Showing results for 
Search instead for 
Did you mean: 

General Scoring Primer and Version 8 Master Thread, pub.5.17.20

Super Contributor

General Scoring Primer and Version 8 Master Thread, pub.5.17.20

Please don’t merge. We did something like this, but it cannot be updated/maintained as original OP is unavailable.


Intro- (Table of Contents/Index in Post 8 to make it easier to find subject matter/understand flow. Link to TOC.) (Also acronym and useful links in Post 8 and 10.)


Quick Reference


**ccquest's workbook calculates utilization, balances, ages & more: Link. **

**Cassie's Negative Reason Code/Score Factor Charts: Link. **


FICO Score 8 select optimal Characteristic Attributes




Payment History

100%, Zero baddies. Penalty fully removed after 7 years, with few exceptions. 60D late or worse is a dirty scorecard.

Aggregate Revolving and Installment Utilization


Number/Percentage of Revolvers with a Balance

Recommend 20% of revolvers or less, IMHO. Metric is not weighted as heavily in Score 8 and varies by bureau/scorecard. EX8 less sensitive. (15% for 5-4-2)

Revolving Balances

Under $100; Never $0 or AZ (All Zero) penalty of 10-25 points; All AUs at AZ is a separate and independent AZ penalty on 8 & 9.

AoOA-Oldest Account

Not a scoring factor. Segmentation factor for clean profiles, believed to be at 3 years.


Age of Oldest Revolving Account

Scoring factor mistaken for AoOA, but if oldest account is a revolver, AoORA=AoOA

AAoA- Avg Age of Accounts

Max award believed to be at 90 months


Avg Age of Revolving Accounts

Scoring factor; similar to AAoA, but revolving activity is always weighted more heavily.


Age of Youngest Account

Scoring factor; awards points at thresholds that are multiples of 3 mo, up to 24 mo? Previously thought a segmentation factor.


Age of Youngest Revolver

Not a scoring factor. Segmentation factor for clean profiles. Threshold at 12 months AoYRA.

Inquiries in Last 12 Months

Zero. Score penalty removed at 365 days.

Total Number of Accounts/Mix

Not a scoring factor. Segmentation factor for clean profiles. 4 TLs for Thick, include 1 loan for diversity points. (Penalty for too few/many accts.) Revolver:Loan Ratio 3:1 or 4:1?


Where did this come from?


@MWGardener19  was the inspiration for this thread. He researched/composed most of this first post, with corrections, alterations and edits by @Birdman7. However, MyFICO Contributor Birdman7 has not thoroughly fact checked the majority of the information in MWGardener19's part of this first post, which is a background and overview. The mechanics start toward the end of this first post with Scorecard Basics by Birdman7. No one knows the exact workings outside of FICO, but this is the BM7's sections have the best information we've been able to discern thus far, IMHO.


Birdman7 began by cleaning up a small Reddit post MWGardener19 brought from a writer who had gathered data from this forum through years of reading and made a small synopsis. Birdman couldn't let the inaccuracies stand, and so much was missing, so he went crazy. This is where it ended up, pretty much a new creation, but the Reddit post was the starting point, so we are acknowledging that properly.


No doubt many will have additions, corrections and criticism. Nevertheless, this is meant as a primer and a reference and I hope to update it as we learn more. Please give your feedback.




Brief background


Fair Issac and Company - FICO - states that Score 8 is the most widely used credit scoring system. Score 8 is one of many credit score models that FICO created and lenders use to evaluate and score credit risk (Score 2, 3, 4, 8, 9; Bankcard 2, 3, 4, 5, 8, 9; Auto Score 2, 4, 5, 8, 9).


There are more scores, and we know that at least 2 new scores are expected to debut at the end of 2020 (Score 10 and Score 10T). FICO began creating their systems over 60 years ago (1956). The FICO scoring system as we know it today is only about 30 years old (1989).


 FICO is a large, global, publicly owned company with over 4000 employees. It is worth noting that 88% of the $1billion in revenue that FICO generates comes from banking and insurance customers. 35% of that revenue is from international customers. Scores purchased by consumers are a small part of their business.


FICO scoring and these forums


As users of this forum have or will come to learn, there are certain observations that can be said about FICO scoring:


- we have come to know GENERALLY how FICO scoring works, AND

- we have come to know A LOT about how certain aspects of scoring works, BUT

- we have come to know that we do not know EXACTLY how all of FICO scoring works.


FICO's approach to performing credit scoring is proprietary, meaning that it is private. The folks at FICO knows how their scoring systems work. The rest of us take increasingly educated guesses at learning the finer points of those scoring models, and we come to forums like this to learn for ourselves how to understand, improve and manage our scores, and to help others to learn as well.


As of April of 2020, these forums have nearly 300,000 users and over 4 million posts. These forums provide and contain great value.  As users have also likely observed, many of the same questions are asked and answered again, and again.


The purpose of this thread is to try to capture what we know, what we think, and what we are still trying to learn more about.  Because there are more than 28 different FICO models that exist, the guidance and answers that apply to one scoring model may be different for another model. Rather than trying to gather all of the collective learning about all of the FICO scores in one thread, this thread will focus on one score: FICO Score 8. We will make notes about differences in other versions for certain metrics.


These forums have MANY VERY smart members with extraordinary depth.  It is my hope and intention that we try to capture our knowledge in one thread to make it easier for forum members - new and old - to have a single stop to get the most complete understanding of what we know about FICO Score 8. Although Score 8 was released in 2009, we are still learning more about it today.


At a high level, Score 8, as with other FICO scores, is used by lenders to understand credit risk. Credit risk measures how likely or unlikely it is for a prospective debtor to default on a credit obligation by 90 days in the following 24 months. Each FICO 8 score draws data from its respective bureau, Experian, Equifax and TransUnion to calculate a score. The Score 8 algorithm is slightly tweaked for each of the 3 bureaus. While scores from identical data across bureaus are close, they are not exact.


Score 8 software systems evaluate credit report data, and use very complex mathematics to calculate and report a credit risk score. (How complex, you ask? The underlying math considers Lorenz curves, Gini coefficients, normalized log Bernoulli Likelihood, multicollinearity testing, and other math that is way beyond simple addition, subtraction, multiplication and division.) [ I believe a lot of the mathematic mentioned are used to analyze datasets to create the scorecards (algorithms) used to generate scores. Fact check? BM7]


FICO Score 8 key differences


Score 8 was designed to be more sensitive to high revolving utilization than earlier versions. It excludes nuisance collections (under $100), and is more forgiving of isolated delinquencies compared to earlier versions, but all other accounts must be in good standing. Whereas earlier scoring versions were more customized for each specific bureau, an objective of Score 8 was to reduce disparity among the scores at the 3 bureaus. (Version 9 is supposed to have even less disparity.) You can still have a score difference of +\- 30 points in Score 8 among bureaus with identical data, per the esteemed Moderator Emeritus Revelate.


 So, how does scoring work in FICO Score 8?


Again, we do not know all aspects with precision. There are a number of things that have been learned. Many thanks to Reddit user rtanaka6, who captured many of Score 8 highlights from years reading this forum in a Reddit Post. It was the starting point for this, but this has really became a new creation. Also, note: this thread is dynamic. I will update it with credit given to posters who provide their expertise so we can recognize and thank them for helping us to get smarter.


 As with other FICO models, there are 5 categories or "ingredients" which are evaluated. Your credit report data is fed into the FICO "blackbox" at the respective CRA (Credit Reporting Agency) and is ran through the appropriate scorecard (algorithm). The scorecard evaluates your data based on these 5 areas and result in a numerical score and negative reason codes. Positive reason codes are meaningless and should be ignored, but negative reason codes are listed in order of precedence and can offer insight into the reasons for your score in any version. FICO Score 8 ratings and scores are:



Score Range

Exceptional (or excellent)

800 - 850

Very good

740 - 799


670 - 739


580 - 669


300 - 579


The categories/ingredients that are evaluated from your credit report, and their weightings are:



Payment History



Amount of Debt



Length of History



New Credit



Credit Mix


Remember credit scores are connected with approvals, but are often not the only factor. Inquiries (more here) are used to: 1) generate your score and negative reason codes, 2) analyze your credit data, and 3) give notice to other lenders you have sought credit. This information is used to determine approvals, interest rates, terms, and starting credit lines. Credit scores matter, but are not the sole reason consideration. Lenders will also ask for data not contained in the CR data to consider, as well, such as income and housing expenses. Please note income is not reported to the bureaus and is not a part of scoring. Lenders consider that separately.


