topic Re: VS 3.0 AZEO vs AZE2 & tiny utilization? in Understanding FICO® Scoring

VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Tue, 12 Feb 2019 15:36:11 GMT

This is probably a great question for TT that understands the VS 3.0 algorithm well, or anyone else in the know.

I find my VS 3.0 scores seem to fluctuate +13 points, -13 points rather often, seemingly always due to reported balance changes; 9 times out of 10 I see 823 --> 836 and then 836 --> 823.

For the last handful of months when experiencing this 13-point yo-yo, I'm always at AZEO or AZE2 (never AZ). The reported balance on the AZEO card or AZE2 cards is always small, (say) $5-$300 or so and always in the 1%-2% utilization range. For example, currently I'm at AZEO and that AZEO card has a $5 reported balance on a $10k limit. While my FICO scores are currently maximized with my profile in this state, my VS 3.0 scores are at the -13 (823) mark.

I guess my question here is what factor is it exactly from the VS 3.0 lens that's causing the 13 point drop? I'm thinking it's one of two things:

1 - A balance that's too small (like .0005%) is "ignored" by VS 3.0 where it's rounded up and "seen" as 1% by FICO algorithms.

2 - VS 3.0 prefers AZE2 to AZEO; maybe the times I've been at the +13 (836) score I've had 2 reported non-zero balances and haven't noticed.

Any insight here is appreciated.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Tue, 12 Feb 2019 17:42:16 GMT

Hi BBS. It would be interesting if you could execute a number of test scenarios. For example, comparing....

(1) Extremely low utilization. E.g. exactly one card with a $5 balance with that card having a 10k limit.

(2) Very low but not crazy low. E.g. total CC debt > $101 and total utilization at 0.52% and individual utilization at > 1.01%. Still AZEO as with scenario #1 above.

If #2 gave regularly gave you a significantly better score, then it would be strong evidence in favor of the rounding hypothesis (or the hypothesis that ultra tiny dollar values affect the score, which could be true apart from utilization percentages).

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Trudy — Tue, 12 Feb 2019 18:24:22 GMT

I find this interesting because my scores only fluctuates 2 -3 points regularly. When I report 1% UTL or less (always rounded to 1%) my score rises to 829 or 830, regardless of number of accounts reporting, never more than 5/12 (3cc & 2 loans), usually 4 reporting. Anytime I hit 2% UTL (aggregate) my score drops to 827. Never any single card over 18%. Although, since monitoring my VS3 scores and my activity this is the 1st time I had a single card report as high as 18%. But I've seen this drop previously with less than 9% on a single card.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Tue, 12 Feb 2019 20:40:01 GMT

I paid the AZEO 5 buck balance to zero today and it will report in a couple of days. A few days after that I've got a couple of other cards reporting, so I'll try and get one to report in the .51-1% utilization range and see what happens. If that doesn't work, I'll go for 1.01%+ the next time.

I've never really looked into my VS 3.0 scores and what impacts them, so this will be a fun little test.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Wed, 13 Feb 2019 09:49:39 GMT

I know we're in very different bins/scorecards, but maybe this data I collected will help.

All cards with a balance was a positive change for me on VS3. Maybe it's the low utilization as well.

(That TU VS3 score actually skyrocketed +84 points to 713 with the SSL at 8.47%, before the loan reported closed and dropped 41 points.)

		mF 1B CK	Δ	CCT 3B	Δ	mF 3B CK	Δ	mF 3B CK CCOM	Δ	mF 3B CK
			Loan utilization changed from 17% to 8.47% 1 new inq at TU on 12/20, now 2 under 1yr		1 new inq at EX/EQ on 12/23 after the CCT 3B report		Loan closed! 2 new revolving accounts 1 of 2 cards reporting 2% util with 1% total util.		Both cards (100%) reporting 7% util with 7% total util. 1 inq aged past 1 year on TU The -3 on EX F8 happened before any new statement balances reported.
		12/04	Δ	12/23	Δ	12/31	Δ	01/18	Δ	02/08
VS3	EQ	658			-17	641	+35	676	+9	685
VS3	TU	634			-5 (1 inq)	629	+43	672	+8	680
VS3	EX							676	credit.com update Feb 14	?
F8	EQ			750	-20	730	-46	684	-3	681
F8	TU	666	+62	728		728	-39	689	+6	695
F8	EX			748	-19	729	-30	699	-3	696

