[mikeinatlanta]

Read the article and IMO it draws conclusions not fully supported by the data, but still contains relevant information in regards to shift points.

This closely relates to a motorcycle project I'm currently working. The bike is being prepped for the Texas Mile. Once the basic prep work is complete it will get a series of dyno runs.

IMO, the relevant data for acceleration is torque to the wheel. In order to determine best shift points we'll take a measure of the raw data torque output at the wheel for each gear. An overlay of that data for all gears will show very clearly which gear provided maximum torque at any given mph. Of course after first gear peak you have a continuing falling torque curve. Shift point is the point at which the falling torque in one gear is matched by the next higher gear's torque.

Basic example would be that you don't want to shift out of third if it is putting down 110 lb torque when fouth at that mph only gives 105. In this case you would hold third until the falling rate hit 105 and then shift to fourth. This method ensures that you are always in the gear that puts maximum torque to the ground. Calculating hp is irrelevant to determining best shift point with this method.

All that said, the comment about flat torque curve giving low average is absurd. I'm assuming he misspoke and meant that a flat torque curve gives a lower peak, not average.