MWGardener19's edited work ends here/

Begin Birdman7's learning

Scorecard Basics


There are various metrics within each category. FICO calls each metric a "characteristic" and the cooresponding value is called an "attribute." To be more granular, we break characteristics into 2 categories: "segmentation factors." and "scoring factors." Segmentation factors determine scorecard. Scoring factors directly affect score and their signal strength (weighting) varies by scorecard.  Scorecards are algorithms by their simplest definition. The scorecards generate a score and negative reason codes from the CR data it's fed based on the scoring factors as weighted per the scorecard.


Score 8 has 12 scorecards: 8 clean and 4 dirty scorecards. Which scorecard you are assigned to depends on the segmentation factors. For clean profiles these segmentation factors include thick/thin (number of accounts); mature/young (age of oldest account); and no new account/new account (recency of new revolvers). For dirty profiles, I believe segmentation factors are severity and recency.


Scorecard assignment is a complex matter to be covered in another post on this thread, maybe another thread; there are many posts scattered throughout these forums that we will try to consolidate here. What is important to know about scorecards is that they impact how your specific score responds to information in your credit report and they are why your profile reacts differently than someone else's for the same event. The scorecard to which you are assigned determines your minimum and maximum scores, which negative reason codes are applicable to you, and the signal strength (weighting) for the various different scoring metrics described below.


Each scorecard is geared to specific credit profiles. Scorecard reassignment (rebucketing) occurs when you change scorecards. This can result in a score boost or drop depending on the situation. For instance, when you go to a higher scorecard, you are basically moving from the top of one ladder to the bottom of another and would now be compared to a sub-population with better credit profiles. Therefore, often you will see a score drop. The reverse also holds true, but be aware there are many exceptions and scorecard theory is complex. What follows is an example of how different scoring factors have differing signal strengths depending on scorecard:





A scorecard example for Payment History, Amounts Owed, Length of History, New Credit, and Mix follow (Keep in mind awards vary by profile and there are far more than 1 scoring factor per category):




Worth repeating, there are 12 cards in Score 8. (The rumors of 14 are false, 2 were pulled over AU issues.) There are 8 clean and 4 dirty scorecards. Clean/Dirty is the first segmentation factor and determines the subsequent segmentation factor path. A clean profile is then segmented by: Thick/Thin, Mature/Young and then No New Account/New Account. We will go over these in more detail as we progress through the categories.


If instead a profile qualifies as dirty, we are not as clear. I believe the profile is then segmented based on severity and recency. Severity appears to be split into the 2 delinquency cards and the 2 PR (public record) cards.  Delinquency cards contain profiles with 60 day lates and any other serious delinquencies other than public records. PR cards include profiles with collections, bankruptcy, or other public records, such as tax liens, plus any delinquencies. A judgment would also put you there, but they are no longer reported.


The below is my approximation of how the scorecards in FICO 8 are segmented:





For delinquencies (not collections) recency appears to go from the date of last update, if unpaid. If paid, recency appears to go from the date paid in full or settled. Therefore, once paid or settled, delinquency recency appears to be frozen.


So one could express their scorecard as clean/thick/mature/no new accounts or dirty/PR/recent. Scorecard theory could have its own thread and there are many threads you can search to learn more. You don't need to be an expert, just know this is part of why different profiles react differently to similar actions and that certain penalties exist in some cards that do not in others. For instance, there is no "new account" penalty in dirty cards, whereas it's a 10-20 point penalty in clean cards. So, one good reason to get cards while dirty. Utilization is also more important in clean cards, for example. Below is a slide from FICO describing Score 8 segmentation very generally:






(Version 9 added a 13th scorecard for those with high revolving utilization. The specifics of how it works or what segments you into it are not yet known, just that it is for those with high revolving utilization.) (The '98 and '04 versions had 8 clean and only 2 dirty; the dirty were not segmented by recency and apparently delinquencies are not banished to a dirty card for 7 years there either.)


We will now examine each Category and its segmentation and scoring factors in detail, sequentially, in sections 1-5, below. I will use the terminology of segmentation factors and scoring factors, as it's more granular than the generic term characteristic, which doesn't indicate whether it's for segmentation or scoring.




Last edited: 2.23.21 5:23pm.

© 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
341 REPLIES 341
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20




Per MF, 7 components make up Payment History :


  • Payment information on credit cards, retail accounts, installment loans, mortgages and other types of accounts
  • How overdue delinquent payments are today or may have become in the past
  • The amount of money still owed on delinquent accounts or collection items
  • The number of past due items on a credit report
    Adverse public records (e.g., bankruptcies)
  • The amount of time that's passed since delinquencies, adverse public records or collection items were introduced
  • The number of accounts that are being paid as agreedLink. 

A. Derogatory Categories


This is the most important category, and absolutely 0 derogatories is necessary for maximum scoring in this category. The algorithm looks for delinquencies, (30, 60, 90, 120... CAs (Collections), BK, CFAs (consumer finance accounts), tax liens, or in Version 9, LM (loan modification, code AC)  Link.  etc.. Ideal is none. For one isolated 30 day late, you will wear a significant penalty.


A 60 day late is considered a major derogatory and will segment you into a dirty scorecard for 7 years on Score 8. (It appears a 60 day late only assigns you to a dirty card for 2 years on TU4 per recent testing by our esteemed Moderator Emeritus Revelate Link.  This appears to be confirmed for EX2 as well, by MyFICO Contributor Ficoproblems247 Posts 10 and 17.  And evidence that EQ5 segments delinquencies at two years. Link. Note that the "new account" negative reason code does not exist on dirty cards, so its presence indicates a clean card.)


Creditors typically report a debtor 30, 60, 90, 120 days late and then the account may be charged off. (At 120 days late a chargeoff is typical for loans and at 180 days a chargeoff is typical for CCs. [See link in Post 7.])  A chargeoff  (CO) does not mean the debt is no longer owed or collectable, just that it isn’t expected to be collected and has been removed from the asset column of the creditor's balance sheet. At this point, the account is closed if it hasn’t already been. If the account still has a balance and is owned by the OC (Original Creditor), the TL (Tradeline) will reflect the outstanding balance.


Please note a CO is an accounting measure and may not have a direct effect on your score. The score loss appears to be tied to the TPOD (total period of delinquency) Link, which is from the DOFD (date of first delinquency) until the last update if unpaid, or from DOFD until paid if paid. That's why when a chargeoff is regularly updated, the score continues to be suppressed, but if it's not being regularly updated, it does not continue to suppress score. Also why if you pay it way later when it hasn't been updating, you can see a drop.


If you have paid the balance on a chargeoff or subsequently do so, the TL should be updated to reflect the $0 balance and to change the Current Status from CO to paid or to paid, settled for less. If the debt is sold, the balance should also be updated to $0. These situations will affect utilization, see Section B, Amount of Debt and can also affect aging (TPOD) see Section C. 1., Aging, below.


B. Collections/CA (Collection Authority)/Dunning Notice/DV (Debt Validation)


If you fail to pay a debt after CO, it could be assigned or sold to a CA (Collections Authority). This is a separate and additional TL to the OC’s, and causes additional penalty and scorecard reassignment to a PR card, if you had no other PRs. (Note: all OC’s don’t report a OC TL, such as cable/phone, but a CA can still appear and penalize.)


