Jeff if you read the thread from start Brian described the methodology and it's designers pretty thoroughly.
You bring up an excellent point others have made as well. About 80% of successful anchoring is technique, and that is is selling it short, especially if ground tackle is in specs for that size boat, regardless of anchor design. All these different tests by different people really tell you is how different anchors hold using one identical technique and tested in one particular manner. I find them interesting as a triangulation point, but not as sole the basis for decision making. In my now pretty extensive anchoring experience up and down both US coasts, in a a wide variety of conditions, using primarily CQR (many charter boats), Delta (main on my Hatteras, also charter boats) and Danforth (used as both main and backup on Hatt, also on many sailboats and charter boats and our Boston Whaler dinghy), I have found one constant in not getting a good set: my taking some kind of shortcut in the process.
I've posted this pic before, but I consider this type of situation to be an example one of my true tests of an anchor. This particular bottom was what I'd call "medium" mud. Clocking tidal current, residual winds maybe max 10 knots. Then, a sustained squall from out of the NW, max gust in the 40's, sustained maybe 25-30. Then after a few hours, back to normal.
Now, this is obviously a very hard thing to replicate in any kind of scientific study. Which why these types of controlled tests can only be used as one data point.
I suspect with the vessel that Fortress used, the Rachel Carson, it is probably possible to develop a route and have the vessel follow that route precisely within + or - 4'-5'. You could then set an anchor have the vessel drive over the location of the anchor and then reload at 180 degrees to the set direction (on the assumption that is worst case scenario. However you need base data to start with, which is what Fortress have provided, and then you need a very large wallet.
Many are critical of the protocol - so come up with a better way, publish it here - let's see how we can improve on what has been done before. 'Real life' is not much use as no-one in real life has the cross section of anchors available (no-one has a large enough wallet) nor the motivation.
The test did not do 180 degree swings, but an anchor that has an approximate average holding capacity of 300lb in 5 pulls is unlikely to suddenly be better after a reverse of pull.
I accept the limitations - its a lot better, if not infinitely better, then it not being done.
But if you want a better test - make some suggestions.