Skip to content
Investigating the effects of feedback
JavaScript HTML
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
fellers.html
jquery-1.8.3.js
jquery-ui.css
jquery-ui.js
jquery.flot.axislabels.js
jquery.flot.js
jstat-core.js
jstat-vector.js
notes.txt
readme.txt

readme.txt

A simple web-page to simulate lots of Feller's walks.
Each Feller's walk is 1000 steps long. At each step you flip a fair coin
and keep a running total (starting at zero), 
  o) adding one to the total if you flip a head,
  o) subtracting one from the total if you flip a tail.
How will the total behave as the walk proceeds?
Most people's intuition is that the total will hover around zero.
Does it? Find out!
https://rawgithub.com/JonJagger/Fellers-1000-coin-tosses/master/fellers.html

Every N steps (on average) intead of flipping the coin, you look at the total:
  o) if its negative (more tails than heads), pretend you 
     flipped a head, and add one to the total
  o) if its positive (more heads than tails), pretend you
     flipped a tail, and subtract one from the total.
This increases the likelihood of the total staying nearer zero.
How effective is this feedback? Find out!

I discovered Feller's walk on page 77 of Don Reinersten's excellent book The Principles of Product Development FLOW.

http://jonjagger.blogspot.co.uk/2012/06/principles-of-product-development-flow.html is a blog entry with some snippets from his book.

http://jonjagger.blogspot.co.uk/2013/01/fellers-walk.html is another blog entry on Feller's walk.

http://jonjagger.blogspot.co.uk/2013/11/fun-with-feedback-frequency.html is yet another blog entry detailing how to use the simulation and revealing how effective the feedback is.
You can’t perform that action at this time.