If you're pumping the hot exhaust air back into the chassis you're raising the ambient temperature in the cases which reduces the efficiency of the cooler.
Their argument was that said increase was negligible to the overall performance, which people don't agree with.
It's really complicated. One of the primary issues is that even testing this meaningfully is neigh impossible. The case used, the number and direction of case fans installed, the processor used, the motherboard used will have significant differences between outcomes. Just because x cooler was 2 degrees cooler than y cooler in this case, absolutely doesn't mean that x cooler will preform that much better in another setup. Moreover, the difference in between the two is almost menial.
The could spend hundreds of hours on every single unit and still come up with data that would barely be reproducible in the wild.
They're probably right that there's no practical difference between coolers above a certain surface area, but if you unapologetically do a half-assed job when you're already barely generating any usable data points as it is, it leaves a bad taste in everyone's mouth.
I think it was mostly due to a conflict of interest between them and noctua who they have collaborations with with things as recent as their noctua edition ltt screwdrivers and ltt noctua casefans