Joint motions don’t have an escape. It’s just your joint moving back and forth. The same joint motion can have different escapes depending on the rest of your technique.
The Al Di Meola motion you’re pointing out is just the wrist moving in what we call a “reverse dart throwing” motion. Slighly up and to the right with respect to its own axis.
If you do this on a guitar with the arm planted relatively flat, you get downstroke escape. Cameron Cooper on YouTube is a really great textbook example of what this looks like:
Notice there is little or no upward pickslant. When the arm is held this way the pick is not really slanted compared to the strings, which is why we don’t call this “upward pickslanting”. It is downstroke escape (DSX), yes — but it is very often not upward pickslanting.
However if you were to place your arm on the guitar like Albert Lee and make the same exact motion, you would get a trapped motion or a USX motion. Because the arm is rotated.
Edit:
There are several reasons we have the table tapping tests:
- To show you that most people have at least one, and sometimes many, joint motions they can make that are fast — even people who have never played fast on a guitar before
- We want to get a reading on that speed so we can make sure it’s “normal”, i.e. not super slow like 120bpm or something
- We want you to feel what “fast and easy” feels like, so you can shoot for that when you have an actual guitar in your hands
The table tapping tests aren’t really there to help you learn specific joint motions. Meaning, we don’t want you to learn “Al Di Meola”-style motion from tapping on a table. But… doing any joint motion fast, and feeling that it’s fast, is a huge help in figuring out what it’s supposed to feel like when done correctly on an instrument. So in that sense, I actually do consider the table tapping tests and important learning opportunity, in the general sense.
TLDR you are trying a motion on a real guitar, and it does not feel fast and easy like tapping on a table, then there is a good likelihood you’re not doing it as efficiently as it could be, and some more experimentation is the way to go.