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Tue, 26 Feb 2019 15:06:23 GMT

@Anonymous wrote:
Hi BBS. It would be interesting if you could execute a number of test scenarios. For example, comparing....

(1) Extremely low utilization. E.g. exactly one card with a $5 balance with that card having a 10k limit.

(2) Very low but not crazy low. E.g. total CC debt > $101 and total utilization at 0.52% and individual utilization at > 1.01%. Still AZEO as with scenario #1 above.

If #2 gave regularly gave you a significantly better score, then it would be strong evidence in favor of the rounding hypothesis (or the hypothesis that ultra tiny dollar values affect the score, which could be true apart from utilization percentages).

Best I can tell VS3 does not have a scoring attribute for # cards with balances. Also, not sure it has an all cards at zero balance penalty either. As I recall EW800 almost always reports AZEO with a token balance on one card and has VS3 scores of 839. So, an ultra low balance is not likely a root cause for your score drop.

I do think it would be insightful for you to use the VS3 simulator - particularly now that your score is lower. Increment your total balance up and record what the simulator predicts. I have plotted this a couple times and see a local spike up with my profile in the 4.0% to 5.5% UT range. VS3 does have multiple scorecards and undoubetedly each scorecard treats attributes a bit differently. Not sure how profiles are treated by scorecard.

It appears probable that something else is triggering the 13 points up/down shift. At one time I posted a link to all VS reason codes - there are quite a few. A review of their reason codes could add insight. I can't find the link but you can likely still find the list through a search.

VantageScore does score inquiries for a full two years and appears to discount derogs earlier than Fico based on EW800's TU recent Fico and VantageScore data. Also, VantageScore always considers AU cards and appears to look at charge cards as well. Below are the primary scoring categories for VS3.

...

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Wed, 13 Feb 2019 15:03:28 GMT

Here is some other info on VantageScore. Interesting comments on charge cards and loans.

Note the differentiation in loans starting with VS2.

I have one loan, a 15 year mortgate on file. Back in 2015 my installment B/L was between 50% and 60%. Below are some of my negative reason statements relative to VS3 (from a 3B TU report). Clearly VS is looking for more loans at a lower B/L ratio. Had one HP on EQ - over 12 months but less than 24 months age. VS3 scores were in the low 830s.

...

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Wed, 13 Feb 2019 15:44:11 GMT

@Thomas_Thumb wrote:
Best I can tell VS3 ndoes not care have a scoring attribute for # cards with balances. Also, not sure it has an all cards at zero balance penalty either. As I recall EW800 almost always reports AZEO with a token balance on one card and has VS3 scores of 839. So, an ultra low balance is not likely a root cause for your score drop.

I do think it would be insightful for you to use the VS3 simulator - particularly now that your score is lower. Increment your total balance up and record what the simulator predicts. I have plotted this a couple times and see a local spike up with my profile in the 4.0% to 5.5% UT range. VS3 does have multiple scorecards and undoubetedly each scorecard treats attributes a bit differently. Not sure how profiles are treated by scorecard.

In using the simulator, it shows the breakpoint being a $108 balance. A $107 balance (or less) shows me the lower score, a $108 balance up to $423 shows me the +13 score. $424 results in a 1 point drop, then there are additional 1-2 point drops at different points higher than that. So, the simulator is telling me that having my aggregate utilization at .00058% - .00227% is ideal. Going just under that range, .00057%, results in a 13 point drop.

