cancel
Showing results for 
Search instead for 
Did you mean: 

Breakpoints on number of cards w/balance?

tag
Revelate
Moderator Emeritus

Re: Breakpoints on number of cards w/balance?


@Thomas_Thumb wrote:

I had only an AU card and a charge card report one month. The AU card was not included in Fico 08 card count but, the charge card was included in count. Neither were included in utilization which was pegged at 0% although $1997 reported on AMEX and something like $560 on the AU card.

 

My Fico 08 bankcard and Auto scores took a hit at 0% UT but, Fico 08 classic held at 850 across the board. I posted the below previously as an attempt to look at Classic Fico 08 and Fico 09 buffers. My conclusion was Fico 08/Fico 09 appears to have 20 points worth of buffer on EQ, TU and 25 points worth on EX

 

Buffers.jpg

 

My wife controls the card I am AU on and strongly believes this forum is a waste of my time as credit is not an issue. She as I believe in PIF after balances report. She allows all charges to report naturally and that's the way it always has been and always will be. So, no chance of that not reporting.

 

As for my cards, I did have a month with only the AMEX charge reporting (no revolvers) - that was when Ag UT reported at 0%. Again, I am not a pre-pay before statement cuts advocate, and therefore may not ever be at zero cards with balances.

 

However, I am mildly curious if I could reach 817 on Classic EQ Fico 04 with only the AU card reporting.

 

 


Utilization is seperate from number of cards reporting a balance metric; whenever I'm done with the Discover testing I can try to repeat your test with just my Zync reporting and see if I get something at 0% but 1 card reporting a balance (Charge cards are absolutely factored there in my and other's experience).

 

What I'd REALLY like you to do for sanity's sake: pay everything to zero except the AU and see if you get a non-trivial drop.  If you don't, the AU counts, if you do, the AU does not.  That's the best way to determine whether the AU is being counted at all honestly, I can't think of any other way to test it personally... but the presentation anywhere doesn't matter, score movements and reason code shifts are really the only acceptable data analysis currently and that'll probably remain the same unless we get a non-trivial change in the market which is rather unlikely; while I'm hearing more muttering about the opaque algorithms which influence our lives the odds are they'll never be disclosed in the name of transparency.




        
Message 11 of 16
SouthJamaica
Mega Contributor

Re: Breakpoints on number of cards w/balance?


@Thomas_Thumb wrote:

I had only an AU card and a charge card report one month. The AU card was not included in Fico 08 card count but, the charge card was included in count. Neither were included in utilization which was pegged at 0% although $1997 reported on AMEX and something like $560 on the AU card.

 

My Fico 08 bankcard and Auto scores took a hit at 0% UT but, Fico 08 classic held at 850 across the board. I posted the below previously as an attempt to look at Classic Fico 08 and Fico 09 buffers. My conclusion was Fico 08/Fico 09 appears to have 20 points worth of buffer on EQ, TU and 25 points worth on EX

 

Buffers.jpg

 

My wife controls the card I am AU on and strongly believes this forum is a waste of my time as credit is not an issue. She as I believe in PIF after balances report. She allows all charges to report naturally and that's the way it always has been and always will be. So, no chance of that not reporting.

 

As for my cards, I did have a month with only the AMEX charge reporting (no revolvers) - that was when Ag UT reported at 0%. Again, I am not a pre-pay before statement cuts advocate, and therefore may not ever be at zero cards with balances.

 

However, I am mildly curious if I could reach 817 on Classic EQ Fico 04 with only the AU card reporting.

 

 


Doesn't she realize it's a better hobby than gambling or drinking?


Total revolving limits 741200 (620700 reporting) FICO 8: EQ 701 TU 704 EX 685

Message 12 of 16
Thomas_Thumb
Senior Contributor

Re: Breakpoints on number of cards w/balance?

Reason codes also supported the conclusion. Those were pasted in other posts but is included below as well.

 

Neither the AU card nor the AMEX are included in the UT calculation.

 

3B 8-2016 at 0%.gif

 

The two accounts with balances are: My open mortgage and my AMEX charge. AU cc  with balance not in count

 

8-6-2016 AT UT%.jpg

 

Note slight drop in EQ bankcard Fico 08 but major increase in EQ bankcard Fico 04. Why the drop in BC Fico 08? Because utilization went to 0%. Why the score increase in Fico 04? Because fewer cards are reporting balances AND because UT is not 0% for BC Fico 04 because the AU card balance is counting toward Ag UT.