Upon first contact, CA is required to notify you in writing within 5 days of your DV (debt validation) rights. (This may be included in the first contact and is called a Dunning Notice.) You have 30 days to request DV. It should be sent CMRRR (certified mail return receipt requested). Until the CA responds, CA is prohibited from attempting to collect and therefore cannot contact you (except filing judicial process, etc...). If the CA fails to timely respond or cannot validate the debt, it may delete the collection. If not, you may seek removal via dispute with the CRAs. If you let the 30 days pass without submitting a DV, you still may send one, but the CA is not required to respond and is not barred from continuing to attempt to collect.


If the debt is sold to a CA, the OC TL balance should be updated to $0; if assigned, the balance should remain on the OC TL, as well as be listed in the separate and additional CA TL. But, be aware that if an OC’s TL is assigned to a CA, the CA is required to delete the collection TL IF the debt is recalled  by the OC. So, the preferred method of dealing with a collection is to attempt to negotiate with OC and have the OC recall it. This will require deletion of the collection, even though the OC’s CO TL may remain and be updated. To address the remaining the CO, see GST (Goodwill Saturation Technique).


However, if the OC TL is sold to a CA, then you must negotiate with the CA and the preferred course of action is to request a PFD (Pay for Delete) from the CA, which is where you negotiate with the hope of paying in return for deletion. Be aware this is against CRA policy and CAs may refuse; however, some will accommodate under various interpretations. It is common for a CA to be reluctant to put it in writing for obvious reasons. It's noteworthy that oral contracts are enforceable; they're just hard to prove if denied. Laws on recording phone calls vary by state, some require the other party be notified, some only require one party to be aware. Consult counsel/law in your jurisdiction. Link.  


For collections (CAs), "[a]s far as your FICO Score is concerned, two things are considered:

  • has a collections appeared on your credit report
  • when it was reported" Link. 

So, whether or not it's paid is not even a consideration, scorewise til 9! But it looks better to prospective lenders if paid, even if late. See Aging below.


C. Derogatory Aging


  i. Scoring


It is believed delinquency penalty is reduced at 6 months, 1 and 2 years for 30, 60, 90 and 120 day lates, but all baddies will affect score for 7 years. Recency and number of baddies is believed to play a critical role. The change at 2 years appears to be scorecard reassignment, which may also apply to paid chargeoffs. 


Link to derogatory aging graph. (Credit: MyFICO Contributor ABCD2199.)


COs (chargeoffs) can be tricky. They are not like lates subsequently brought current, where the TPOD remains constant at 30, 60, 90, or 120. With a CO, TPOD continues to grow and suppress scores until it's paid, if regularly updated. If instead an OC fails to regularly update the OC TL, the algorithm doesn't know whether the delinquency period has grown, so TPOD is frozen and no further score suppression occurs; however, if the TL is updated at some later time, the algorithm realizes the TPOD has increased and you may be penalized appropriately to catch up, (TPOD catchup penalty) Link. Link. Link. (Credit: the esteemed MyFICO Legendary Contributor RobertEG.) This typically occurs when someone pays an old late, only to be surprised by a score ding. Could also occur when updated by a sale to a CA, when paying a CA where the debt was assigned, or from a dispute updating the delinquency, if the TL was not regularly updating. However, this drop can be offset or outweighed by changes in utilization or other changes.


There are 3 separate fields in particular that are relevant, Payment Status, Current Status Link and Date Updated. (Credit: Our esteemed MyFICO Legendary Contributor RobertEG.) Payment Status shows the highest level of delinquency that has ever occurred on the account (30-CO), Current Status shows the current status, whether CO, paid, or paid, settled for less, and Date Updated is obvious.


When a CO is regularly updated (Date Updated Field), it continues to suppress scores because the TPOD is increasing, similar to going from 30 to 60. Because the Current Status field remains CO, the algorithm knows it's still delinquent through the last update and calculates increasing TPOD (Date Updated Field - DOFD) and penalizes appropriately. Thresholds, if applicable, are not known. If the Date Updated field is not updated, the algorithm cannot determine TPOD has increased and therefore can't further suppress/penalize scores. (Be happy when they aren't updating, but know you may eventually pay the piper.)


When an update does occur, the OC updates the Date Updated field and TPOD is increased, if Current Status is still CO. Updates typically occur due to updating the Current Status field to paid or paid settled for less, or from a dispute, as stated above. Once paid, subsequent disputes should not cause dings because Current Status is paid and TPOD is therefore frozen. (Once a CO is paid, one can try the GST in hopes of having the OC remove the lates/account. Can't hurt, but might take 100 tries! Persistence.)


CAs are different and do not appear to help score if you pay them, only if they are deleted or age to some unknown threshold (PR scorecard recent > mature 2 yr?). "Your score weighs collections on your credit report according to when the collection occurred. Generally, the more recent the collection, the more it's going to hurt your FICO Score." Link. So over time a collection appears to reduce its penalty. Recency is determined from when it "occurred."


I believe "occurred" means opened, not updated. (Some think if regularly updated, it can't age to a mature PR card to lessen the penalty, but an MF blog says paying a CA in ver. 8 and older lacks positive scoring impact, so color me skeptical.)


Tax Liens and BKs should also reduce penalty over time when you shift out of the recent PR scorecard into the mature PR scorecard (believed to switch scorecards at 5 years on the mortgage scores). Judgments are no longer reported, but that could always change.


  ii. Removal


Regarding removal, the (FCRA) Fair Credit Reporting Act, §605(a) applies (15 USC § 1681c). Link.(page 28 by page, 22 by pdf.)


For open accounts, scattered lates are treated differently than strings of lates. With scattered lates, they are each removed at 7 years individually. With strings, it depends on bureau. "...Experian...excludes...delinquencies in a common “string” 7 years from [DOFD]...[t]he other two CRAs have no official, published policy interpretation, and have...excluded based on [DOFD] OR have treated each...delinquency as its own separate adverse item of information, and thus have not excluded...until each has reached...7 years..." [1] Link. Link 2.  The particular subsection for monthlies is a catch-all provision and lacks specificity. §605(a)(5), FCRA. Please note US Gov't insured/guaranteed student loans or national direct student loans have specially lengthened reporting periods. See sections 430A(f) and 463(c)(3) of the Higher Education Act of 1965, 20 U.S.C. 1080a(f) and 20 U.S.C. 1087cc(c)(3), respectively.


For a charged off account and/or an account that was sent to collections, the entire TL(s) (OC, and CA if applicable) must be removed at 7.5 years from the DOFD, but is typically removed at 7 years. See §605(a)(4) & (c)(1), FCRA. (5yrs NY) A closed account with delinquencies should stay for 10 years post-closing, as long as it was not charged off or sent to any type of collections. The lates will age off but the account will remain for 10 years. Closed accounts without delinquency should remain for 10 years, But from update or closure? See: link. However, MF Contributer SouthJamaica reports accounts being dropped as soon as a year after being closed. 


Paid tax liens are removed at 7 years from payment. §605(a)(3), FCRA. Bankruptcy is removed 10 years from date of entry of the order for relief or the date of adjudication. §605(a)(1), FCRA.

Early exclusion: link. 

  iii. CR Removal Quick Reference


So, §605(a)(1): BK - 10 years from order/adjudication

                     (3): Paid Tax Liens - 7 years from payment

                     (4): COs and CAs - 7.5 years from DOFD/7 years from CO/Collection; 5 yrs NY Link. 

                     (5): Catch-all for any/all other negative information - 7 years (starting date ambiguous).


  iv. Judicial SOL (Statute of Limitations, varies by state)


Note: Whether a debt is judicially actionable is based on the state's SOL (Statute of Limitations) of the jurisdictions in which the creditor could sue the debtor (The Plaintiff has the choice of forum with limited exceptions). This could be: 1. the state the creditor is incorporated/organized in, 2. the state you entered into the agreement in, 3. your home state, or, 4. if you have moved, your new home state. (Consult qualified counsel, Conflict of Laws and Jurisdictional Law are complex and beyond the scope of this thread and forum.) This Right of Action SOL is separate and distinct from the federal exclusion SOL that determines how long negative data may be reported on your CR by the CRAs, which is the above stated time periods per the FCRA. 


  v. Various

Loans can also give points over time without regard to installment utilization thresholds and " age is an actor that influences score outside of B/L ratio on installment loans..." [2]
Link.  (Credit: Thomas_Thumb) The points could derive from this category or the Length of History category's loan metrics.


