging from personal relationships to politics and finance. How people maintain unrealistic optimism, despite frequently encountering information that challenges those biased beliefs, is unknown. We examined this question and found a marked asymmetry in belief updating. Participants updated their beliefs more in response to information that was better than expected than to information that was worse. This selectivity was mediated by a relative failure to code for errors that should reduce optimism. Distinct regions of the prefrontal cortex tracked estimation errors when those called for positive update, both in individuals who scored high and low on trait optimism. However, highly optimistic individuals exhibited reduced tracking of estimation errors that called for negative update in right inferior prefrontal gyrus. These findings indicate that optimism is tied to a selective update failure and diminished neural coding of undesirable information regarding the future."
So, is this scientific challenge to your ability to get better at estimation enough for you to quit tracking actual time spent on stories/tasks?  The evidence is that your bias will keep you from learning except in the cases that prove you are faster/better/more awesome than you estimated you were.  Which leads many managers to decide that there teams are sand-bagging and just not as good as "we were in the older days," a different human behavior to study.

See Also:
Estimates? We Don’t Need No Stinking Estimates!
"How a hashtag (#NoEstimates) lit the nerdy world of project management aflame — or at least got it mildly worked up." A nice summary of the dust-up.  Imagine if the tag would have been #LeanEstimates?

The #NoEstimates blog by Woody Zuill

Your Brain won't allow you to Believe the Apocalypse could actually happen

Why your Agile Teams are Bad at Estimating - by Heidi Araya

The planning fallacy, first proposed by Daniel Kahneman and Amos Tversky in 1979, is a phenomenon in which predictions about how much time will be needed to complete a future task display an optimism bias (underestimate the time needed).





Post a Comment

Most Popular on Agile Complexification Inverter

David's notes on "Drive"

- "The Surprising Truth about what Motivates Us" by Dan Pink.

Amazon book order
What I notice first and really like is the subtle implication in the shadow of the "i" in Drive is a person taking one step in a running motion.  This brings to mind the old saying - "there is no I in TEAM".  There is however a ME in TEAM, and there is an I in DRIVE.  And when one talks about motivating a team or an individual - it all starts with - what's in it for me.

Introduction

Pink starts with an early experiment with monkeys on problem solving.  Seems the monkeys were much better problem solver's than the scientist thought they should be.  This 1949 experiment is explained as the early understanding of motivation.  At the time there were two main drivers of motivation:  biological & external influences.  Harry F. Harlow defines the third drive in a novel theory:  "The performance of the task provided intrinsic reward" (p 3).  This is Dan Pink's M…

What is your Engagement Model?

What must an Agile Transformation initiative have to be reasonably assured of success?

We "change agents" or Agilist, or Organizational Development peeps, or Trouble Makers, or Agile Coaches have been at this for nearly two decades now... one would think we have some idea of the prerequisites for one of these Transformations to actually occur.  Wonder if eight Agile Coaches in a group could come up with ONE list of necessary and sufficient conditions - an interesting experiment.  Will that list contain an "engagement model"?  I venture to assert that it will not.  When asked very few Agile Coaches, thought leaders, and change agents mention much about employee engagement in their plans, models, and "frameworks".  Stop and ask yourselves ... why?

Now good Organizational Development peeps know this is crucial, so I purposely omitted them from that list to query.

One, central very important aspect of your Agile Transformation will be your Engagement model.  

Exercise:: Definition of Ready & Done

Assuming you are on a Scrum/Agile software development team, then one of the first 'working agreements' you have created with your team is a 'Definition of Done' - right?



Oh - you don't have a definition of what aspects a user story that is done will exhibit. Well then, you need to create a list of attributes of a done story. One way to do this would be to Google 'definition of done' ... here let me do that for you: http://tinyurl.com/3br9o6n. Then you could just use someone else's definition - there DONE!

But that would be cheating -- right? It is not the artifact - the list of done criteria, that is important for your team - it is the act of doing it for themselves, it is that shared understanding of having a debate over some of the gray areas that create a true working agreement. If some of the team believes that a story being done means that there can be no bugs found in the code - but some believe that there can be some minor issues - well, …

Refactoring - examples from the book

Martin Fowler's book Refactoring:  Improving the Design of Existing Code has a simple example of a movie rental domain model, which he refactors from a less than ideal object-oriented design to a more robust OO design. Included in this Refactoring_FirstExample.zip Zip file are the Java source code files of the Movie, Rental, and Customer classes. Along with a JUnit CustomerTest class. Using these example source files you too can follow along with the refactoring that Fowler presents in the first few chapters of his book.


Metrics for a Scrum Team (examples)

What metrics do you collect to analyze your scrum team?

We live in a world of data and information.  Some people have a mindset that numbers will diagnose all problems – “just show me the data.”  Therefore many directors and senior managers wish to see some list of metrics that should indicate the productivity and efficiency of the Scrum team.  I personally believe this is something that can be felt, that human intuition is much better in this decision realm than the data that can be collected.  However, one would have to actually spend time and carefully observe the team in action to get this powerful connection to the energy in a high-performing team space.  Few leaders are willing to take this time, they delegate this information synthesis task to managers via the typical report/dashboard request.  Therefore we are asked to collect data, to condense this data into information, all while ignoring the intangible obvious signals (read Honest Signals by Sandy Pentland of MIT).
What if …