Paul Hammant's Blog: Misapplying Bayes-Theorem to Agile Versus Waterfall
Thomas Bayes lived from 1701 to 1761. Apart from his life in the Presbyterian church as a minister, he contributed something lasting to probability - a theorem in his name. Here’s a formula for it, cribbed from Wikipedia:
For a reasonable intro to usefulness of the theorem in real life, see this 2.5 minute video clip by Luana Micallef that uses mammography screening for breast cancer as an example. You won’t regret taking the time to watch that.
Misapplying Bayes’ Theorem to Agile vs Waterfall.
If you’re planning on having an 18 month project from start to the first go-live, then it is best choose Waterfall as your methodology. Why? Well, there’s a chance with that duration that you’ll lose a percentage of your staff en route, and need the documents/artifacts as a way of reducing the risk and cost of acclimating (British: Acclimatizing) new starters to what’s currently in progress.
If you’re planning on a 30 day project (from start to its first go-live), then choose Agile. Why? In 30 days it is unlikely that you’ll have staff quitting, and you’ll be able to skip the interim documents/artifacts and use the unit/functional tests you made along the way as living documentation.
[ Here’s the misapplication ]
If you choose Waterfall, your project will take 18 months, and you’ll lose a bunch of the staff along the way. If you choose Agile, your project will take 30 days, and nobody will quit part-way through.
Well, it’s almost nothing to do with Bayes’ Theorem, so I’ll get my coat.
Feb 20, 2013: This article was syndicated by DZone