D. Amount Still Owed on Delinquent Accounts


Per MF, the amount still owed on delinquent accounts is factored into the Payment History category. So once an account is charged off, the amount may be considered under the Payment History category rather than the Amounts Owed category. However this may not be. The same article says the amount owed on collections are considered and we know that paying a CA without deletion offers no points. The only way I can reconcile this is if the initial penalty is determined in part from the CA amount.


At two years sleeping, COs may be taken out of revolving utilization and only accounted for by the Payment History category. Link. 


E. Multiple Delinquency Penalty


TU8 and EX8 are more forgiving of an isolated late. When they see frequency, an additional penalty is imposed. However, 800s are possible with 60 day lates.


F. Recency


Score 8 seems to penalize you more heavily for recent delinquencies/derogatories. I believe this is due to the 2 new dirty scorecards I believe segment based on recency. (5-4-2 appears to reassign to clean cards with penalties at 2 years.)


G. Number of Accounts Paid as Agreed


Typically on young or thin profile with a new account(s). The more positive payment history, the better.


Last edited: 2.23.21 5:13pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 2 of 342
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20






**ccquest has created an awesome Excel sheet that will calculate utilization, balances and ages easily for you at Link.**

A. Utilization

and aggregate utilization are scoring factors.


          1. Revolving

i. The major recognized Aggregate revolving utilization thresholds are believed to occur at 10%, 30%, 50%, 70%, 90%, and 100% (Some scorecards may also have lower thresholds.)

ii. The major recognized Individual revolving utilization thresholds are believed to occur at 30%, 50%, 70%, 90%, and 100% (Some scorecards may also have lower thresholds.) (Individual utilization is usually but not always same as statement balance. Credit: SouthJamaica)

For max points, you should remain under the lowest thresholds. As you go up, penalties are assessed, more per threshold for aggregate than individual. Typical rounding is used, e.g., you should be able to have a 9.4% balance without going to the next interval of 10% - 29%, but 9.5% = 10%.

See the chart below for a graphical example of the effects of aggregate revolving utilization: 




For PLOCs (typically revolvers) and proof of True chargecards, Link. HELOCS usually factor in revolving utilization on 5-4-2, but are factored separately and as a ratio to all revolver balances in version 8 at TU an EX. See Codes 62 and 64.


Also know that EQ8 additionally tracks bank/national revolving account utilization without other revolving accounts (Code 89) and then other revolving utilization separately (Code 90).


          2. Loans


- One metric of loan utilization is aggregate loan utilization of all open loans (current total balances divided by the total original loan amounts) The largest threshold (biggest point award) comes when B/L is <9.5% B/L. (Updated to 9.5% based on new knowledge from Fico experts.) This was discovered by our esteemed Moderator Emeritus Revelate Original thread and his strategy devised is worth 15-35 points (credit to MyFICO Contributor Saeren for top end of range of 35 points on an SSL.). See SSLT thread for detailed information. A smaller award is believed to have a threshold ~65%. Link. Changes seen at other suspected thresholds are believed to be due to changes in total installment balances. There's also a report of a loan threshold ~88% from Remedios.

Long story short, acquire a long-term loan, paydown to less than 9.5%, do tiny autopay for activity. Catch is, it requires a FI that doesn't advance the maturity date and won't work if you have other loans, as installment utilization is based on the aggregate. 


Here is a detailed post from May 2020 of a member going through the process of acquiring an SSL and executing the strategy at Navy. SSLs are/were available at Navy NASA FCU (link in post 7), and SSFCU, and PLs (Personal Loans) are believed to work at Alliant, WF (3k min, $75 fee) and USB. PenFed? There are others and these may change. 


-Be aware that while paying aggregate loan utilization to <9.5% gives a nice award, there is also evidence that time open can offer points for installment loans [2]. (Credit: Thomas_Thumb)


Another great discussion of the intricacies of loan utilization by our esteemed Thomas_Thumb: Link. 

It appears all 3 look at aggregate loan utilization while EX and TU also track individually mortgage, auto, and HELOC utilization, as well as a special HELOC balance to revolving balance utilization.


B. Number of Accounts Reporting a Balance/AZEO


The number of accounts reporting a balance impacts score independent of utilization. The higher the number of accounts with a balance, the higher the penalty. (Only at EQ8, the Number of bank/national revolving accounts with balances is also a separate scoring factor. Code 23.) Number of accounts reporting a balance is not as heavily weighted on Classic 8, and less so on EX8. Ex.:


EQ BC Fico 8 vs cards reporting a balance.jpg

Credit: Thomas_Thumb


This metric is much more influential for the mortgages scores (especially EX2), which is believed to have lower thresholds. 

There was debate as to whether the account metric's thresholds were raw numbers or percentages. According to fico experts, it can be either/both. They implied it could be a number on thin/young scorecards and a percentage on thick/old scorecards or ? Link. 


Many people think Score 8 has a 33%, a 50%, and maybe a 80% or 100% threshold for this metric as a percentage rather than number. Well, there appears to be one at 20% on EQ8, link, (the weighting may vary by bureau) and the mortgage scores and I've seen too much evidence of the 100% threshold.


There could be 20%, 40%, 60%, 80%, and 100% thresholds or 20%, 50%, and 100%. More testing is advised and hopefully other members can add data for the thresholds. (MyFICO Contributor Trudy's dps lead us to discover this metric includes closed revolvers on EX2. See link. Confirmed. We must still determine whether it is applies to Score 5/4/8/9.)


No matter how many revolvers you have, if you only have only one report a balance, then you will be at the lowest possible number/percentage that your profile will allow. (You can't change the fact a loan has a balance, so you can trigger too many accounts with a balance if you have too many loans.)


This has given rise to the infamous AZEO (All Zero Except One) concept. You’ll often hear this recommended, especially for mortgage scores. It doesn’t mean you have to only let one revolver report a balance to be at your best scores, but it guarantees you will be below the lowest threshold possible for your profile for these metrics (which may or may not be one revolver).


Additionally, AZEO is typically recommended as one national bankcard with a small balance. This actually potentially optimizes 3 or more metrics. Number/percentage of accounts with a balance, revolving utilization, and the revolving balance metric(s). 


So, it’s a easy way to give good advice to achieve the best scores, but you may be able to have several revolvers with small balances report and still be at your best scores. You just have to test your individual profile. Also, you don't have to pay interest, just pay after it reports, but before the due date.


A note of caution about your AZEO card. Use a national bankcard with a CL no higher than $31,000, because the mortgage scores exclude cards with higher credit limits. Link. Avoid retail cards, credit union cards, and charge cards, as they can cause unintended consequences.


Typically the more revolvers you have, the more you can have with a balance and still be at your best score. This author recommends at least 5 revolvers at AZEO, 1 loan with B/L under 9.5%, and no inquiries and no new accounts in the last 12 months for reaching your best scores. Also noteworthy, this is a metric that is more reactive on TU and EQ. Many report no changes at EX.


While not counting towards revolving utilization (except on EX2, where the high balance is used as the CL), true charge cards (American Express charge card) do count toward one's number/percentage of revolvers with a balance metric in Score 8. The reason it's not advised as your AZEO card is next.


C. All Zero Point Loss- (AU Test included)


When all revolvers report $0 balance (AZ), there is a loss of ~10-20 points. If one has AU cards (that are not discounted by the anti-abuse algorithm) and they all report $0 balances, an independent AU AZ point loss will also occur. (A retail card used for AZEO seems to give a partial AZ loss.) EQ8 may penalize for a retail balance, maybe EX? 


This allows to test whether your authorized user cards are being discounted by the anti-abuse portion of the algorithm on versions 8 & 9. (FICO included this to address TL renting.) Have all AU accounts report zero and see if you experience a score loss. When an AU reports a balance the points should return. If they do, they're counting. If not, they are not counting for 8 and 9, but they will still count for the mortgage scores. That part of the algorithm didn't exist back then, so they always count for mortgage scores, though lenders realize the score is artificially inflated.


