BR's test was continually adding test dust into the oil throughout the test. Just like an engine is constantly making debris as it's running. Of course the level of contamination BR was adding was way accelerated compared to an actual engine, otherwise the test would have taken a very long time to fully load the filters. Every engine produces debris as it runs. You've seen enough filter C&Ps here and the debris inside of oil filters to realize that.
The filter the leak gap measurements were on may or may not be the "norm", but it was an example actually measured. Some could have less or more of a leak gap depending on manufacturing consistency and quality. Fact is, basically all that were checked had some level of gap between the leaf spring and end cap.
Would you use a filter that's 84% @ 20u on your engine? Why would you chose a 99% @ 20u filter instead of a 84% @ 20u filter? If you think the engine is so clean, and you can justify lower efficiency filters because if it's not caught the first time maybe it will be caught the 2nd, 3rd or 4th time around, then why not just use lower efficiency oil filters to start with?
The other thing that nobody mentions is a filter with a big leak gap can let debris much larger that 40-50u past the media because the leak gap is much larger than that, whereas a low efficiency filter (like 99% @ 40 or 50u) with no internal leakage will catch most of the debris that large and larger.