topic Re: Score Increased When #Accounts Reporting Bal Dropped in Understanding FICO® Scoring

Score Increased When #Accounts Reporting Bal Dropped

Iusedtolurk — Thu, 11 Mar 2021 13:48:36 GMT

I always look at all the data points when anything happens so I can look for those changes in future similar circumstances.

Today I was checking my EX and noticed my score had went up +9 points. (always a good feeling to see a score increase)

The only differences between yesterday and today:

Today aggregate 20.0%/dollars owed 4472

Yesterday aggregate 20.6%/dollars owed 4590

Difference -118

Today # of accounts with Balances 11/12

Yesterday # of accounts with Balances 12/12

As the amount of dollars owed and aggregate utilization is practically the same I can come to the conclusion that going from 12/12 100% Accounts reporting with a balance to 11/12 91% was cause for the +9 point increase.

Re: Score Increased When #Accounts Reporting Bal Dropped

dragontears — Thu, 11 Mar 2021 14:12:35 GMT

Just to be clear, when you say "accounts with balances" is that just revolving accounts or does it include installment accounts?

Re: Score Increased When #Accounts Reporting Bal Dropped

Iusedtolurk — Thu, 11 Mar 2021 15:46:28 GMT

@dragontears wrote:
Just to be clear, when you say "accounts with balances" is that just revolving accounts or does it include installment accounts?

@dragontearsAll revolvers (12) no current open installments.

Re: Score Increased When #Accounts Reporting Bal Dropped

Curious_George2 — Thu, 11 Mar 2021 16:01:21 GMT

Thanks for the DP and congrats on the 9 pt gain! I have a few questions.

Is this EX8?

We have some indications FICO uses open and closed accounts as the denominator for calculating the percent of accounts with a balance. You said you have no open loans, but how many accounts are on your report, including open and closed loans and cards?

Even though your total dollars owed on cards didn't change by a lot, there's actually a chance it crossed a threshold. @Anonymous has been hunting for a threshold in that vicinity. I think his current theory is that for TU8 it's between $4385 and $4684. Maybe you just found approximately where it is on EX8 (I don't know if it's necessarily the same across bureaus).

Re: Score Increased When #Accounts Reporting Bal Dropped

Iusedtolurk — Thu, 11 Mar 2021 16:32:17 GMT

@Curious_George2 wrote:
Thanks for the DP and congrats on the 9 pt gain! I have a few questions.

Is this EX8?

We have some indications FICO uses open and closed accounts as the denominator for calculating the percent of accounts with a balance. You said you have no open loans, but how many accounts are on your report, including open and closed loans and cards?

Even though your total dollars owed on cards didn't change by a lot, there's actually a chance it crossed a threshold. @Anonymous has been hunting for a threshold in that vicinity. I think his current theory is that for TU8 it's between $4385 and $4684. Maybe you just found approximately where it is on EX8 (I don't know if it's necessarily the same across bureaus).

@Curious_George2 I have a total of 12 open revolvers and 3 closed installments (unfortunately all CFA).

Being as how @Anonymous was trying to nail down a possible dollar amount between perhaps $4385 and $4684 my change from $4590 to $4472 would still fall inbetween those dollar amounts.4385<4472<4590<4684.

I believe tomorrow 2 other accounts should report reducing the 4472 to 4419 still inbetween Birdmans dollar amounts but also I should be showing 9/12 accounts with balances. Curious to see if any affects will be seen.

I've read about thresholds of 1/3 or less than 1/2 of accounts reporting balances having a point effect but I just have a feeling that a person may get penalized pointwise having all accounts reporting a balance. I really don't expect a point change tomorrow once I show 9/12 but it will be interesting to look forward to.

Edit: Forgot to include...Yes this is EX8

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Fri, 12 Mar 2021 03:29:29 GMT

@Iusedtolurk wrote:
I always look at all the data points when anything happens so I can look for those changes in future similar circumstances.

Today I was checking my EX and noticed my score had went up +9 points. (always a good feeling to see a score increase)

The only differences between yesterday and today:

Today aggregate 20.0%/dollars owed 4472
Yesterday      aggregate 20.6%/dollars owed 4590
Difference     -118

Today            # of accounts with Balances 11/12
Yesterday   # of accounts with Balances 12/12

I have always found 100% of accounts reporting a balance to cause quite wild swings across most FICO score models.

I just switched from 4-of-4 to 3-of-4 this month and saw EX8 +6pts. +2 from aging to AAoA 2yr0mo on the 1st, followed by +6 with the balance change for a total of +8pts on EX8 this month.

