Historical FICO 8 testing for me was always on revolvers. I mean always, no matter how many I had or how many installment lines open at the time.
What's a little strange, and admittedly I was in a derogatory scorecard was I always got the line at 1/3 of revolvers reporting a balance; every single time I tested it from 7 open revolvers to 9 open revolvers to substantially more now. Also Yes-Its-Me who's analysis I trust more than most also saw identical results on a clean file, though I think she might've been on a young credit scorecard at the time.
So there may be some scorecard dependence, which to be fair wouldn't exactly surprise me for things like accounts with balances (or revolvers) or even inquiries TBH; we get segmented into different tranches for a reason, presumably because different initial conditions produce different results based on similar inputs.
My theory anyway from a data analysis perspective, TT or Cassie might be able to explain that better I only have a passing knowledge of statistics and probability... I wasn't really studying enough at the time but upper-division Probability in the Math department at UCI was really the only class, ever, which I felt utterly befuddled in. Admittedly I withdrew before the drop deadline that quarter and I may never know if I could've wrapped my brain around that or not, I doubt I'll ever chase a Math degree again TBH.
@Birdman7.....so EQ5 has an 818 maximum, and TU4 has an 839 maximum...why such weird maximums rather than 850, or 900? I have never heard that before. I thought almost every scoring model had either 850 or 900 as the maximum, but I admittedly have never really focused on the lesser used scores either.
I heard it explained to me once upon a time from someone who was (allegedly) working on some of the very early algorithms that some of the top scoring items were mutually exclusive. I don't put a lot of credence on this but it's worthy of rumor mentioning at least.
The algorithms have a full score range of 850 for Classic scores or 900 for industry options; however, many of the scores prior to FICO 8 had a real world max of less than 850; we documented the ranges that we knew in Jello's old excellent post on FICO versions. Not certain I'd call the mortgage trifecta lesser used to be fair, but in virtually every mortgage score disclosure they put the real world min / max for the given score, and it's not 850.
This was a complaint so in FICO 8/9 they extended it to the full listed range to remove the confusion.
Thanks to you and Birdman, my learning something new every week is now covered...I had no idea any scores had these strange maximums. I agree also that lesser used is not really descriptive of the fico scores used for mortgages. Earlier versions is a better fit, as these scores can be a lot more vitally important if seeking a mortgage loan. In my case, I do not expect to ever have another mortgage, so I do not reaaly see as much need to monitor them. A wealth of knowledge here even if it does at times cause me to ask questions that demonstrate my ignorance of some of the credit info. Ignorance in this case is not an ugly word, and is descriptive of a lack of knowledge in a given subject matter...everyone is ignorant about some subject matters. I would not ever use it when referring to anybody else though because the word has been mis-used as an insult far too much. It is a shame though, because no other word really is as descriptive if used to explain why I can't do brain surgery or tell others how to do it.