I know in the past TT you've suggested that 5% or so aggregate utilization may be ideal for VS 3.0, but personally my aggregate has never been that high (usually .1%-2%) when I've seen my top 836-837 scores.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Wed, 13 Feb 2019 22:45:24 GMT

@Anonymous wrote:
@Thomas_Thumb wrote:
Best I can tell VS3 ndoes not care have a scoring attribute for # cards with balances. Also, not sure it has an all cards at zero balance penalty either. As I recall EW800 almost always reports AZEO with a token balance on one card and has VS3 scores of 839. So, an ultra low balance is not likely a root cause for your score drop.

I do think it would be insightful for you to use the VS3 simulator - particularly now that your score is lower. Increment your total balance up and record what the simulator predicts. I have plotted this a couple times and see a local spike up with my profile in the 4.0% to 5.5% UT range. VS3 does have multiple scorecards and undoubetedly each scorecard treats attributes a bit differently. Not sure how profiles are treated by scorecard.

In using the simulator, it shows the breakpoint being a $108 balance. A $107 balance (or less) shows me the lower score, a $108 balance up to $423 shows me the +13 score. $424 results in a 1 point drop, then there are additional 1-2 point drops at different points higher than that. So, the simulator is telling me that having my aggregate utilization at .00058% - .00227% is ideal. Going just under that range, .00057%, results in a 13 point drop.

I know in the past TT you've suggested that 5% or so aggregate utilization may be ideal for VS 3.0, but personally my aggregate has never been that high (usually .1%-2%) when I've seen my top 836-837 scores.

It looks like you found an anomoly at a very low utilization point. I did not pick up the very low UT% spike in my below test - but that was not a focal point in my simultion. I've never been down in the 0.1% range on UT because I only pay balances after statements cut. I'm typically in the 1.0% to 5% range but, been as high as 8.5% since tracking.

Although you never report in the 5% range I'd be interested in seeing if the simulator gives you a score spike in that 4% to 5.5% range. It's a narrow band so you would need to check at 0.2% increments. I'd like to see if someone else sees this spike. It's also possible the spike anomoly is influenced by balance history. Six months from now you may see a spike UT% shift just as I did when comparing against a simulation from a couple years ago.

BTW - It does appear the UT% simulation result has some correlation to real world results as you have seen.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Wed, 13 Feb 2019 22:19:45 GMT

I'll go through tonight and throw a bunch of different numbers into the simulator and report back.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Thu, 14 Feb 2019 17:21:14 GMT

Below are the actual simulation data points for the above graph. Steady at 829 between 0.0% and 0.8%. Thus, my earlier comment on no negative impact with zero balance. Peak for this simulation (1/2019) appears to be around 5.5%.

Ag Utilization	VS 3 Score	Ag Utilization	VS 3 Score
0.00%	829	10.55%	824
0.03%	829	10.92%	823
0.40%	829	11.65%	822
0.77%	829	12.21%	821
0.96%	827	13.13%	820
2.07%	826	13.50%	819
2.63%	825	14.62%	817
3.70%	824	15.54%	816
4.44%	823	16.28%	814
5.04%	822	17.39%	812
5.74%	831	18.32%	810
6.11%	830	19.43%	808
6.66%	830	20.37%	806
7.26%	829	23.31%	800
8.33%	827	24.98%	796
9.25%	826	27.76%	789
9.62%	825	30.91%	783

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Thu, 14 Feb 2019 02:39:58 GMT

@Thomas_Thumb wrote:
VS3 does have multiple scorecards and undoubetedly each scorecard treats attributes a bit differently. Not sure how profiles are treated by scorecard.

That's the whole problem with figuring out these various scoring models - they all have a variable number of bins, which is standard for every logistic regression model. In data analytics, it's generally recommended to have between 10 and 20 bins to get the best predictors (information values). But every so often, the model (FICO/VS3) gets updated with new data (from CRA's) and the number of bins/scorecards/buckets can change out from under anyone trying to figure out exact point gains/losses.

