Thread: Gadgetman
View Single Post
Old 05-06-2013, 05:00 PM   #87 (permalink)
Master EcoModder
Join Date: Jul 2008
Location: Silly-Con Valley
Posts: 1,454
Thanks: 185
Thanked 251 Times in 189 Posts
Originally Posted by divedaddy03 View Post
One big thing guys keep talking about the A-B-A test (and others) too...but I'm not sure what that means. Can somebody please explain?
Here you go:

The simple version is:
- Try to set up conditions in as repeatable and consistent way as you can.
- Take one set of data without the modification.
- Take one set of data with the modification.
- Take a second set of data without the modification.

In this case "A" would be with no mod, and "B" would be with the mod. ABA helps to try to isolate any changes to the thing you are trying to test, rather than random variation. If the two "A" tests give results that are close to each other and the "B" result shows a clear improvement over both, then you have good evidence that the mod was effective. If not, then you don't.

Since driving style and external conditions (engine temps, air temperature, rainfall, weather, tire pressures, uphill or downhill, etc.) all have an impact on MPGs, you need to set things up with all of those controlled to as close to the same as you can get. For most of us, that means picking a calm day or time of day, doing several runs each direction in quick succession on a given stretch of road, using cruise control, and so on. It's kind of time consuming, quite frankly, but if you want good data you have to put the time in.

  Reply With Quote
The Following User Says Thank You to some_other_dave For This Useful Post:
MetroMPG (05-06-2013)