To avoid AZ loss and maximize scores, many people use AZEO, where only 1 national bankcard (and 1 AU account, if applicable for Scores 8 and 9) reports a small balance $5-$20. This AZ loss is a frequent post of why did I lose 15 points?


Charge cards are not revolvers and will not save you from the AZ losses. Correct, they count for number of accounts with a balance, but will not prevent AZ loss. That's because the AZ metrics are based on revolving balance, I believe.


Oh yeah, be careful if you use Chase as an AZEO card and pay it to $0. They automatically off-cycle update $0 balances, so it can cause an unintentional AZ loss. (Credit: SouthJamaica.) However, they can also be the perfect AZEO card as long as you leave a couple dollars when making payment! Also convenient if another card accidentally reports, as Chase will always report the $0, don't even have to ask.


D. COs


The exact mechanisms of how unpaid COs affect utilization are currently under study. We do think an unpaid chargeoff has an effect on utilization. We're just trying to confirm and understand it. However it is possible the only penalty comes from the Payment History category instead after chargeoff.


Chargeoffs with balances were considered maxed out cards by common wisdom, but this does not appear to be correct. Further study is ongoing and DPs are appreciated!!. 


At two years sleeping, COs may be taken out of utilization and only accounted for by the Payment History category. Link. See section A for other possible effects due to Payment History.


E. Revolving Balance Metrics-Aggregate and ABORT

We don't know where all the thresholds are, but we do know score is influenced by the aggregate balances on revolvers.


At EQ8, Amount owed on bank/national revolving accounts is additionally tracked. Code 66. TU and EX additionally track mortgage and installment balances separately where EQ doesn't track them at all. TU does not separately track retail balances while the others do. Revolving balances are weighted less heavily than revolving utilization. 


I believe the estimator indicates thresholds at $500, $1000, $5000, $10,000, and $20,000, but it lumps revolving and non-mortgage loans into it. Take that for what it's worth, we haven't had enough control tests to know, but it appears most of the thresholds are at $25,000 and below, or maybe they repeat perpetually? Because of this, those with very large TCL's may see score changes from this metric and mistake them for utilization because 1% for them may be $5,000 or $10,000 or more. Apparently there is a measure of installment balances as well, but that is not something heavily weighted or easily tested.


Average Balance on Revolving Trades (ABORT) also appears to be a Scoring Metric from fico slides (not sure if it applies to all versions). An average balance under $100 but above $0 appears to be optimal, if the slide is believed to be accurate. 


F. Loan Balance Metrics

The amounts owed on mortgage, open mortgage, and installment loans are considered at TU and EX. Apparently not at EQ; it relies on loan utilization I guess. 


G. Amount Owed on All Accounts Metric

All bureaus consider this.


Last Edited: 2.23.21 5:14pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 3 of 342
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20






A. AoOA - Age of Oldest Account is a segmentation factor in clean profiles

"AoOA is not a Fico scoring factor. It is a scorecard assignment segmenter. If an increase moves you to a different scorecard your score may change due to a shift in weighting of the factors used in scoring and the assigned min/max scores associated with the scorecard." [4] (Credit: Thomas_Thumb) Link 


A mature profile is more stable and preferred. Penalties are less severe, but awards are smaller. The Score 8 threshold was previously theorized to be between 10-15 years. But it has been found at 3 years! Link.   (It also appears EX2 segments at 2 years based on CassieCard's testing and data. Link.  ) TU4 and EQ5 also appear to segment at 2 years. The reassignment to mature typically has a 10-15 point loss, but for some can be little or none. It's really profile specific because the metrics are weighted differently. 


There is only one threshold for AoOA per Version, it's a segmentation factor, and it causes scorecard reassignment; it is not a scoring factor. If you see a point gain at a suspicious AoOA threshold, it's probably an award from AoORA because that is probably the same as your AoOA.

If a credit card is your oldest account, as it is for most people, AoOA=AoORA. These metrics therefore run the same for most people. The lack of realization of AoORA has caused delay in determination of the thresholds for AoOA and AoORA.


B. AAoA - Average Age of Accounts is a scoring factor

It appears these awards are greater on young scorecards, meaning those whose AoOA is < 3 years old for Score 8.

Awards seem to be at multiples of six months

- 3-15 point annual increase

- Maximum award believed to occur by 90 months.


There are 12 and 18 month thresholds on young scorecards, unconfirmed reports of thresholds at 24 and 30 months, and likely thresholds at 60, 66, 72, 78, 84 and 90 months. (K-in-Boston confirmed and reconfirmed the significant threshold at 84 months on a clean/thick/mature/new account profile.) I'm sure there are more. Link with links. 


C. AAoRA- Average Age of Revolving Accounts is a scoring factor- under study


Negative reason codes point to this metric.


D. AAoOLA-Average Age of Open Loan Accounts is a scoring factor under study-


Negative reason codes point to this metric on some variants.


E. AAoLA-Average Age of Loan Accounts is a scoring factor- under study- 


Negative reason codes point to this metric. Link.

AoORA-Age of Oldest Revolving Account is a scoring factor under study


FICO never gives away exact thresholds, but we've had a reports of thresholds at 24 & 30 months: link, one at 36 months on a PR scorecard worth 6 points on versions 8 & 2: link, one at 72 months: link, 18 to 29 points, and we also have 2 reports of a possible threshold at 9 years: link. Consider the following from FICO:




Since most people get a revolver first, typically AoORA = AoOA. Therefore, gains are mistaken for AoOA (the segmentation factor) when it's actually AoORA (the scoring factor) that's responsible.

2 similar clean young profiles with similar AoOA and other statistics have significantly different scores, why? One started with a card and the other started with a year loan. AoORA DOES make a difference. Which do you think had the higher score? The one who started with a revolver. 


G. AoOOLA-Age of Oldest Open Loan Account is a scoring factor under study-


Negative reason codes point to this metric on some variants. LinkLink. 

H. AoOLA-Age of Oldest Loan Account is a scoring factor under study-


Negative reason codes point to this metric on some variants.  Link. 


I. AoYA - Age of Youngest Account (Number of Months since most recent account opening)


Scoring Factor. AoYA = AoYRA, unless the youngest account is a loan. Appears to award points at thresholds that are multiples of 3, up to 24 mo? AoYA is a scoring factor and therefore directly gives/takes points; AoYRA is a segmentation factor and affects scores indirectly via scorecard reassignment.  Different categories too! (Kinda like AoOA and AoORA.)


This metric appears to be the source of reported gains at 3/6/9 months AoYA on young files, NOT AoYRA. Notice it does not discriminate against loans, as the "New Credit" category AoYRA segmentation factor does. 




4. NEW CREDIT CATEGORY (No New account/New Account)






A. AoYRA - Age of Youngest Revolving Account is a segmentation factor in clean profiles. (Formerly wrongly thought to be the same as AoYA.)


For Score 8, if you have 0 revolvers under 12 months of age, you are in a "No New Account" scorecard. If you have a revolver under 12 months of age, you're in a "New Account" scorecard. All other things constant, There’s typically a ~10-15 point difference. 5/4/2 under study, Contribute DPs!


This is one cause of losing points when a new revolver reports. If you already had one < 12 months, you'll only see changes related to AAoA, AAoRA and utilization when it reports; however, if you had 0 < 12 months of age, you wear a significant penalty as you reassign scorecards, plus the potential changes mentioned, plus the initial HP, discussed soon.


There is debate as to whether this metric includes loan accounts; however, my testing, link, and that of others has conclusively demonstrated to this author's satisfaction loans are not included in AoYRA. AmEx chargecards, don't know. Someone who gets one and doesn't have a revolver under 12 months, please contribute!


The metric AoYA includes loans and is a scoring factor, but is in the Length of History ingredient. 


