That may be what you’re trying to do, but “movement size” isn’t an input you control - you only control force and speed, which it sounds like you already know. In these examples, based on how soft the acoustic sound of your picking is, it sounds like force is being modulated down. You set the click, that sets your picking speed. Then to make the movement “small”, you use the lightest possible touch for that pre-selected speed. That’s what sounds like is happening.
Again, if this works and you get results, that is all that matters. There is absolutely a rationale for not using too much force (or too much pick, which requires more force), to make the movements natural. I think this is really what the “small movements” crowd is really saying, without perhaps realizing it. And in this respect, yes, I agree. Don’t swing wildly using a lot of unnecessary power - that’s just going to make your movements feel awkward.
What I’m suggesting by using more power is an alternate “athletic training” style approach, where you don’t think too much and just try to move as fast as possible. If that produces a faster result, then you have something else to work with. It doesn’t have to be elbow - it can be the same mechanic. If it doesn’t, no worries, continue on!
I can only speak for my technique, but I sense no difference between “tremolo” and “regular picking”. Any speed I can pick, I can use it for individual synchronized fretted notes. The hyperpicking technique we have looked at may be different. Maybe it’s harder to control - I can’t say because I can’t do it yet. However for every other kind of motion mechanic I can do, they are controllable at every speed in my experience. Pushing your maximum speed higher just makes all the other speeds that much easier, no matter whether you are fretting notes or playing tremolo.
I didn’t count the notes, I counted the hand movements. So you get a free pass! 
Kidding aside, we’re talking about movement speed, not playing accuracy, so the hand movement is really what matters. Dead notes are ok for now. I just counted again using a three-second sample. Over that interval you played 35 notes, which is 11.6 per second, which is 700 per minute, divided by four is 175bpm sixteenths. This is about what it sounds like to me just by listening, so I think this is accurate.
It will! When we film players, accented notes are usually physically larger movements than other notes. Maybe more pick is involved too, maybe not - this is harder to measure. But the movement size is easy to see under the camera, and this definitely happens. This is something I learned about myself when we began filming slow-motion examples for our seminars. I didn’t know I was doing this at first. Viewers pointed out that when I advised using “conceptual accents” for chunking practice, they were not actually conceptual, they were real. You could actually see the accents in my playing where the pickstrokes are bigger, especially on single string licks like Yngwie patterns where speed was the primary variable.
What I can tell you just anecdotally is that focusing on the initial note of a grouping and hitting it harder feels like having an internal metronome. It is much easier to synchronize one movement out of every four (or six), than it is to synchronize every note individually. What we know from interviewing motor learning experts is that these movements eventually become stored as a single circuit that fires all the notes. In other words, you think “phrase” and the phrase plays. It may be that the accented note helps with this storage process. We don’t know, but it would make for a cool research study.
Either way, what I notice also is that players who do not do this type of practice almost always have synchronization problems, even if they have very fast tremolo speeds. I think this is why some players perceive “tremolo” as a separate type of picking movement that cannot be controlled.
So yes, hitting that note harder is somewhat more athletic and requires power. Maybe this is one aspect of being synchronized at faster speeds that needs more study. But I wouldn’t overlook the athletic aspect, even if it seems challenging at first.
This is not what you are actually doing. If you watch your 2wps example, you will see that the forearm moves when you do upstroke string changes, and doesn’t move when you do downstroke string changes. In other words, parallel forearm is not “neutral”, it is “primary up” two-way pickslanting. This is what John McLaughlin and Andy Wood do, and it’s fine. Every fast pickslanting player we have interviewed has a primary orientation. It only make sense. This way, any time you want to do a downstroke string change, no orientation change is necessary. You only have to worry about upstroke string changes.
This is what I was getting at with understanding your forearm position. You have already made choices for yourself, and these choices create limitations and also possibilities. It doesn’t really matter if that choice is supinated / primary down type orientation, parallel or pronated primary up - it just matters that you make one, and you understand what it is doing.
One thing which I will mention is that if you use a truly parallel setup, wrist-based crosspicking will not work. You must be strings-pronated or strings-supinated for that to work. Otherwise, you will need to involve some forearm or fingers to make up the difference.
Oh please. We’re appreciative of everyone that posts here because it allows us to keep doing what we’re doing. We learn, you learn, everybody learns.
You’re doing great so far I’m really just providing some details from things we have seen. Keep up the good work!
).
. Luckilly there are not too many things that I’d like to play that require that speed.