Originally Posted by NSFW
So the question remains, what happens if you switch headers and don't change AFR?
He actually has both cases covered.
The EFI test was tuned with 1 header vs. tuned with the other header and AFR is right in line in both cases.
The Epic test shows a real world expectation from someone adding headers to an existing map, since they stated Cobb OTS stage 2 maps in both cases.
Both have their place, as both are realistic expectations for what a customer would do. A customer will either add the header and re-tune or a customer will add a header and not retune. The useful information is that there is a gain in both cases, but there is some danger in adding the header and not retuning, since AFR seems to lean out as a result. This is all very good and useful information to have.
Science is an interesting thing because one person's idea of what "holding everything else constant" means is different from another. For example, I don't think holding AFR constant is the correct way to do a "only one thing changes" comparison. It is possible for exhaust modifications to change EGTs at a given AFR, and would argue that if the header allows for cooler EGT at a given AFR, then EGT should be raised some by leaning AFR to create an "equal" margin situation.
I think KillerB has gone to greater lengths than most manufacturers to quantify the benefits of his product. I think there is enough information available that an educated potential customer can see the tests are reasonably well done. Perhaps not perfect
but there is enough information available to demonstrate there are genuine power gains to be had in the upper RPMs from the product.