AAoA thresholds? UPDATE 5/2/19

[Anonymous]

---

@Anonymous wrote:
 But I promise you that it’s more than a couple points BBS. What other difference is there between you and anyone else on your scoresheet? why are you the special case that only gets that very small spread? Because you cannot optimize. I’m sorry but I don’t see any other reason IMHO.

---

See, I don't think I'm a special case.  I just don't agree with the potential number of points you're assigning to "optimization" even though it can't be quantified and we're both speaking from our gut opinions.  If I see 8 points in going from 100% of accounts with a balance to 25% of accounts with a balance, which crosses multiple threshold points I'm sure, I personally don't feel crossing that final "optimal" threshold point is going to be worth some [relatively] huge amount.  IMO, it's not going to be even worth the 8 points gained from crossing the multiple earlier thresholds.  That's just my opinion.  Less than 8 points.  Whether that's 3, 5 or 7 is really all S&Gs talk to me, as I'm not going to consider that much outside of a couple of points. 

I guess another way to look at it is what is considered "optimized" on other scoring models verses 1 notch off from optimized and look at the score change associated with crossing that 1 threshold. 

Also, what is the optimized percentage threshold for TU 4?  Do we know it?  12.5%?  8.9%?  I honestly don't know.  Do we have any data points from anyone on the forum that tested TU 4 in crossing whatever that threshold point is?  Say, going from 3/20 accounts with a balance to 2/20?  If so, not only what was their gain, but what was their gain relative to the gain in going from 100% (or 50%, whatever, a measurable/known number) down to 3/20 accounts with a balance?

[Anonymous]

I think AAoA is one factor per algorithm. I don't, at this point yet, buy into the fact that TU FICO 04 is considering the two aging factors (revolving/installment) separately AND that both have a higher precedence than your # of accounts with a balance. For the industry options, yeah maybe. The classic algorithm and we are just now hearing about it 15 years later? No, hard to reconcile that. I wanna see your reason codes from a MF 1B for TU. If it's there, you'll convince me.

[Anonymous]

We think it's either 15% or 20%. With what you said, I agree, it's not gonna be worth that much, but, I wouldn't doubt the same spread as others, whatever that is, 16-18? Probably 20 or less..idk, what do other ppl get?

[Revelate]

Reason codes are indeed listed in order of impact.

#1/2 are usually the big score drivers (like scorecard segmentation level though seeking credit can get up there on clean files) and 3/4 are often more minor... which is why having accounts with balances in the #2 slot above recent late even was such a shock on my TU FICO 4 score... and why I gained so many points (20) reducing that.

Unless there is substantial scorecard dependence on accounts with balances, and for the record I don’t think we have found that on literally any algorithm (different models behave differently but FICO 8 for example is the same for all scorecards AFAIK) which is why I am so confused at BBS’s non-movement in comparison on TU FICO 4: it is an outlier, wish there were a similarly pretty person that could confirm that one as I’m skeptical TBH.

[Anonymous]

Ok I get that they are listed in order of impact, but I just don't see 2 average ages having more impact than # of accounts on TU 04. I mean 15 years later all of a sudden length of revolving accounts becomes a factor in a '04 algorithm and reason statements appear.

Where have they been all these years that no one has noticed or thought to mention it? I think one of those is fluff, to borrow BBS's terminology. FICO 9? Sure. '08? Maybe. Industry Option? I don't doubt. But I think Fair Issac is keeping his secrets safe here by tossing a red herring. Maybe it was planned in R&D but not implemented until later, IDK...but I need more evidence first.

Hey, I may be totally wrong. Would not be the first time, probably not the last, but that's my take at the moment... and no, I haven't had much sleep , lol!

[Anonymous]

What is the normal spread btw all accounts with a balance and optimal on TU FICO 04? Anyone know? And IF it's scorecard dependent, at the time of testing BBS, you were clean/aged/thick/no new, right?

Also, who has had BoA, or one of these other CMS's that display the code in question, many years and is a credit vet? When was the first time anyone can remember seeing this negative reason code about oldest/average age of REVOLVING accounts?

I don't think this is something that has been around 15 years. And I don't recall any authorized updates to TU FICO 04's firmware, lol. Let's see if we can trach the genesis of this, because if this is a part of TU FICO 04, Why doesn't MF and all the other CMS's display this code?

Again, maybe I'm way off. I can be corrected. But one thing about it, we're gonna learn on the way, I promise ya that!

[Anonymous]

---

@Anonymous wrote:
What is the normal spread btw all accounts with a balance and optimal on TU FICO 04? Anyone know? And IF it's scorecard dependent, at the time of testing BBS, you were clean/aged/thick/no new, right?

Also, who has had BoA, or one of these other CMS's that display the code in question, many years and is a credit vet? When was the first time anyone can remember seeing this negative reason code about oldest/average age of REVOLVING accounts?

---

Didn't CGIDs recent thread that talked about the "revolving" accounts negative reason code find that it was limited only to TU, as no one was able to ever reference it on EX/EQ?  Maybe I'm recalling that wrong.

Yes, my file was clean/thick/aged and no new accounts at the time of testing; It was only a couple of months ago.  My oldest open revolving account is 4 years in age, just to throw that piece of data out there with my average age of revolving accounts landing probably in the 2-2.5 year range, I'm too lazy to figure it at the moment.  Youngest revolver, also my AoYA, something like 21-22 months.  The oldest closed revolver, also my AoOA, right around 18 years.

[Anonymous]

---

@Anonymous wrote:
I wanna see your reason codes from a MF 1B for TU. If it's there, you'll convince me.

---

I'm not a MF expert by any means, as I've only grabbed scores from them a couple of times ever, but aren't negative reason codes omitted from MF with scores above a certain point, say 800?  That being said, I don't believe I'd be able to see any negative reason statements associated with my TU4 score from them?  I'm a big fan of sources like my mortgage lender that still provide negative statements all the way up to likely a point off from a perfect score.

[Anonymous]

CGIDs thread found it all over 9, on all flavors of TU FICO 8, and only on the classic flavor of TU FICO 04. But that’s not what I’m getting at. I’m not questioning where it’s at, I’m questioning whether it is live in TU FICO 04 or just fluff.

If so, that was the first iteration of it. I do not believe that in the first iteration of it, it was live and yet we’ve never heard anything about it or any of its reason codes for 15 years until now. I just find that hard to believe. I may be wrong.

Yes you’re absolutely correct, MF does not show reason codes above 800. I don’t know what I was thinking. I haven’t had enough sleep the last few days to be honest BBS, my thought process is not at its best. And you got a point that is awesome they still provide reason codes despite the fact they are that high.

Another thing because of the buffer, I don’t think you could realize the full point spread anyway. I got a think about this some more BBS and get some sleep and I don’t know.

[gorgon]

Two more FICO8 data points for my Experian file: 

1) Didn't lose any points going from five to six inquiries from an app on May 16th. (Triple pull from Cap1 - Approved for SavorOne - New account has not reported yet)

2) Time since negative went from 11 months to 12 months.  +14 points and rating went from Fair to Good