Card count.jpg

Reason statements 8-2016 3B.jpg

 

 

Fico 9: .......EQ 850 TU 850 EX 850
Fico 8: .......EQ 850 TU 850 EX 850
Fico 4 .....:. EQ 809 TU 823 EX 830 EX Fico 98: 842
Fico 8 BC:. EQ 892 TU 900 EX 900
Fico 8 AU:. EQ 887 TU 897 EX 899
Fico 4 BC:. EQ 826 TU 858, EX Fico 98 BC: 870
Fico 4 AU:. EQ 831 TU 872, EX Fico 98 AU: 861
VS 3.0:...... EQ 835 TU 835 EX 835
CBIS: ........EQ LN Auto 940 EQ LN Home 870 TU Auto 902 TU Home 950
Message 13 of 16
Anonymous
Not applicable

Re: Breakpoints on number of cards w/balance?


@Thomas_Thumb wrote:

I have looked at impact of # cards reporting relative to EQ which appears to tweak the Fico models to be more sensitive to # (%) of cards reporting balances. See below table.

 

My profile is rather stable with all AAoA, oldest and youngest account thresholds having been exceeded years ago. I maintain Ag UT in the 1% to 6% range and don't believe UT% is influencing scores. The lone INQ may have an effect (note EQ 04 score difference with 3 cards reporting with/without INQ).

cards reporting updated.gif

 

Edit Correction: The EQ Bankcard scores are Fico 08, I had a mental lapse and originally listed them as Bankcard Fico 04. (my EQ Bankcard 04 scores actually swing - just like the EQ Classic 04 scores. I added the BC 04 scores to the left of the BC 08 scores).

 

For my profile, EQ Fico 04 score is affected by % of cards reporting once the 50% level is exceeded (greater than 50%).

a) 2 of 6 => 3 of 6, 0 point drop

b) 3 of 6 => 4 of 6, 5 point drop

c) 4 of 6 => 5 of 6, 17 point additional drop

d) 5 of 6 => 6 of 6, 23 (or 10) point additional drop ... - INQ likely influencing drop by up to 13 points

 - score 796 at 3 of 6 reporting with 1 INQ vs 809 at 3 of 6 reporting with 0 INQ

 

Side note AU card not counting in # reporting for Fico 08. However, this card faithfully reports a balance every month. So, for Fico 08 subtract one from # reporting and total count (example, Oct-16 would be 3 of 5)


Thank you for this post and such detailed data. For your profile, 50% appears to be the only breakpoint. What I'm really surprised at is the disparity between 0 & 1 INQ!

Message 14 of 16
Revelate
Moderator Emeritus

Re: Breakpoints on number of cards w/balance?

TT, why do you continue to state that the FICO consumer calculation of revolving utilization is accurate?

 

The odds that the front-end Consumer developers are the same ones that work on the FICO business algorithms are next to zero, like not even by infinite universe theory.  The odds that they actually know the algorithm any better than we do is also near zero, that is a closely guarded piece of IP, and I can tell you I've personally found numerous errors in the MF interface which were corrected upon informing them of it and their verifying it with their internal resources.  Why should this be any more reference?

 

Your conclusions cannot be supported by this or any other interface, and repeatedly we've asked you to simply test it by zeroing out every one of your accounts with the exception of the AU which it sounds like your wife leaves a balance on anyway... there's really no good excuse Smiley Sad.

 

You've stated a desire to figure things regarding the algorithm out, so then test them to completion; I simply don't understand your reluctance on this with something so easy to accomplish.




        
Message 15 of 16
Thomas_Thumb
Senior Contributor

Re: Breakpoints on number of cards w/balance?


@Anonymous wrote:

@Thomas_Thumb wrote:

I have looked at impact of # cards reporting relative to EQ which appears to tweak the Fico models to be more sensitive to # (%) of cards reporting balances. See below table.

 

My profile is rather stable with all AAoA, oldest and youngest account thresholds having been exceeded years ago. I maintain Ag UT in the 1% to 6% range and don't believe UT% is influencing scores. The lone INQ may have an effect (note EQ 04 score difference with 3 cards reporting with/without INQ).

cards reporting updated.gif

 

Edit Correction: The EQ Bankcard scores are Fico 08, I had a mental lapse and originally listed them as Bankcard Fico 04. (my EQ Bankcard 04 scores actually swing - just like the EQ Classic 04 scores. I added the BC 04 scores to the left of the BC 08 scores).