B. Inquiries


Please read the Inquiry Master Thread. (I'm skipping an exhaustive examination of inquiries because the information is in the linked thread, you need to read it. It also goes over some scorecard info.)


- HPs may count for 0-20 FICO points each, though they are typically <5 points on mature/thick profiles, and higher for young/thin cards. (MyFICO Contributor Cassiecard experienced a 20 point drop for one inquiry on a young/thin scorecard. Proof. )






I believe they cost more on dirty profiles, too. ~ the 9th or 10th inquiry, is a saturation point and there's no further penalty. 


- Hard Inquiries are believed to be “binned.” This means there may be a score loss for the 1st, but not the 2nd, maybe for the 3rd, but not the 4th, etc..  Exact bins are not known and may vary by scorecard.


- Inquiry penalty points are returned at 365 days, unless it falls in a bin. Inquiries will remain on your report for a up to 25 months, but are only scoreable for 1 year.


The purpose of a HP rather than a SP from a creditor is to put the world on notice of your credit seeking behavior, so as to slow your roll and protect their interests, so that you do not overextend yourself or so that at least other lenders have their eyes wide open in lending to you. So be glad when you get an SP CLI.


          i. Buffering/De-duplication of Installment Inquiries


The point loss is immediate for most non-installment HPs, but FICO ignores installment HPs (IF coded correctly) from the preceding 30 days. (Buffering.)

Installment HPs of the same type within 45 day windows are counted as 1 for scoring purposes by the algorithm (14 days for EX2). This is referred to as de-duplication and is designed to allow for rate-shopping. De-duplication does NOT combine installment inquiries across types. A mortgage pull and an auto pull will not be de-duplicated. But 10 auto inquiries within 45 days will only penalize you as if it were 1, scorewise.








Please note that when applying for CCs, most lender computers simply see the raw number of inquiries, not de-duplicated. This causes auto denial for inquiry-sensitive lenders. A solution is not to apply to lenders that do not allow reconsideration (looking at you CapOne). If you apply and are denied for too many inquiries (credit-seeking), a quick call to UW explaining the multiple inquiries are from rate-shopping the same loan will usually cause manual reconsideration.


Soft inquires are inquiries done for various reasons, have no scoring impact and can only be disclosed to the consumer. Examples include: promotional, AR (account review), consumer disclosure, insurance, employment. The type of SP determines the data given.

Promotional inquiries, for instance, do not give account information, just contact demographics. AR gives everything except SPs and consumer disclosures give everything for example. 






C. Spree Penalty "Too many accounts recently opened" reason code 


This is a scoring factor that considers how many revolvers < 12? months old exist. Details under study. The below are data I've gleaned: 

- too many new accounts opened within {0, 30, 60, 90?} days of opening an account may cause an additional "spree" penalty. This penalty is believed to cease at 12 months - CassieCard - but all the details are not known. Probably doesn't affect mature profiles.


Note: "New Account penalty points" are from scorecard reassignment WHEN a new revolver reports and you had no revolver under 12 months of age upon it reporting. (AAoA, AAoRA, AoORA, and utilization changes all factor in at once along with everything else upon scorecard reassignment.) Any losses from subsequent revolvers reporting while having one <12 months of age are from AAoA hits, AAoRA hits, and/or balance/utilization changes; these changes also occur in the former situation upon the first card reporting, but because the algorithm accounts for them in one swoop, it is impossible to determine exactly how much of a changes each factor contributes. However, if no aging or utilization thresholds are crossed, that makes educated guesses easier. 


There is no new account penalty for a 2nd or subsequent card within 12 months, though losses may come from other metrics (like the spree penalty). 


Getting another revolver 6 months later might not cost you a new account penalty, but extends the time you are under it. As a result, I recommend getting what you need in 12 month cycles before the first one reports, as you now have a 12 month penalty and no use in making it longer. Plus you may as well have the best score when you do app, so waiting 12 months allows that award, plus it lets your HPs and new accounts to age and the spree penalty to reset.


Last edited 2.23.21 5:15pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 4 of 342
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20







A. Number of accounts is a segmentation factor. (Thick/Thin)


If you have 4 TLs on your CR, you are believed to be in a thick scorecard in Version 9, and maybe all versions were not certain. A thick profile is preferred, penalties are less severe and the score is therefore more stable. Number of TLs has other effects, see 3, below.



B. Mix Diversity


For FICO, there are 6 recognized account types (and having at least one of each in the first two categories on your CR is known to satisfy the "Diversity" scoring factor and give bonuses; it appears the addition of a third category or a sufficient number of bankcards could give additional Mix Diversity points.


The fifth category is not known to give any additional bonus, and the sixth category is negative and penalizes.


i. Revolving (CCs, most PLOCs, HELOCs);


ii. Loans (automobile loans, personal loans, student loans, recreational and vehicle loans, Credit builder loans [SSLs], etc.);


iii. Open-ended (True Chargecards, although these are classified as a separate category for mix, there is no evidence that it offers additional points).


iv. Mortgage Loans Mortgage loans, while installment, are in a category of their own. 


v. Retail Accounts While these are typically revolving, they are in their own category and don't seem to offer any additional points.


vi. Consumer Finance Accounts (CFAs) These are loan tradelines that are considered negative. Finance companies are directed at those with lower credit scores and it is believed each bureau has its own list of those companies. If an institution has the word finance or financial in it, there's a good chance it may be a CFA.

A card/loan on record, open or closed, appears to give a mix diversity bonus, link, 
but open ones can have increased benefits via other categories. 


A lot of people think a mortgage is necessary to get an 850, or that not having one hurts their score. We know from the experts that both statements are false:


Tom Quinn (FICO, VP of Scores):  "There is no characteristic in FICO Scores that penalize a user for not having a mortgage loan." ( Link )


Tom Quinn: "There is no requirement to have an open mortgage to get an 850 score." ( Link )


C. Number of Bankcards


Numer of bankcards is a scoring factor for Mix, but the exact ideal number is unknown. Closed revolvers may count. (TU4 seems to be the most sensitive to a high number of total accounts as a Scoring Factor.


For Score 8, it is believed you are leaving significant points on the table unless you have at least 3 revolvers. Upon acquiring your first cards, points are awarded and the too few bankcard penalty is reduced. The exact amount is unclear as it occurs upon reporting, which may change many scoring factors and may result in scorecard reassignment if it results in AoYRA < 12 months of age upon reporting where it was previously 12+ months.


When the scorecard reassignment results in a drop, its referred to as the "new account" penalty, which can be exacerbated by aging hits. It can be reduced by increased TCL (Total Credit Limit), if it causes thresholds to be crossed. The drop can be partially offset by the "too few cards" penalty being reduced by your first several bankcards or by an award for your first loan (Mix points). Changes in score can also result from the percentage/number of accounts reporting a balance changing, if that changes when a new revolver reports.


Number of bankcards indirectly affects other metrics, as opposed to number of accounts, which is a segmentation factor at all 3 bureaus. The exact optimal number of bankcards is unknown. I recommend no less than 5 revolvers. I believe there may be a small benefit to having more than 5, that can increase to a point. I don't believe there is a scoring benefit to having more than 10. The increased benefits may stop before 10; impossible to know at this point. I think 5-7 is great, but this is jmho. There is an example Scorecard in the first post under Scorecard Basics where you can see an example of # of bankcards and resulting points.


D. Revolver:Loan Ratio


Revolver:Loan Ratio is a scoring factor at EQ8 and although the exact ideal ratio is unknown, it's believed to be 3:1 or 4:1. Link.  Code 84.


E. Account Exclusion/Removal Timeframes


Chargedoff accounts are removed at 7 years. Link.  Accounts with a delinquency have the delinquency age off at 7 years and the account may remain until 10 years, like accounts that are closed with no delinquencies which are "supposed to” remain on CR for 10 years, but not always. MyFICO Contributer SouthJamaica reports accounts being dropped after a year being closed, IIRC.


So, while a closed loan satisfies the credit mix diversity scoring factor, who knows when it will disappear? Plus, as discussed in Amounts Owed Section B, supra, having an open loan with B/L equal to or <9.5 offers 15-35 bonus points.


