Originally Posted By: demarpaint
This might answer you and Trajan. I recall seeing where compression results were posted, hyping another product used here, so what? I can write before and after numbers on paper and post them here, and show an improvement. What does it mean? It only means something if you believe me. MPG claims are attacked so can compression results. After all it is the word of the person posting the results, and if you feel that person is honest or not. After all this is tne Internet, and I'm also a skeptic.
I can also do a wet vs. dry compression test and depending on the condition of the engine get different results with a little squirt of oil, or maybe a different tool for the after test. If I'm the dishonest type, leave that bit of information out. But you already knew that.
All of this is dead on. It's so nice to read a post from you with which I agree totally.
Before I answer these concerns, let me say this: The idea isn't to have an open-and-shut case (or "proof" as you might call it). As you rightly point out, that is impossible to expect from dialog over the Internet. The idea here is to
increase the odds of being able to tell fact from fiction -- not only to convince others and root out shills on the Internet, but to make sure we're doing things better in our own lives. We can do this by bringing more evidence to the table, and by keeping everything open and above-board so others can scrutinize and criticize.
Raw numbers may be just as fakeable as words, but they provide some key advantages. When you have a series of before-and-after measurements, there is statistical math you can do that can be quite revealing. If that math makes the results look too clean or too random, that could mean it's more likely the data was faked. Excess variation within each set ("before" and "after") could also mean that any difference in the averages is likely to have been due to chance. Anyone with access to basic statistical software could repeat the analysis to confirm or refute, or they could look at it from another angle and bring something new to the table.
Again, it's not a slam-dunk, but it's leaps-and-bounds better than eyeballing something and then saying "take my word for it." And yes, it would be possible to fake a data set that could fool even an experienced statistician. But it's not something most people would know how to do, and it requires a fair bit of effort.
Here's another layer of sophistication: pictures and videos. For the trial I outlined, the person conducting it could post pictures or videos from each fill-up to a publicly available website on the day it happens, showing an odometer reading and the amount filled. If that person has a GPS-enabled smartphone, those pictures or videos could be geotagged. People on BITOG can then follow the experiment in real time. This still leaves doors open for mistakes or trickery, but it closes a lot as well.
Again, no one expects an absolutely unassailable case. Just something better than rough averages and vague, subjective testimonials.