When I first found that installment ratios or utilization (whatever you wish to call it) mattered for FICO 8, I was in a dirty scorecard at the time with a 60D, collection, and tax lien.
The highest FICO 8 score I've seen reported was a 90D late at a 792; they had an 850 file otherwise though it was amazingly strong and the late was nearly 7 years old.
I'm sitting with a 60D late at 772 on TU FICO 4 which to me is a stupidly high score for the serious deliquency scorecard. My FICO 8 there is currently 760 without inquiries, sadly I don't have additional reason codes there but I don't know that there's anything materially wrong with it other than the lates and not enough credit history to drive it higher.
There are multiple derog scorecards though, public record and public record + deliquency are both lower than deliquency only to my knowledge in terms of max scoring. I don't think it's possible to hit 770 with a BK or collection for example on any FICO algorithm other than maybe FICO 9... my TU FICO 9 is 795 now and that's still in the recent late category too according to the reason codes. Also FICO 9 breaks out revolving history length in the reason codes even for the Classic score: I don't think that exists at all in FICO 4/8 except for FICO 8 BC (Bankcard industry option) in the dirty scorecards. Dirty = derogatory in my definition aka not the top 8 pretty(ish) scorecards.
Recent late is at least 18 months apparently since I picked it up April 2018 and it's still on there as of today 11/1/19. I'm guessing that might hang out till the 2 year mark.
@Anonymous wrote:
@Revelate wrote:...
The highest FICO 8 score I've seen reported was a 90D late at a 792; they had an 850 file otherwise though it was amazingly strong and the late was nearly 7 years old.
792 is quite impressive for a dirty card, though still below the 800 you said last year you thought was impossible.
Do you think bankruptcies are a dirty segmentation factor distinct from other public records (collections & tax leins)? If so, that would account for 4 cards. i.e.:
1. no bankruptcy or public record - max 792
2. public record - max 760-770?
3. bankruptcy - max 750?
4. bankrupcy & public record - max ?
I don't think the public record and BK scorecards are separate. BK / Collections / tax liens (when they were a thing) / judgements / etc are all about the same I think, they certainly score similarly and the reason codes are supiciously alike to my knowledge.
As far as I know for certain there's a delinquency and a public record scorecards.
I think they might get split based on age, but I misspoke a little earlier it's possible that a deliquency is just a modifier on the public record scorecard, I'm not 100% sure that it's an additional segmentation device when we're talking FICO 8.
We just don't really know, it's like the "recent late" bit which is still around after 18 months everywhere on TU. That might be a split between recent late and old late, and might be a simlar between recent PR and old PR, at least for FICO 8.
I had an interesting datapoint from a few years ago: with a tax lien on EQ already that was aged, I got another one tacked on brand spanking new. EQ FICO 8 moved around 40 points south, but EQ FICO 5 only moved 5 points.
As a result I don't think the age metrics are quite the same between the two, namely I think the older FICO 04 and prior models may not really factor age at all for PR's, but FICO 8 does. When that tax lien came off (filed in error, well sort of, I'd actually already paid it before it was filed) both scores rebounded to the prior values.
We just don't have great info as to how things are split.