You can even run your own 'mini-FICO simulation' using a programming environment I work in a lot called RStudio: https://www.listendata.com/2015/03/weight-of-evidence-woe-and-information.html

It appears probable that something else is triggering the 13 points up/down shift. At one time I posted a link to all VS reason codes - there are quite a few. A review of their reason codes could add insight. I can't find the link but you can likely still find the list through a search.

Here's a list of all the Vantage Score 3.0 reason codes I could find, in a nice HTML table rather than a PDF.

Taken from: http://tenantdata.com/downloads.html (An Experian PDF dated 2013)

Also from: https://www.reasoncode.org/ (The 'All Codes' link at the bottom. )

--Code---	Vantage Score 3.0 Reason Code / Text Table
4	The balances on your accounts are too high compared to loan amounts
5	Too many of the delinquencies on your accounts are recent
6	You have too many accounts that were opened recently
7	You have too many delinquent or derogatory accounts
8	You have either too few loans or too many loans with recent delinquencies
9	The worst payment status on your accounts is delinquent or derogatory
10	You have either very few loans or too many loans with delinquencies
11	The total of your delinquent or derogatory account balances is too high
12	The date that you opened your oldest account is too recent
13	Your most recently opened account is too new
14	Lack of sufficient credit history
15	Newest delinquent/derogatory payment status on your accts is too recent
16	The total of all balances on your open accounts is too high
17	Balance on previously delinquent accts are too high compared to loan amts
18	Total of balances on accts never late is too high compared to loan amts
21	No open accounts in your credit file
22	No recently reported account information
23	Lack of sufficient relevant account information
29	Too many of your open bankcard or revolving accounts have a balance
30	Too few of your bankcard or other revolving accounts have high limits
31	Too many bankcard or other revolving accounts were opened recently
32	Balances on bankcard/revolving accts too high compared to credit limits
33	Your worst bankcard or revolving account status is delinquent/derogatory
34	Total of all balances on bankcard or revolving accounts is too high
35	Your highest bankcard or revolving account balance is too high
36	Your largest credit limit on open bankcard or revolving accts is too low
39	Available credit on your open bankcard or revolving accounts is too low
40	The date you opened your oldest bankcard or revolving acct is too recent
42	The date you opened your newest bankcard or revolving acct is too recent
43	Lack of sufficient credit history on bankcard or revolving accounts
44	Too many bankcard or revolving accounts with delinquent/derogatory status
45	Total balances too high on delinquent/derogatory bankcard/revolving accts
47	No open bankcard or revolving accounts in your credit file
48	No bankcard or revolving recently reported account information
49	Lack of sufficient relevant bankcard or revolving account information
53	The worst status on your real estate accounts is delinquent or derogatory
54	The amt of balance paid down on your open real estate accounts is too low
55	Open real estate acct balances are too high compared to their loan amts
56	You've been playing reindeer games on myFICO forums.
57	Too many real estate accts with delinquent or derogatory payment status
58	The total of all balances on your open real estate accounts is too high
61	No open real estate accounts in your credit file
62	No recently reported real estate account information
63	Lack of sufficient relevant real estate account information
64	No open first mortgage accounts in your credit file
65	Lack of sufficient relevant first mortgage account information
66	Your open auto account balances are too high compared to their loan amts
68	No open auto accounts in your credit file
69	Lack of sufficient relevant auto account information
71	You have either very few installment loans or too many with delinquencies
72	Too many installment accts with a delinquent or derogatory payment status
73	The worst status on your installment accounts is delinquent or derogatory
74	The balance amount paid down on your open installment accounts is too low
75	The installment account that you opened most recently is too new
76	You have insufficient credit history on installment loans
77	The balances on your accounts are too high compared to loan amounts
78	Balances on installment accts are too high compared to their loan amounts
79	Too many of the delinquencies on your installment accounts are recent
81	No open installment accounts in your credit file
83	Lack of sufficient relevant installment account information
84	The number of inquiries was also a factor, but effect was not significant
85	You have too many inquiries on your credit report.
86	Your credit report contains too many derogatory public records
87	Your credit report contains too many unsatisfied public records
88	One or more derogatory public records in your credit file is too recent
90	Too few discharged bankruptcies
93	The worst status on your student loan accts is delinquent or derogatory
94	The balance amount paid down on your open student loan accts is too low
95	You have too many collection agency accounts that are unpaid
96	The total you owe on collection agency accounts is high
97	You have too few credit accounts
98	There is a bankruptcy on your credit report
	Exclusion Scores
0001	File/trade indicating consumer is deceased
0004	No information on credit report (other than maybe inquiries)
9000	System Exclusion

