Oh I think I was just misunderstanding you then. That was actually the video I was trying to find earlier by the way!
The reason the velocity increase creates the additional tumble or swirl, or both (it actually depends heavily on RPM I believe) is in the head/valve design. Basically, with low velocity, the particles kind of slide off the valve and the ports and just "seep" into the piston. You notice there's only a slight tumble in the cylinder without the tumble valve in this case. Its not stimulated to do so, and as such a lot of the fuel that is injected (in this case into the cylinder directly) doesn't spread out, it kind of follows gravity and stays lower next to the piston. This leads to a cooler piston of course, but you have an obvious lean -> rich trend. The rich areas will generate excess soot, the lean areas excess NOx, and the flame front wont be as uniform (which this simulation doesn't seem to show, it appears to just be a CFD of the air and fuel particles), and that means a loss in cylinder pressure and torque.
PI does a little better in this regard, but still suffers nonetheless. And actually DI can cheat a little bit with their piston bowl design itself. A lot of DI engines have a unique bowl design in the piston, and this is to encourage this fuel mixture process you see happening. Basically the bowl redirects fuel from being shot at the piston up into the cylinder air itself I believe, and I think this is why things like the Ecoboost truck dont have IMRC, but I will admit, I'm not an expert on the whole process in a DI engine on fuel spreading and piston design, being much more familiar and having read much much more on the design for PI stuff.
But basically, the idea here is that valve is speeding up the particles into a jet, like putting your finger over your hose to make it spray. Say you're cleaning off dirt on your garage door. If you just let it trickle out, the water runs down the garage door. If you spray it out with a higher velocity, the water will bounce off your garage more in a jet, even though its the same amount of water. Same idea here. The IMRC basically causes the air to jet up, and it bounces off the head/valve design creating the charge motion that helps the fuel particles spread around. Same idea here as say adding food coloring to water. If I just poor it in, it slowly seeps in. If I add some velocity and stir it up, I significantly accelerate this mixing process.
So yes, they're skipping steps and may be cutting it short saying: Velocity = Power & Torque here, and you're right thats misleading, and I'm not trying to encourage that view point. Its actually the side effect of what they can do with the increased flow velocity, that is, promote better charge mixture, which promotes a better burn, which means better everything: power, emissions, etc.
Also, the restriction doesn't mean it'll be less CFM. You can have 300 CFM of water move through a 3" pipe, or through 6" pipe, but in the 3" pipe itll be moving faster. That's what theyre doing.
I know what you're thinking, that causes a pressure drop, and you're right, it does! But remember, the ECU targets a set Desired Airflow. Based on current VE and desired airflow, it can figure out a needed manifold pressure and throttle plate aircharge needed, and adjusts the throttle to reach this goal. So if you keep airflow desired the same, but now add that pressure drop, it opens the throttle slightly to counteract that.
Essentially, you operate the engine in a condition knowing that pressure drop is there, because you're using that to increase the overall efficiency of the process. You're trading unused throttle for increased performance at the same airflow.
At high velocities and flows you get at WOT, you naturally get the tumble, and no longer want the IMRC there increasing the velocity more (and causing that big pressure drop!) You want WOT, and you want the lowest pressure drop you can get, so thats why you fold them out of the way. You also are lucky in this situation to not need them anymore to create the velocity you needed to get a jet of air, as the natural amount of air you're shoving through does this for you
I hope I'm explaining this clear enough, and I really thank you for not getting upset with me as I commonly encounter whenever something like this happens on other forums! Thats why I'm back and making another post, because you weren't being argumentative just to be mean, and I really appreciate that. Now I just have to hope I can explain it so you see
how it increases power and economy. But I do agree as well, its not as simple as velocity = power here, but what they use that velocity to do, and they do kind of leave that out of their explanations, but they're not outright lying that the IMRC doesn't help them increase power and efficiency at these lower airflow values.