Even with only 2 cards on my file, I could cause a large drop on several scores just by having them both report a balance, always under 9% on everything. Early on in my first year of revolving history, the difference was as much as EQ8 19pts/TU8 14pts.

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Fri, 12 Mar 2021 03:35:15 GMT

OP, do you know what scorecard you're on?

On some scorecards such as clean/thick/aged some have reported no score changes at all on EX8 related to number of accounts with a balance. I'm one of those people. Whether I'm at AZEO, all accounts with a balance or anywhere in between my EX8 score doesn't budge, where my TU8 and EQ8 do fluctuate.

Re: Score Increased When #Accounts Reporting Bal Dropped

Iusedtolurk — Fri, 12 Mar 2021 03:53:33 GMT

@Anonymous wrote:
OP, do you know what scorecard you're on?

On some scorecards such as clean/thick/aged some have reported no score changes at all on EX8 related to number of accounts with a balance. I'm one of those people. Whether I'm at AZEO, all accounts with a balance or anywhere in between my EX8 score doesn't budge, where my TU8 and EQ8 do fluctuate.

@Anonymous clean/thick/(young) mature/no new accounts

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Fri, 12 Mar 2021 03:56:26 GMT

Gotcha. The young mature seems to be the cause of strange/uncharacteristic things lately in some circumstances.

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Sat, 13 Mar 2021 07:27:27 GMT

@Iusedtolurk You know that is a good thing you did by putting the qualifier young mature in there because it apparently does make quite a difference. Even though you're on a mature card, there's a big difference in whether or not your credit history is say under 10 years or beyond 10 years.

And I'm just throwing that number out there I don't know what the real number is that makes a big difference, but we do see quite a lot of gains I believe from age of oldest Revolver as it makes its way towards 10 years.

I would suggest for future reference when we ask for scorecard we also ask for age of oldest account so we can determine how mature their profile is.

and yes, there does seem to be a score change at 100% of accounts open with a balance, however we have found for EX2 it includes closed accounts. And I have found Thresholds There and I'm in the middle of a test and I will post my results when I get done.

Who knows if that applies to the other versions. And who knows if that was conflated and possibly included a balance Threshold?

Re: Score Increased When #Accounts Reporting Bal Dropped

Iusedtolurk — Sat, 13 Mar 2021 10:11:58 GMT

@Iusedtolurk wrote:
I always look at all the data points when anything happens so I can look for those changes in future similar circumstances.

Today I was checking my EX and noticed my score had went up +9 points. (always a good feeling to see a score increase)

The only differences between yesterday and today:

Today aggregate 20.0%/dollars owed 4472
Yesterday      aggregate 20.6%/dollars owed 4590
Difference     -118

Today            # of accounts with Balances 11/12
Yesterday   # of accounts with Balances 12/12

As the amount of dollars owed and aggregate utilization is practically the same I can come to the conclusion that going from 12/12 100% Accounts reporting with a balance to 11/12 91% was cause for the +9 point increase.

UPDATE: Was waiting on the decrease in #of accounts with balances to see affects:

3-11 aggregate 20.0%/dollars owed 4472

3-10 aggregate 20.6%/dollars owed 4590

Difference -118 dollars

3-11 # of accounts with Balances 11/12

3-10 # of accounts with Balances 12/12

CHANGE EX8 +9 Points
-------------------------------------------------
3-12 aggregate 20.0%/dollars owed 4654 (had a cli so didn't increase aggregate)

3-11 aggregate 20.0%/dollars owed 4472

Difference +182 dollars

3-12 # of accounts with Balances 10/12

3-11 # of accounts with Balances 11/12

CHANGE EX8 SAME
-------------------------------------------------

3-13 aggregate 19.8%/dollars owed 4608

3-12 aggregate 20.0%/dollars owed 4654

Difference -46 dollars

3-13 # of accounts with Balances 09/12

3-12 # of accounts with Balances 10/12

CHANGE EX8 +1

Summary

When accounts with balance were 100% 12/12 and dropped to 11/12 picked up 9 points. dollars owed 4472 util 20.0

When awb dropped from 11/12 to 10/12 there was no change. dollars owed 4654 util 20.0

When awb dropped from 10/12 to 9/12 picked up 1 point. dollars owed 4608 util 19.8%

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Sat, 13 Mar 2021 13:08:06 GMT

@Anonymous wrote:
@Iusedtolurk You know that is a good thing you did by putting the qualifier young mature in there because it apparently does make quite a difference. Even though you're on a mature card, there's a big difference in whether or not your credit history is say under 10 years or beyond 10 years.

