cancel
Showing results for 
Search instead for 
Did you mean: 

Further Proof FICOs & FAKOs Have ZERO correlation

tag
Anonymous
Not applicable

Re: Further Proof FICOs & FAKOs Have ZERO correlation


@Anonymous wrote:

@Aahz wrote:

In this past week I paid off two cards with 0% Balance Transfers down to zero.  Paying off $11,000+ of my $30,000 in debt I was carrying and dropping utlity from 12% to 7%

 

EQ FICO8 742 -> 761

TU FICO8 752 -> 772

 

EQ Vantage (CK) 732 -> 703

TU Vantage (CK) 726 -> 703

 

Yes, my Credit Karma FAKOs took massive hits for paying off more than a third of my credit card debt.

No, there were no other changes.  I track my reportsactivity obsessively.


I had a similar issue with CK a few weeks ago.

 

I paid balances down to  less than $300 on one card, they were previously 1800 total spread over 5 out of 8 cards.

 

FaKos dropped TU 40 pts and EQ  30.   AND I had a paid judgement removed from Equifax.   

 

 

I think they penalize for paying cards completely off or down.  They must.   There is no other explanation

 

 


Yep! It's like they punish you for not staying in over your head in debt.  Maybe the 3 CRAs that create this model should stick to reporting - it's what they are good at... scoring is what Fico is good at.  JMHO.

Message 11 of 22
Anonymous
Not applicable

Re: Further Proof FICOs & FAKOs Have ZERO correlation

I find it odd that VS 3.0 would give someone a ding for dropping thier utilization from 12% to 7% or whatever.  Maybe they have different scoring thresholds, like how FICO has 1% to 9% as being the "optimal" range and 10% to 19% is the next best thing... maybe for some reason VS 3.0 reverses these.  We all know that FICO dings you good for reporting 0% aggregate utilization; maybe VS 3.0 dings you for anying from 0% to 9% perhaps suggesting that one isn't showing enough credit use?  I don't know, just thinking out loud here. 

 

Personally, I'd find it tough to listen to someone argue that someone showing 10% - 19% aggregate utilization is LESS risky than someone else with 1% - 9% aggregate utilization, especially with the monster total credit lines that people can obtain relatively easily these days.  On $150k in credit lines someone with 19% aggregate utilization verses 5% utilization is showing over $20k more in debt.  To me that represents more risk, not less.  But maybe VS 3.0 sees it opposite.

Message 12 of 22
Anonymous
Not applicable

Re: Further Proof FICOs & FAKOs Have ZERO correlation


@Anonymous wrote:
Thomas_Thumb, can you explain how removing incorrect derogatory info off of your credit report result in a 91 point drop in Vantage scoring with a high AAoA? Nothing else changed on my CRs. It just seems plain wrong to me and doesn't do a consumer like me any good to have a 91 point drop. If my lender used that score, I would be screwed! It feels like Vantage punished me because I improved my CRs.

I may have found the answer - rebucketing.  Still dont like Vantage though.  I have natural adversion to them. LOL. Thanks for you explanation though.

Message 13 of 22
Thomas_Thumb
Senior Contributor

Re: Further Proof FICOs & FAKOs Have ZERO correlation


@Anonymous wrote:

I find it odd that VS 3.0 would give someone a ding for dropping thier utilization from 12% to 7% or whatever.  Maybe they have different scoring thresholds, like how FICO has 1% to 9% as being the "optimal" range and 10% to 19% is the next best thing... maybe for some reason VS 3.0 reverses these.  We all know that FICO dings you good for reporting 0% aggregate utilization; maybe VS 3.0 dings you for anying from 0% to 9% perhaps suggesting that one isn't showing enough credit use?  I don't know, just thinking out loud here. 

