The Quincunx demonstrates that adjusting a process to “correct” it to the controls actually makes a process more out of control. I always remember it by recalling Bob Barker’s plinko board!
While I’m amused by grown adults dropping things down a board with pin in it, amusement was not it’s chief intent. The Quincunx was invented by Sir Francis Galton to demonstrate the central limit theorem. In particular the Quincunx board demonstrates that the normal distribution is approximate to the binomial distribution.
A Quincunx helps you to:
The Quincunx is a wonderful teaching tool as it is a great way visually to show the concept of randomness and with enough repetition, a bell-shaped curve. Obviously it helps you see the concept of the central tendency.
What isn’t obvious at the first is that it helps you understand the danger of adjusting your process while in the moment to the center. Watch the demonstration below. Doesn’t that make it clear how specification limits are used to monitor and then control variation?
The Quincunx demonstration reveals several things:
First, you can’t judge the process by sampling a part every so often. That would be like dropping just one chip down the board and then speaking to how all would fall.
Adjusting your process – moving where you drop the chip from – gives you a very different possible landing zone. It’s for that reason that the more you adjust your process to the center (nominal), the more variation you will create in your results. If you want less variation, keep a consistent process!!!
The counter to that is if you see a process generating a nice bell-shaped curve, it is probably not being influenced by outside forces like the six Ms.