And I'm just throwing that number out there I don't know what the real number is that makes a big difference, but we do see quite a lot of gains I believe from age of oldest Revolver as it makes its way towards 10 years.

I would suggest for future reference when we ask for scorecard we also ask for age of oldest account so we can determine how mature their profile is.

and yes, there does seem to be a score change at 100% of accounts open with a balance, however we have found for EX2 it includes closed accounts. And I have found Thresholds There and I'm in the middle of a test and I will post my results when I get done.

Who knows if that applies to the other versions. And who knows if that was conflated and possibly included a balance Threshold?

When you say "10 years" is it 10 years AoOA (revolver or any?) or 10 years AAoA?

I see you mention 10 years for AoORA, is that what you mean regarding that aging metric? And does it matter if it is open or closed?

Re: Score Increased When #Accounts Reporting Bal Dropped

Revelate — Sat, 13 Mar 2021 13:13:15 GMT

Has there been any evidence of closed accounts counting for anything beyond EX FICO 2?

I really would think just based on what I've seen data wise and also how the consumer credit market has changed that counting on closed tradelines probably changed in the decade between EX 2 and the FICO 8 models. Much like I'm pretty certain that installment acocunts don't count for accounts with balances under F8 either.

That said EX FICO 2, when I went from 3/35 total accounts to 4/35 total (open and closed) accounts with balance I did see a small drop so if we were thinking there's a 10% line in there, whee!

3/19 to 4/19 open accounts for reference, which sadly, is another pretty 20% number haha.

Re: Score Increased When #Accounts Reporting Bal Dropped

Curious_George2 — Sun, 14 Mar 2021 07:06:39 GMT

@Revelate wrote:
Has there been any evidence of closed accounts counting for anything beyond EX FICO 2?

I really would think just based on what I've seen data wise and also how the consumer credit market has changed that counting on closed tradelines probably changed in the decade between EX 2 and the FICO 8 models. Much like I'm pretty certain that installment acocunts don't count for accounts with balances under F8 either.

Good point. The places we were seeing it were limited to the EX2 family of scores. (Link) Classic and Bankcard, specifically. Thanks for remembering and clarifying that.

I don't recall previously hearing the theory that installments don't count in accounts with balances in the 8s. I find that approach a bit more intuitively appealing.

Re: Score Increased When #Accounts Reporting Bal Dropped

Revelate — Sun, 14 Mar 2021 18:29:41 GMT

@Curious_George2 wrote:
@Revelate wrote:
Has there been any evidence of closed accounts counting for anything beyond EX FICO 2?

I really would think just based on what I've seen data wise and also how the consumer credit market has changed that counting on closed tradelines probably changed in the decade between EX 2 and the FICO 8 models. Much like I'm pretty certain that installment acocunts don't count for accounts with balances under F8 either.

Good point. The places we were seeing it were limited to the EX2 family of scores. (Link) Classic and Bankcard, specifically. Thanks for remembering and clarifying that.

I don't recall previously hearing the theory that installments don't count in accounts with balances in the 8s. I find that approach a bit more intuitively appealing.

I think that might actually be in the FICO 04 model too honestly, FICO 98 appears to be all accounts which to your point makes little to no sense when you can't have an open installment loan without a balance so I don't precisely know **bleep** insight they're seeing on this one... data analytics have admittedly improved in the last two decades and I know the algorithm has changed over time. It was commonly accepted that EQ FICO 5 under the old scorewatch was revolvers: that might be incorrect but I don't think so... EX 2 and TU 2 (TU 98 to be explicit, both of these scores are FICO 98 models though the TU one is no longer for sale) were outliers in that regard I believe.

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Sun, 14 Mar 2021 23:46:18 GMT

@Anonymous wrote:
@Anonymous wrote:
@Iusedtolurk You know that is a good thing you did by putting the qualifier young mature in there because it apparently does make quite a difference. Even though you're on a mature card, there's a big difference in whether or not your credit history is say under 10 years or beyond 10 years.

And I'm just throwing that number out there I don't know what the real number is that makes a big difference, but we do see quite a lot of gains I believe from age of oldest Revolver as it makes its way towards 10 years.

I would suggest for future reference when we ask for scorecard we also ask for age of oldest account so we can determine how mature their profile is.

and yes, there does seem to be a score change at 100% of accounts open with a balance, however we have found for EX2 it includes closed accounts. And I have found Thresholds There and I'm in the middle of a test and I will post my results when I get done.

