You just defined the reason that they do the 2-second pre-send of the resistance adjustment. It is a simple, and singular change that is meant to address the lag issue with trainer resistance changes.
The problem is that the delta of power change, trainer in use, and even the practice of the rider can all impact the actual results of the changes. I have gotten to the point that my changes (along with the 2-second sends) align pretty well with the interval target blocks.
But regardless of that, I now ignore the summaries. I only focus on hitting the power (via proper cadence application and letting the trainer do it’s job) during the interval. I make a simple call at the completion of the interval based on what I saw from the yellow live power line vs the blue interval block. I usually end up with a slight late, but overshoot at the start, and a bit of a hang at the end.
Those result in the basic goal of the interval demand on the body and accomplish the desired loading. Splitting hairs over seconds early and late are really pointless if you step back and look at the goals.
The implementation of the 2-second early send works reasonably well without the extra demand of making it trainer specific or user adjustable. It is the law of diminishing returns and making it “better” now would yield only minimal improvement in results for the rider.