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Thu, 14 Feb 2019 03:17:21 GMT

TT, I used the simulator to come up with a data set similar to what you provided in Post 11, except I stopped just above 10% aggregate utilization. I did not see any projected score spike the way you did in the 5%-6% or so range as you can see from my numbers:

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Mon, 25 Feb 2019 22:36:46 GMT

So I went from a $5 reported balance on a single revolver with a $10k limit to a $297 reported balance on a single revolver with a $10k limit and my score moved back from 824 --> 836. I would say that this is pretty strong evidence that a very small AZEO balance doesn't maximize VS 3.0 scores. I think I read on here that there's no penalty for all $0 balances on revolvers with VS 3.0. Does that mean no penalty verses a tiny (less than ideal) balance, or no penalty from a small but not trivial balance that is ideal from a VS 3.0 standpoint?

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Tue, 26 Feb 2019 00:12:39 GMT

VantageScore has a comprehensive list of negative reason statements. I don't recall seeing a "no recent activity on revolving accounts" penalty/reason code like what Fico has published (I'll recheck). Not sure what triggers the VS3 spike up but, location of the spike does appear to shift.

The below simulation showed a constant score from absolute zero up to about 0.8% followed by steady decline up to just over 5%. Then there is an abrupt spike up followed by second steady decline. In the below simulation, the spike up was clearly not associated with going from an all zero penalty to a balance reporting.

Location of the spike appears to be profile/scorecard dependent. Could prior utilization/balance affect how current utilization/balance is evaluated?

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Tue, 26 Feb 2019 00:34:10 GMT

List of VS3 negative reason codes (same as posted above - with individual links)

VantageScore 3.0 codesv3

04: The balances on your accounts are too high compared to loan amounts

05: Too many of the delinquencies on your accounts are recent

06: You have too many accounts that were opened recently

07: You have too many delinquent or derogatory accounts

08: You have either very few loans or too many loans with recent delinquencies

09: The worst payment status on your accounts is delinquent or derogatory

10: You have either very few loans or too many loans with delinquencies

11: The total of your delinquent or derogatory account balances is too high

12: The date that you opened your oldest account is too recent

13: Your most recently opened account is too new

14: Lack of sufficient credit history

15: Newest delinquent or derogatory payment status on your accounts is too recent

16: The total of all balances on your open accounts is too high

17: Balances on previously delinquent accounts are too high compared to loan amts

18: Total of balances on accounts never late is too high compared to loan amounts

21: No open accounts in your credit file

22: No recently reported account information

23: Lack of sufficient relevant account information

29: Too many of your open bankcard or revolving accounts have a balance

30: Too few of your bankcard or other revolving accounts have high limits

31: Too many bankcard or other revolving accounts were opened recently

32: Balances on bankcard or revolving accounts too high compared to credit limits