 

Personally, I'd find it tough to listen to someone argue that someone showing 10% - 19% aggregate utilization is LESS risky than someone else with 1% - 9% aggregate utilization, especially with the monster total credit lines that people can obtain relatively easily these days.  On $150k in credit lines someone with 19% aggregate utilization verses 5% utilization is showing over $20k more in debt.  To me that represents more risk, not less.  But maybe VS 3.0 sees it opposite.


As with Fico, VantageScore 3.0 likes to see aggregate utilization in the single digit range. Monitoring of VS3 over the past couple years indicates optimal VS3 score for my file with all credit cards reporting is achieved when aggregate utilization is between 4.5% and 5.0%. I lose a few points at 4% and 7%. Drop increases when utilization reaches double digits.

 

Key point to remember is that VS counts AU accounts in full. For me and as I recall for you, an AU credit card was ignored by Fico 08. Perhaps you had some activity on your AU account that is ignored by Fico 08 but which VS3 reacted to. Additional note: Fico 08 does not consider charge cards in their aggregate utilization calculation. Not sure about VS3. See below table based on simulation results.

 

[All credit cards, including AU reported balances both times. AMEX charge did not have a balance. Side note: the VS3 simulator appears to use actual # cards reporting balances as a baseline for simulations. So, if # cards changed between pulls, simulation won't consider it]

 

Actual scores and simulations based on Transunion profile data. Interestingly actual prior (3% UT) and current month 9% UT) scores were 830 and 829 - which matches simulation predictions. Actual VS3 EQ scores were 829 both pulls.

 

VS3 graph.jpg

 

Ag UT%VS3 Score
1832
2831
3830
4829
5834
6833
7831
8830
9829
10827
11826
12824

 

P.S. Fico 08 is discriminatory regarding AU accounts. I suspect my AU account is ignored because there has been no activity under my name in 15 years. Hefty charges on the AU card pushed VS3 Ag UT to 9%. Without AU card, Ag UT is under 3% (which is what I would expect to see on MyFico - if I paid for an updated report.)

 

 

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 14 of 22
Anonymous
Not applicable

Re: Further Proof FICOs & FAKOs Have ZERO correlation

I agree with TT, I have tracked FICO and Vantage scores for 3 individuals for 2 years.  Myself, my mother and my wife.

 

They all correlate though the scores are sometimes very close for some individuals and for other individuals as much as 100 points different.

 

It's all because of AU accounts.  My wife has 7 AU accounts, one joint auto loan with me and 3 individual credit cards, while I have NO AU accounts and my mother has 2 joint credit cards with me and 2 individual retail store cards and one other bank card.

 

My wifes scores are quite different Vantage3 Vs FICO8.

 

Six months ago I was AU on my wifes' chase Freedom card and I removed myself and requested that Chase remove me from all three CRA.   My Vantage score went up 40 pionts and there was a high UTIL on that card, my FICO 8 stayed the same. The reason my Vantage score went up was because this was a new account and really lowerd my AAoA on Vantage but there was no effect on FICO 8.

 

My Util for all 3 accounts has varied quite a bit also, though it has never been zero on any CRA report.  UTIL varies as expected, it seem mostly related to aggregate UTIL and not an individual card. At times, I have gone to 89-95& on my own individual cards, though I think never above 40% aggregate.  Of course, this lowered my scores by about 100 points on all models. 

 

My wife is similar, though its more volitle on hers.

 

My mother I went to 66% UTIL on an individual card and about 30% aggregate, her scores went down about 50 points, but her AAoA is 13 years and inquires were zero at that point.

 

Some things that can cause confusion, different CRA catch date at different times,  last week EQ began tracking a new card of my mine while TU hasn't reported it yet, but i expect them to sooner or later.

 

A new AMEX card on my wifes account I don't seem to see it on any CRA, though I havent thoroughly checked, I tend to just look at the score unless I get curious.   I don;t see it on the reports when I looked a month ago, not sure why.

 

I keep a spreadsheet of different scores and they change, they correlate but it can take as much as 6 weeks for scores to move, so looking at them in one point of time it would appear they don't correlate, but smooth it out and they do, keeping in mind they are never the same.

