This is veeery anecdotal if anything. What reflector was used with what bulb and ballast? Data is clearly lacking. Was different reflectors used? Cooled hood with glass or free hanging barn style?
To many variables to give a clear answer IMO. If you want to compare lights correct and fairly they have to be run at the same distance. Otherwise you compare apples to oranges in my book.
If the results were somewhat close, I would agree that those things REALLY matter, but its only a few inch height difference. They just used a different standered for the base line than you think should have been used.
You have to be willing to agree that no matter how close, or how good of a hood, that HPS has, it will never come even close to the average or useable PPFD that a LED array can produce.
Above you mentioned "numbers that are invalid and all over the place."
But the mapping for the sp3000 were real world tests done inside of mylar barriers, those PPFD measurements are really what you get when you use those lights. Even if you think the HPS map is understating its output.