They say in the video that the ECM is "mapped." Coming off a corner at mid-revs and full load, 99.99% chance it's a mega-knock event, shattering a piston, leaving the rod to beat the hell out of the block. You hear it happen, where the piston goes away, but there's a short time that the rod small end stays in the cylinder. As they keep driving forward, it comes out and you hear it turn into a sledge hammer. We see this play out here, over and over... Always people screwing with the stock ECM.
Stock ECM has
thousands of tables. It does really well to protect the engine. I see boost change from 1.8 bar on one pull to 1.3 bar on a different pull, as the ECM changes the target boost according to all kinds of things like IAT, how much load has been applied to the engine for how long, and on and on.
Here's a 2-3 gear pull I logged today with RPM, Target Boost, and Actual Boost. You see it goes to a maximum of around 1.6 bar @ 5000rpm with a commanded target of about 1.5 bar. By 7,000 rpm, the target falls away to about 1 bar, and actual is following, down to 1.2. Easy power gain is simply to change the target boost table to stay at 1.8 bar all the way across the board. Problem is, there are a hundred compensation tables trying to dial it back in, and/or fail-safes that will apply when there's too large of a deviation from target to actual. 10.5:1 compression direct injected turbo engine has a narrow window that it needs to run in and live, and with the stock ECM, that window is always on the move, and no one really knows the whole picture. Solution? Many "tuners" turn off the things that are moving that window around while trying to save it....