33: Your worst bankcard or revolving account status is delinquent or derogatory

34: Total of all balances on bankcard or revolving accounts is too high

35: Your highest bankcard or revolving account balance is too high

36: Your largest credit limit on open bankcard or revolving accounts is too low

39: Available credit on your open bankcard or revolving accounts is too low

40: The date you opened your oldest bankcard or revolving account is too recent

42: The date you opened your newest bankcard or revolving account is too recent

43: Lack of sufficient credit history on bankcard or revolving accounts

44: Too many bankcard or revolving accounts with delinquent or derogatory status

45: Total balances too high on delinquent/derogatory bankcard or revolving accts

47: No open bankcard or revolving accounts in your credit file

48: No bankcard or revolving recently reported account information

49: Lack of sufficient relevant bankcard or revolving account information

53: The worst status on your real estate accounts is delinquent or derogatory

54: The amount of balance paid down on your open real estate accounts is too low

55: Open real estate account balances are too high compared to their loan amounts

57: Too many real estate accounts with delinquent or derogatory payment status

58: The total of all balances on your open real estate accounts is too high

61: No open real estate accounts in your credit file

62: No recently reported real estate account information

63: Lack of sufficient relevant real estate account information

64: No open first mortgage accounts in your credit file

65: Lack of sufficient relevant first mortgage account information

66: Your open auto account balances are too high compared to their loan amounts

68: No open auto accounts in your credit file

69: Lack of sufficient relevant auto account information

71: You have either very few installment loans or too many with delinquencies

72: Too many installment accounts with a delinquent or derogatory payment status

73: The worst status on your installment accounts is delinquent or derogatory

74: The balance amount paid down on your open installment accounts is too low

75: The installment account that you opened most recently is too new

76: You have insufficient credit history on installment loans

77: Newest delinquent or derogatory status on installment accounts is too recent

78: Balances on installment accounts are too high compared to their loan amounts

79: Too many of the delinquencies on your installment accounts are recent

81: No open installment accounts in your credit file

83: Lack of sufficient relevant installment account information

84: The number of inquiries was also a factor, but effect was not significant

85: You have too many inquiries on your credit report.

86: Your credit report contains too many derogatory public records

87: Your credit report contains too many unsatisfied public records

88: One or more derogatory public records in your credit file is too recent

90: Too few discharged bankruptcies

93: The worst status on your student loan accounts is delinquent or derogatory

94: The balance amount paid down on your open student loan accounts is too low

95: You have too many collection agency accounts that are unpaid

96: The total you owe on collection agency accounts is high

97: You have too few credit accounts

98: There is a bankruptcy on your credit report

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Tue, 26 Feb 2019 02:37:43 GMT

Interesting that your simulation and mine are different, TT. I would think that you and I would possess similar scorecards and thus see similar simulations. You having a spike around 5% is dramatically different than mine around .25%. I know nothing of the different VS 3.0 scorecards though, so perhaps ours are more different than I'm assuming.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Thomas_Thumb — Tue, 26 Feb 2019 15:05:19 GMT

Perhaps past balance/utilization influences peak location. You typically report AZEO with a trivial balance. I let all cards in use report full monthly charges and then just pay statement balances. Typically I have 4 cards reporting monthly balances.

It would be interesting to see how NRB525's VS3 scores look on a simulation. He reports balances on multiple cards and some of those balances are nontrivial.

Re: VS 3.0 AZEO vs AZE2 & tiny utilization?

Anonymous — Tue, 26 Feb 2019 16:34:58 GMT

@Thomas_Thumb wrote:
Perhaps past balance/utilization influences peak location. You typically report AZEO with a trivial balance. I let all cards in use report full monthly charges and then just pay statement balances. Typically I have 4 cards reporting monthly balances.

It would be interesting to see how NRB525's VS3 scores look on a simulation. He reports balances on multiple cards and some of those balances are nontrivial.

Agreed that would be good to hear. I wouldn't say that I "typically" report AZEO with a trivial balance though. I have a bunch of times for testing purposes and as you know recently I did in order to grab my 28 FICO scores. I've also done the converse, where I did that test last year in taking all of my 8 revolvers to a reported balance in order to see the FICO score impact. I would say it's more typical that 2-3 of my cards have a reported balance at any given time and usually they're nontrivial with my Citi card (daily driver) showing a reported balance in the $2000-$3000 range and 1-2 other cards with a few hundred bucks reported.