 

BTW, I tend to not even notice changes of a few points, I look at 20+ points as significant.

 

Message 15 of 22
Anonymous
Not applicable

Re: Further Proof FICOs & FAKOs Have ZERO correlation

To the above, I forgot that my wife's Chase Freedom also had a high UTIL, so both how new the card was (lowered my AAoA) and it was charged up to 90% lowered my Vantage 3 score by 40 points.

 

However, no affect on Fico 8.

Message 16 of 22
Thomas_Thumb
Senior Contributor

Re: Further Proof FICOs & FAKOs Have ZERO correlation


@Anonymous wrote:

To the above, I forgot that my wife's Chase Freedom also had a high UTIL, so both how new the card was (lowered my AAoA) and it was charged up to 90% lowered my Vantage 3 score by 40 points.

 

However, no affect on Fico 8.


So, you were an AU on your wife's card - yes? VantageScore and the older Fico mortgage models always include AU cards - not so with Fico 08. Card utilizations of 90% and above are "max out" territory and impact on score under such conditions can be quite significant.

 

P.S. I am AU on my wife's card but, Fico 08 excludes it. Really not sure how a determination is made on inclusion/exclusion of AU accounts.

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 17 of 22
Aahz
Established Contributor

Re: Further Proof FICOs & FAKOs Have ZERO correlation


@Anonymous wrote:

 

They all correlate though the scores are sometimes very close for some individuals and for other individuals as much as 100 points different.

 

It's all because of AU accounts.  My wife has 7 AU accounts, one joint auto loan with me and 3 individual credit cards, while I have NO AU accounts and my mother has 2 joint credit cards with me and 2 individual retail store cards and one other bank card.

  


My original post disproves the first sentence I quoted.  If FICO8 and VS3 correlated in any way, then it would not be possible for my EQ scores to have a 10 point difference in one week and a 58 point difference the next week, particularly with the scores heading in opposite directions as they did.  

 

There is NO correlation between these two scores.  There is NO formula that can be applied to a FICO8 to determine a VS3 or vice-versa.

 

And, in my case, it has absolutely nothing to do with AU accounts as I am not an AU on any cards.

Message 18 of 22
Thomas_Thumb
Senior Contributor

Re: Further Proof FICOs & FAKOs Have ZERO correlation

Aahz -

 

I think the point you are making, which I agree with, is scores from one model can not be used to predict scores for the other model. However, in aggregate scores do move in the same direction but certainly not always on an individual basis. Really, there is not a true relationship between Fico 04 and Fico 08 either. I have seen some rather large score changes in EQ Fico 04 while EQ Fico 08 and EQ VS3 remained stable.

 

Score table 9-2016.jpg

 

It is virtually impossible to have a zero correlation betweem models. That would require data to be completely random in distribution - like a shot gun blast. The two models do show correlation. Taken as a whole those with higher Fico scores will have higher VantageScores based on the below scatter plot. If a correlation is above 0.8 (or below -0.8), the relationship is generally considered strongly correlated. The below shows a rather strong correlation particularly in the 20% to 70% score range.

 

correlation table.jpg

 

correlation graph.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 19 of 22
Aahz
Established Contributor

Re: Further Proof FICOs & FAKOs Have ZERO correlation


@Thomas_Thumb wrote:

Aahz -

 

I think the point you are making, which I agree with, is scores from one model can not be used to predict scores for the other model.


Yes, that is my sole point here.  I posted this thread as a nice, simple, very clear piece of evidence which can be pointed to when people try to say "My FAKO is XXX, so I assume my FICO is similar" or ask "What's my FICO if my FAKO is XXX?" (which far too many people do far too often).

 

I have no horse in the "which is better/more accurate" race.  Just happen to stumble into a near perfect example.

Message 20 of 22
Advertiser Disclosure: The offers that appear on this site are from third party advertisers from whom FICO receives compensation.