 

For my profile, EQ Fico 04 score is affected by % of cards reporting once the 50% level is exceeded (greater than 50%).

a) 2 of 6 => 3 of 6, 0 point drop

b) 3 of 6 => 4 of 6, 5 point drop

c) 4 of 6 => 5 of 6, 17 point additional drop

d) 5 of 6 => 6 of 6, 23 (or 10) point additional drop ... - INQ likely influencing drop by up to 13 points

 - score 796 at 3 of 6 reporting with 1 INQ vs 809 at 3 of 6 reporting with 0 INQ

 

Side note AU card not counting in # reporting for Fico 08. However, this card faithfully reports a balance every month. So, for Fico 08 subtract one from # reporting and total count (example, Oct-16 would be 3 of 5)


Thank you for this post and such detailed data. For your profile, 50% appears to be the only breakpoint. What I'm really surprised at is the disparity between 0 & 1 INQ!


As a follow-up to the above I wanted to look at TU scores relative to # accounts reporting balances. The TU mortgage score is derived from Fico 04 same as EQ. In the case of TU, I have no inquiries on file so that factor is not a consideration.

 

Here is the TU comparison table:

Cards reporting TU.gif

Observation: [ TU Fico 04 much less sensitive to # cards reporting than EQ Fico 04]

 -TU Fico 04 dropped 12 points between 2 of 6 and 6 of 6 cards reporting.

 - By comparison EQ dropped 45 points (perhaps 32 points adjusting for inquiry at 6 of 6)

 2) Able to report balances on 2, 3 and 4 cards without impacting Fico 04 score (in certain cases)

 3) Reporting balance on 6 of 6 cards negatively impacted all Fico scores except Classic 08.

 4) Unable to explain drop in Classic Fico 04 with only 2 cards reporting on two occassions

 5) Lowest scores ever on Fico 08 (Classic and Bankcard) in 10/2016 ... but highest Fico 04 scores.

 

Drop in TU Fico 08 scores also seen on Discover card.

However, no drop in EX scores as reported on Discover creditscorecard and AMEX.

 

Discover TU scores.jpg

AMEX Fico scores.jpg

 

EQ Fico 08 bankcard score (as reported through Citi) does not show any significant drop in October. However, the reason statements are rather interesting. Could Fico 08 have the ability to factor in prior month # accounts with balances against current month # accounts where payments were made? (if prior # with balances > current month # payments, then risk). Fico 04 certainly does not to that.

 

Also worth noting: EQ Bankcard Fico 08 was at a low of 876 on an 8/6 3B report with 2 cards (AU and AMEX charge) showing balances but, at an all time high of 892 on 8/23 when AMEX had been paid off and a single non AU credit card had posted a balance.

 

Citi EQ Fico date stamp.jpg

 

Additional information, question:

 1) AU card had a "high" balance report on October statement. This resulted in a 8.45% Ag UT% in October based on CK's summary date. 

 2) The AU card balance was paid 5 days after the DUE date. The payment was well before 30 days so, it won't show on credit reports as a 30 day late.

 3) Is it possible TU has tweaked Fico 08 to factor in # accounts with balances current month against # accounts with balance prior month or # accounts showing payments current month against # accounts with balances prior month? If so the late payment (although less than 30 days) would show fewer accounts having a payment vs accounts with having had a balance - which could be a risk indicator.

 

CK reported UT% (includes AU card).jpg

Fico 9: .......EQ 850 TU 850 EX 850
Fico 8: .......EQ 850 TU 850 EX 850
Fico 4 .....:. EQ 809 TU 823 EX 830 EX Fico 98: 842
Fico 8 BC:. EQ 892 TU 900 EX 900
Fico 8 AU:. EQ 887 TU 897 EX 899
Fico 4 BC:. EQ 826 TU 858, EX Fico 98 BC: 870
Fico 4 AU:. EQ 831 TU 872, EX Fico 98 AU: 861
VS 3.0:...... EQ 835 TU 835 EX 835
CBIS: ........EQ LN Auto 940 EQ LN Home 870 TU Auto 902 TU Home 950
Message 16 of 16
Advertiser Disclosure: The offers that appear on this site are from third party advertisers from whom FICO receives compensation.