>this is a simpler explanation for the observed phenomena (apply occam's razor here). certainly more than one explanation might conjointly apply, but yours is the more complicated explanation and thus implores more observation and measurement.And I thought Occam's Razor as guidance made the basic physics of the joints making wobbling unavoidable was the simpler explanation. I think we have different exposure to this product. I actually examined this desk at the store. If you look at the pdf[0] of the assembly, you'll notice 2 bolts that attach the pedestal to the posts. It doesn't matter how much one tightens the bolts because when you extend the desk past 40 inches with a heavy weight on the top (20+ pounds), that T joint will flex and deform. (See 1st video I cited for example).
As analogy, here's example of unstable weight distribution on the end of a joint (car and trailer): https://www.youtube.com/watch?v=6mW_gzdh6to
One doesn't need to examine 1000 different cars towing a heavy weight on the back to determine that it's unstable and will fishtail. The basic physics of the configuration will make it fishtail. That's the type of Occam's Razor I'm using. Yes, there will be variances in how the trailer hitches were installed to the vehicle, and also variances in the trailer ball diameter -- but those are not the dominant factors. To continue the analogy, some drivers with misconfigured weight distribution may not notice any fishtailing because they drove slower than 40 mph, or they didn't make any sudden steering overcorrections etc. Same situation with some desk users reporting no wobble by using it a shorter height, or placing it against the wall, or just not subjectively noticing it.
If one makes a desk that lets people put 20+ pounds at the end of a 40-inch extended pole that's attached to 2 bolts -- without triangle stiffeners, or welds, or cross-bracing, or whatever, ...it's going to wobble at least little bit. Based on hands-on experience with the desk, I contend the underlying physics of the assembly design overwhelms any deviations in manufacturing tolerances. Or put another way, if we consider the entire Gaussian distribution of manufacturing variances, all of the desk samples will still exhibit wobble to some degree as shown in the 1st Youtube video. Some may wobble less; some may wobble more. E.g. the telescoping cylinders (not CNC milled) that the users do not assemble may have variances that add to the total wobble.
I noticed others in this thread mentioned that the newer IKEA model IDASEN "solves" the wobble. (Or minimizes it.) If we look at its alternative design[1], lo and behold, we see that desk has added diagonal bars on the legs for extra support. Yes, it makes sense that structural triangles are stronger and stiffer than just 2 bolts fastening a T joint of the BEKANT.
[0] https://www.ikea.com/us/en/assembly_instructions/bekant-sit-...
[1] https://www.ikea.com/us/en/catalog/products/S29280934/