Wrong. You have no idea about scientific testing. If all the variables remain the same and you're only changing one thing it's still a valid test regardless of the test environment. It doesn't matter if it's in or out of a case. All that affects is the absolute temperature. The whole point is that the loop order doesn't matter.
Of course it matters if it's in or out of the case. Which way does the air come from when you drive your car down the road? The same principle applys.
Also I'm a plumber, going on 10 years. Water cooling a computer correctly is child's play compared to what I do for a living bud. It's ok man. I water cooled wrong for years until I finally pulled my head out of my ass and learned.
Plumbing school, scientfically, helped me understand the principles better, including some of the equations that go along with those principles. The problem being, a computer water cooling is so small, it's actually difficult to refine the math involved on paper.
Most, and I mean pretty much all, heating or cooling systems do it all at once before it goes into the system and then branches. Kinda like blood in your body.
So all the cooling happens in a cooling tower (cooling) or boiler (heating). It doesn't cool off, go down a pipe, and then cool something and then cool off again and then cool somwthing. The more time water gets to spend being cooled or heated before doing its job the better. This is why we have pump curves for efficiency.
I have stood over my last setup with thermal probes, thermal sensors all over the case. Measuring everything until I found the best scenario. Best setup for ME was: back to back 360 rads before cooling any cores. And EVERY fan was pulling room air. Pump was half speed as full speed was a few degrees warmer on the cores under load. In the winter under max load (folding@home) my cpu was never higher than 25c and the video never broke 45c. My cpu idled at 2c over ambient.
So tell me my man, who do ya think is doing it right? The problem you're having is A/ fan direction. B/ thinking internal case air is fine when it's totally not.