Over the years here I've seen weird score changes as a result of score paydowns in the myFICO alerts specifically with Experian and a bunch of others have reported similarly. I get the daily reports and scores now from Experian, and had a weird score change but now I've got daily data points and access to other scores... and I've got a diverse set of accounts with balances right now.
There's also been a bunch of debate as to whether it's all accounts, some accounts, or just revolvers or even just CC's over time.
So this thread is about determining what balances count for the number of accounts with balances ding, CC's, Amex Charge, HELOC, installment, even a PLOC of all things, and if possible where the dings for a comparitively low number of balances on the 3 major FICO models of the last two decades on Experian... bearing in mind EQ/TU will be different potentially, almost assuredly in the squirrley nature that is EX FICO 8 and balances.
I have a 30D and a CFA on Experian, so I'm in one of the lower top 8 scorecards, not the severe derogatory scorecards.
The initial data: Green marks good (decrease in balance percentage, or increase in score) Red marks bad (increase in balance percentage or drop in score... just trying to make the data easier for breakpoint analysis).
| Total Open accounts | 19 | ||||
| Installment | 2 | ||||
| Credit Cards | 14 | ||||
| Amex | 1 | ||||
| PLOC | 1 | ||||
| HELOC | 1 | ||||
| Balances | 11-Dec | 12-Dec | 13-Dec | 14-Dec | 15-Dec |
| Installment | 2 | 2 | 2 | 2 | 2 |
| Credit Cards | 4 | 3 | 3 | 3 | 2 |
| Amex Charge | 1 | 1 | 1 | 0 | 0 |
| PLOC | 1 | 1 | 1 | 1 | 1 |
| HELOC | 1 | 1 | 1 | 1 | 1 |
| Account % non-zero | |||||
| True CC | 28.6 | 21.4 | 21.4 | 21.4 | 14.3 |
| CC/Amex | 33.3 | 26.7 | 26.7 | 20 | 13.3 |
| CC/PLOC | 33.3 | 26.7 | 26.7 | 26.7 | 20 |
| CC/PLOC/Amex | 37.5 | 31.3 | 31.3 | 25 | 18.8 |
| CC/PLOC/HELOC | 37.5 | 31.3 | 31.3 | 31.3 | 25 |
| All Revolvers | 41.2 | 35.3 | 35.3 | 29.4 | 23.5 |
| All Accounts | 47.4 | 42.1 | 42.1 | 36.8 | 31.6 |
| EX FICO 8 | 746 | 743 | 743 | 743 | 746 |
| EX FICO 3 | 687 | 687 | 687 | 687 | 688 |
| EX FICO 2 | 728 | 732 | 732 | 732 | 732 |
I think at a quick glance we can utterly rule out an Amex charge card counting under FICO 8 for number of accounts with balances as if it did count, the increase in balance resulting in a drop, and then the Amex to zero should've produced the 3 point swing that I saw after the next CC reported $0. I also think that blows away the all accounts with balances theory as well.
Also can see that FICO 8 took a small dip on the payoff off one balance, and as it was a CC every single possible calculation went down which does demonstrate that a balance paydown can apparently result in a score decrease on Experian. The fact that it's exactly matched in magnitude a few days later suggests there's something awkward in Experian's dataset that lead them to this conclusion.
Finally FICO 2 appears to be calculating on different thresholds.
I put FICO 3 in there just because it is a FICO 04 model like EQ FICO 5 and TU FICO 4, but I don't know any lender that actually uses it.
I haven't done any more analysis up there but wanted others to take a look at it and maybe tease some more out of it. I will continue to update this as my balances change on my CC's and pay off both the PLOC and HELOC over the next few months.
