How to properly test modifications (AKA: don't believe every MPG claim you read)

[j-c-c]

I have made claims here in the past, where following all the parameters to do 'proper" testing just was not practical in trying to control all variables. My test, rather than do a sample to extrapolate, like a 200 mile round trip over the same section of roadway while controlling as many conditions as possible, I'd do a year's worth trips over all the same roads in nearly all conditions say for many tens of thousands of miles as a baseline for an average.
I'm satisfied with my results.

[j-c-c]

IMO there are two main trains of thought on DIY testing, control methodically all the variables known and/or possible, or make a test controlling what is possible but as long and extensive as allowable, to allow the "random" variables to hopefully cancel in order to achieve a credible result.
Pick your poison.

[freebeard]

IM[H]O there are two main trains of thought on responding to AI trolls.

Don't feed them.

Why not troll them back, they never respond.

[j-c-c]

I have found from years back they don't recognize the use of color in the verbiage to indicate emphasis on a topic.

[Riyuzaki]

Originally Posted by ssnsvibe09

What is your measurement techniques and what sort of accuracy you looking for?

I invested in a high end "consumer" (fully aware the Professional will add a digit to the price) OBD interface that can read the latest CAN data if I am willing to pay for the software.... I hoped the data provided would be 95% accurate. Time will tell.

Well.... That has me thinking... and more questions....

I got 25 years repairing digital feedback control systems that measure stuff in the micro range at a >1000 Hz rate and be able to read that accuracy in a digital word and references. I was expecting to see the same kind data on the OBD but within the milli range at best.

Does the OBD interface provide raw standardized data as detected on the CAN bus?

If not standardized, where is it interpreted?

If not raw, what is filtered out? What is the resolution for the filtered data?

My minimum sample rate that I am hoping for in 1 Hz but I expect 10 Hz, is that achieveable?

I have not logged data from a car ECU yet and I hope to be able to generate references that will us the best driving technique for different circumstances throughout my commute.

Wow, really appreciate the depth you're coming from—25 years in digital feedback systems is no joke! I get what you mean; going from high-frequency, micro-level accuracy to OBD's relatively vague data can feel like a step backward. OBD often gives sanitized or filtered data, and yeah, it’s not always “raw CAN.”

Standard PIDs are limited, and once you hit manufacturer-specific stuff, it's a bit of a black box unless you dive into reverse engineering. As for logging—1 Hz is definitely doable, and in some cases 10 Hz is too, but it really depends on the interface, protocol, and how much access you get.

Your goal of building personalized efficiency profiles for your commute is honestly inspiring. Would love to see where you take it. Keep us updated!