Who knows if that applies to the other versions. And who knows if that was conflated and possibly included a balance Threshold?
When you say "10 years" is it 10 years AoOA (revolver or any?) or 10 years AAoA?

I see you mention 10 years for AoORA, is that what you mean regarding that aging metric? And does it matter if it is open or closed?

@Anonymous i'm saying we should ask for the age of oldest account, because that is what the scorecard change is based on in version 8. And it seems soon after that three-year Scorecard change the new account penalty is close to nonexistent.

But you do raise a good point, because many points are gained from the age of oldest Revolving Account up to 10 years, it seems, so we should also ask that. We haven't determined where the buffer derives from, so it could be that or it could be AOYA or some other metric.

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Mon, 15 Mar 2021 00:02:58 GMT

@Revelate no sir, there is no evidence that AWB (accounts with a balance) includes closed accounts anywhere except EX2 so far, as @Curious_George2 stated. That is hampered by the fact there's no daily scores on the other two mortgage models obviously.

I have no idea whether AWB includes loans on the newer models, but I would love to know. That's one of the things on my list to determine. So many things, yet so little time.

Nevertheless even if the 2004 & 2008 versions exclude loans from AWB, does it still include closed Revolvers? Did that change from the 1998 version?

All data points from all members are appreciated.

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Mon, 15 Mar 2021 00:09:09 GMT

I'm actually doing some extensive testing on AWB right now. It's pretty easy with EX2 and consistent, but it's much more difficult to follow with the delayed alerts on the other two mortgage models, but I need to bring myself up-to-date and watch it closely.

I may be able to determine whether it's including closed accounts on those other models, if I'm lucky.

Here's another thing to consider some people have had only one bankcard with a balance and still received the too many AWB code, is this on version 8? If so that might be indicative that it's counting Loans.

Re: Score Increased When #Accounts Reporting Bal Dropped

Revelate — Mon, 15 Mar 2021 01:46:07 GMT

@Anonymous wrote:
I'm actually doing some extensive testing on AWB right now. It's pretty easy with EX2 and consistent, but it's much more difficult to follow with the delayed alerts on the other two mortgage models, but I need to bring myself up-to-date and watch it closely.

I may be able to determine whether it's including closed accounts on those other models, if I'm lucky.

Here's another thing to consider some people have had only one bankcard with a balance and still received the too many AWB code, is this on version 8? If so that might be indicative that it's counting Loans.

Pretty sure it's still a ratio. Too many accounts on a file with 3 tradelines /shrug. Even then probably would only see it on EQ from my data, maybe Cassie has it in her old data? Hard to say with a new file or even a dirty one too many things that might push that off the table.

I'll be interested to see what you find, I mean I remember EX 2 being all accounts was a surprise when it was originally postulated and AZEO always worked for F8 to my knowledge and even EQ5 too... you'd think with so many people getting anecdotal data we would've gotten this one right but other things were mistaken so this might be too.

Re: Score Increased When #Accounts Reporting Bal Dropped

Anonymous — Mon, 15 Mar 2021 02:16:19 GMT

@Revelate wrote:
@Anonymous wrote:
I'm actually doing some extensive testing on AWB right now. It's pretty easy with EX2 and consistent, but it's much more difficult to follow with the delayed alerts on the other two mortgage models, but I need to bring myself up-to-date and watch it closely.

I may be able to determine whether it's including closed accounts on those other models, if I'm lucky.

Here's another thing to consider some people have had only one bankcard with a balance and still received the too many AWB code, is this on version 8? If so that might be indicative that it's counting Loans.
Pretty sure it's still a ratio. Too many accounts on a file with 3 tradelines /shrug. Even then probably would only see it on EQ from my data, maybe Cassie has it in her old data? Hard to say with a new file or even a dirty one too many things that might push that off the table.

I'll be interested to see what you find, I mean I remember EX 2 being all accounts was a surprise when it was originally postulated and AZEO always worked for F8 to my knowledge and even EQ5 too... you'd think with so many people getting anecdotal data we would've gotten this one right but other things were mistaken so this might be too.

@Revelate well when we had the experts they explained AWB could be a number or a ratio, with the implication thin files could be a number. So I'm afraid the ratio argument might not hold up for a thin file, imho. But I believe we have seen that situation with people that had more than just a few accounts.

EX2 was definitely a surprise as was the fact version 8 and 9 segment on revolvers and many other revelations we've made over the last couple years, but AZEO it's still good for all models except 8/9 requires a modified AZEO, if you have authorized user accounts that are not discounted by the anti-abuse portion of the algorithm.