I don't want to get into a big long stats post because those threads decline pretty quickly into garbage but I semi-routinely keep stats on my teams including turnover rate and scoring efficiency. Against the press my turnover rate over time has been lowest using "normal" tempo and my scoring efficiency has been highest using "normal" tempo. Not a huge sample size but enough to make me believe.
My theory is this: the engine doesn't track each ball movement but that it weighs a bunch of factors into a "likelihood of turnover" and that tempo is a factor in two ways:
1) because it affects the length of a possession and 2) because I believe there's an over-the-top correction to turnover rate that is based solely on tempo, regardless of possession length.
First on possession length, I could see it working something like this: the engine looks at the total length of a possession and divides that by some average number of ball movements (either passes or dribbles) required to acheive that possession length. This means that the longer my possession goes, the more ball movements my team has to make. Also, I think the engine looks at who I have on the floor (bh/pass/ath/spd/off iq) and the players on the floor for the defense (ath/spd/def/def iq) and calculates a "probability of turnover per ball movement" for each possession.
Once I have that probability per ball movement and the number of movements required I know the odds of a turnover in any possession and its pretty easy to use a random number generator to tell me whether or not one occurred in any given possession. And honestly, I don't think that would be all that complicated from a coding perspective though I'm not a programmer.
If I have a crappy bh/pass team, and the odds of a turnover per movement is say 15%, the probability of a not having a turnover in a "3 movement" possession (uptempo) would be just over 60%, in a "6 movement" possession it would be just over 40% and in a "9 movement" (slowdown) possession it would be just over 20%.
That might tempt you to go uptempo but I also think tempo (aside from length of possession) is a factor. I think its well-documented that uptempo, on its own, causes an increase in turnover rate (i.e. aside from possession length). And as well, lowers your offensive efficiency due to poor shot selection.
I could be waaaaaaaaaaaaaaaaay off but this is my going-in assumption.