Here is a great discussion of the details of credit mix by our own esteemed Thomas_Thumb: Link. 




If your identity was stolen and an account tradeline is not yours, you may use the procedure laid out in this link, rather than using a dispute. Otherwise...


A. Direct Disputes


A direct dispute is an option a consumer has to dispute directly with the creditor (furnisher). One would contact their creditor directly and file said dispute. Sometimes this can get a matter corrected without involving the CRAs, which is the distinction.


If you've done a CRA dispute first, do not advise the creditor/furnisher of that via direct dispute or they may summarily dismiss it without investigation. Link. 


B. CRA dispute


A debtor also has the option of filing a dispute with any or all of the CRAs, which are then required to reinvestigate the allegations within 30 days, if they meet certain requirements. The same issue raised successively may be summarily dismissed without reinvestigation. If you submit additional information before the re-investigation is completed, they are allowed an additional 15 days. There's also a 5 day period for mailing. 

if they cannot verify it, they must remove it. If they refuse you can file a complaint with the CFPB or initiate a civil action. However, if they're able to subsequently verify the information it can be reinsertedinserted at a later date. link. 


When you do these by computer or even by hand, if they can scan it in with OCR, it's all handled by computers automatically. That is not to your benefit. In my opinion to give yourself the best chance, hand write your dispute include supporting documentation and mail it CMRRR. (certified mail return receipt requested.) keep copies of everything.


CRA’s must consider supporting documentation. Link. 


C. MOV (Method of Verification) Request


If a CRA dipute comes back denied, as most will, you have the right to submit a Method of Verification request which is essentially a request for how the CRA verified your dispute as accurate. This is nice as your burden in 4 is to disprove the reasonableness of the reinvestigation, so a information  as to how they concluded your account as accurate will go a long way in you showing how the investigation was not reasonable in section D. Check these for more info: Link. Link. 


D. CFPB (Consumer Financial Protection Bureau)/Judicial remedies


If you can't get no satisfaction elsewhere, you may file a complaint with the CFPB or initiate a suit in court contesting the reasonableness of the furnisher's investigation, if you meet the requirements. Link. The CFPB has the responsibility and authority to enforce the FCRA and FDCPA. They may initiate judicial action against a CRA but it is rare. Likewise if they direct the CRA to do something, they are usually motivated to do it, yet they may in some cases filed suit against the creditor to enforce the CFPB's order. Link.  (Consult qualified counsel).





A. Locking


Congress passed legislation requiring CRAs to allow consumers to freeze their CRs without charge. While frozen, no HPs are possible, BUT new accounts CAN be added. Consequently, a new line of credit is unlikely and ID theft is reduced because most creditors will not extend credit absent a HP. Another advantage is, it may cause one to rethink applying for credit, as it adds another step.


B. Freezing


The CRAs, obviously wanting to avold the legislation, devised an alternate method, "locking" your CR. They monetized this and marketed it as easier than freezes, despite the fact that it is no better than freezing. In fact, locking affords the consumer less protection, as it is not subject to the legislation's protections and requirements for freezes. I believe all but Experian now offer locks for free.


C. Fraud Alerts

If you place a fraud alert on your profile, a lender will not extend credit to you until they verify your identity by contacting you via the telephone number on your credit report.


If you do not have a telephone number on your report, you're going to get a hard pull for nothing. So don't place a fraud alert unless you have a current number on there or don't apply for credit and expect mote than a hard inquiry, if you don't have a number there where you can be reached and have a fraud alert active. 

D. Identity Theft Exclusion


If an account appears on your credit report as a result of identity theft, the FRCA provides a remedy. If you file a police report alleging you did not authorize the account, it is a result of identity theft and obtain a police report, the CRA's are required to exclude it from your credit reports. This remedy does not involve the creditor or involve any contact with the creditor. The police report is provided to the CRA and they exclude the account. Link. (Credit: The esteemed Legendary MF contributor RobertEG.)



8. SEARCHING THE FORUM (Search Secrets)


Sorry, but the built-in forum search software leaves much to be desired and even more to be found. But not to fret, Google can search just this site for you and even filter it by time period as well as keywords. Try it out, bet you'll find more than you thought.


Here's an example by @CassieCard sharing with the Community how to do so:


General: site:[domain] [keywords] after:YYYY-MM-DD before:YYYY-MM-DD
Example: scoring primer after:2019-01-31 before:2020-10-08
Exact Match Keywords: "lost 20 points" after:2020-10-01
Exclude Keywords (boost): experian -boost before:2020-09-28

No spaces next to any colon (Smiley Happy
Search a range of scores for recent approvals: 630..680 approval after:2020-10-01
(Uses Google range operator, dot-dot between numbers.)


Play with it, make a macro, but search the site with a deep search and limit it by date and keyword for relevance. Oher tricks might come....if those DPs start rollin in....Smiley WinkSmiley Happy


Edited 2.23.21 5:17pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 5 of 342
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20



Negative reason codes are generated by the fico algorithm at the same time as your score. These codes give a window into the reason for your score, the reason for score changes, and how you can improve your score. They are listed in the order of precedence:




The following shows an example of how the algorithm generates the negative reason codes:




Positive reason codes are meaningless per existing knowledge. It seems as if they are the negative codes reversed. The only things you know you can use to reliably tell what the algorithm itself is doing it seems are score changes and negative reason codes, as we know they are the direct output from the algorithms.


Negative reason code charts are helpful, as they tell you which codes do or do not exist in versions and variants. A great deal can be determined from this. However, keep in mind that all negative reason codes for a particular version do not exist in every scorecard of that version. This is one of the clues as to what scorecard you may be in.


Here's my theory for what it's worth: FICO starts with ~500 characteristics and they run their math on the datasets at the bureau. It tells them the 40 most predictive characteristics and how they should be weighted among the sub-groups.

This creates the scorecards for that version at that bureau. That's why data can be identical across bureaus and result in a different score. Each bureau's dataset was different, so the resulting scorecards are different and have different characteristics being measured.

We see that one bureau measures certain characteristics where others don't and there's several instances of this as can be seen in Cassie's Scoring Factor thread in the differential post.

I believe each negative reason
 corresponds to one or more characteristics being measured. For instance, the negative reason code that refers to the oldest or average age of your revolving accounts is signaling that one or both of those characteristics is not in the optimal range, and therefore improvement could award more points.

CassieCard has created html reason code tables for you at:





The mortgage Scores react very differently than 8/9. The mortgage scores were actually the general purpose score for the version when they were issued.


For example, EX2 is based on the 1998 version and is used as Experian's mortgage Score. It also has a bankcard and auto Version which is used. 

TU4 & EQ5 (and EX3) are based on the 2004 version and likewise have auto and bankcard variants.


So Experian's mortgage score is going to react differently than TransUnion and Equifax's because they're based on different underlying versions. Even with that said, TransUnion and Equifax's scores still have differences even though they were created from the same version, because their scorecards were generated based on the datasets from the respective bureaus, which differ.

However we know there's going to be a bigger difference between the Experian Score versus the other two because they are based on different versions. One thing about all the mortgage Scores is the number of accounts with a balance is a very important metric, more so than utilization. Version 8/9 made utilization more important, that's a big difference between the old and newer scores.


Another important feature of the mortgage scores is they have a dollar amount cutoff where revolvers over that amount are not factored into utilization & balances. This can cause you to run into unintentional AZ loss by only having a balance on a revolver above the cutoff amount. The Experian cutoff is different from the other two which is the same because they're based on the same 04 version. 

Why are we using such old scores for mortgages? Well because Fannie Mae and Freddie Mac require mortgages to be underwritten on those scores I believe in order for them to purchase them, so all lenders therefore want them to qualify as my understanding. 

The best thing you can do for your mortgage Scores is AZEO. Authorized user accounts always count full monty, just like a primary account. 

Mortgage scores also have a higher threshold for new accounts segmentation, meaning new accounts hurt for a lot longer. We have an initial report that EX may have a 17 month threshold. Still searching for the other two and we also don't know whether it's just revolvers or if it also includes loans on the mortgage scores.


Would love data on that as well. How can you help? Look at your negative reason codes and see if you have new account there. Whenever it drops off and you have a decent score change and reason code shift on a 1st, let me know because you may have experienced scorecard reassignment. Inquiries seem to cost more there too, I believe. 

Another thing, the young/mature segmentation point is at 2 years rather than 3 on the mortgage scores. We're not sure about the thick/thin line, but we do know derogatories are treated differently because they only had 2 dirty Scorecards then as opposed to 4 starting with version 8/9. 

My theory is the mortgage scores had a public record scorecard and a delinquency scorecard, and Version 8/9 broke those into recent and mature. It appears on the mortgage scores after 2 years, you went back to a clean card with penalty whereas you stay in dirty cards for 7 years on the newer scores. 

I'll try to add to this some more but just wanted to give some general information to start building it up. Suggestions and comments are welcome. 

Good discussion by TT of differences among bureaus relevant to a old mortgage loan being paidoff and closed. (Only loan.) Link. 

CFAs can cost ~20 points across bureaus on PR card and ~40 points on a delinquency card, according to: link. 

A. EQ5


-not sensitive to lack of open loan.


-very sensitive to number of accounts with a balance. -TT Link. 

-revolvers with CLs of $35,000 or higher are excluded and should not be used for an AZEO card. 




B. TU4


-sensitive to total number of accounts. -BM7


-revolvers with CLs of $35,000 or higher are excluded and should not be used for an AZEO card.


C. EX2


-revolvers with CLs > $31,000 may be excluded and should not be used for an AZEO card. (Cutoff is over $31,000 but below $34,900.)-TT


-6 year/72 month AAoA threshold on clean/thick/mature/no new account scorecard, +7 for classic, +15 for auto, +7 for BCE. Link. -BM7


-sensitive to # of accounts with a balance. -BM7


-Closing an old mortgage had a strong impact on auto, but not Classic or BCE. -TT Link. 

-has a 30% threshold for number of accounts with a balance, but the denominator includes all accounts including closed accounts. (Trudy) Link. This is not the lowest threshold. 



So I figured it was time to add a section for version 9, since many banks have adopted it and there are plenty of questions about it.


Version 9 was built to be an improvement on version 8. We do know all the segmentation points for version 9, except the additional 13th Scorecard that was added for high utilization. That we do not yet understand.


Segmentation points are the same as version 8, except that we are sure of the file being segmented as thick with 4 accounts on version 9. I would argue we are not totally sure for the other versions. So 60 day or worse put you in a dirty card, 3 years makes you mature, 4 accounts make you thick, and 12 months AoYRA, you go to a no-new scorecard.


So what's different what's new? Paid collections are now ignored in addition to nuisance collections which are those under $100. Medical collections are weighted less. A more recent dataset was used to create version 9.

Many of the changes are opposite that of version 8 and are counterintuitive. For instance, the new accounts reassignment point change seems to go in the opposite direction from version 8.


I will add to this as I think of other things or as we learn other things. Please contribute data points, corrections, or suggestions.


Edited 2.23.21 5:18pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 6 of 342
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20

12. Bibliography 


Aging delinquency strings versus monthlies:

Attribution for: "...Experian...excludes...delinquencies in a common “string” 7 years from [DOFD]...The other two CRAs have no official, published policy interpretation, and have...excluded based on [DOFD] OR have treated each...delinquency as its own separate adverse item of information, and thus have not excluded...until each has reached...7 years..."

[1]. RobertEG Jan 2020, Permalink. 


Attribution for "installment loans can give points over time irrespective of thresholds"...."As mentioned up thread and shown by a graph, loan age is an actor that influences score outside of B/L ratio on installment loans..."

[2]. Thomas_Thumb May 2017, Permalink 


Attribution for "Utilization less than or equal to one of 9%, 29%, 49%, 69%, or 89% works just fine to avoid the next higher interval."

[3]. CassieCard Feb 2020, Permalink 


Attribution for: AoOA is not a Fico scoring factor. It is a scorecard assignment segmenter. If an increase moves you to a different scorecard your score may change due to a shift in weighting of the factors used in scoring and the assigned min/max scores associated with the scorecard.

[4]. Thomas_Thumb Jan 2018, Permalink 


13. References/Permalinks:


more about Scorecards:

Common abbreviations:


Adding an installment loan -- the Share Secure technique:


May 2020 example of SSL execution:


NASA FCU $1500 60 month SSL:

Balance threshold @ $147?


Tracking the First Year with Credit Cards:


All Zero Penalty to AZEO (credit angelwingz):


Evidence/References for "AU cards at Zero = Separate Penalty"


If you are just starting your credit:

HOWTO: Get a +62 point FICO 8 jump on a clean file with thanks to @CassieCard


An Application of Credit Scoring: Developing a Scorecard Model :

(An introduction to logistic regression and WOE/IV tables, with R code.)


Installment loan thresholds by Revelate:  


Thomas_Thumb new credit points expected:  


Example of 20-year-old hitting 800 with two years history:  


SJ utilization breakdown regarding Statement balance timing:  


PFDs can be done with OCs or CAs:  


Charge off required at 120 days on loan and 180 days on revolver:  


Explanation of how a charge off can drop score upon being paid if not regularly updated:


Explanation of the difference between payment status and current status:


Discussion of good secured cards:  


CAs that do PFD's:


Which banks pull which CRA  


3-year threshold for score 8 AoOA  


2 year threshold for EX2 AoOA  


A balance change can affect score a couple days before it’s reflected in the CMS front end.  


Loan modification is a serious derogatory in fico 9


Accounts in dispute removal for mortgage  

AmEx 3X CLI thread


Explanation of why you cannot rely on values calculated and displayed by CMSs' frontends AKA "Fluff" (Credit to BBS for coining the term, I believe):

800+ is possible with 60 day derogatories

List of links to soft pull preapprovals


Miscoded INQ causing 30 day buffer trigger from cc app?


850 buffers discussion



Utilization re: closed cards


Bankruptcy to 700 in 24 months!


A PR scorecard appears more sensitive to utilization compared to a delinquency scorecard


AAoRA threshold found at 9 years  


Adding a second and third card and the SSL and Paydown


12 mo threshold for EX2-new still looking


1.5 year average age threshold


CA May not add interest or fees unless permitted by the original debt agreement

Notice requirements for service of process/default judgments without personal service

Last edited: 2.23.21 5:20pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 7 of 342
Super Contributor

Re: General Scoring Primer and Version 8 Master Thread Pub.5.17.20








Where did this come from?

Quick Reference

Brief background

FICO scoring and these forums

FICO Score 8 key differences

Rating/Score range

So, how does scoring work in FICO Score 8?





Payment History



Amount of Debt



Length of History



New Credit



Credit Mix


Start Birdman7

Scorecard Basics




1. PAYMENT HISTORY CATEGORY (Clean/Dirty)          

A. Derogatory Categories

B. Collections/CA/Dunning notice/DV

C. Aging Derogatories

         i. Scoring

         ii. Derogatory Removal

         iii. CR Removal Quick reference

         iv. Judicial SOL (Statute of Limitations)

         v. Various

  D. Amount Still Owed on Delinquent Accounts
  E. Multiple Delinquency Penalty

  F. Recency

  G. Number of Accounts Paid as Agreed




2. AMOUNT OF DEBT CATEGORY                                     

          A. Utilization

                 i. Revolving

           a. Aggregate
           b. Individual
             ii. Loan

          B. Percentage/Number of Accounts Reporting Balance/AZEO

          C. All Zero Point Loss-(AU Test included)

          D. CO

          E. Revolving Balance Metrics-Aggregate and ABORT

          F. Aggregate Loan Balances metrics

          G. Amount Owed on All Accounts Metric





       A. AoOA-Age of Oldest Account is a Segmentation Factor in clean profiles

B. AAoA - Average Age of Accounts is a scoring factor

C. AAoRA- Average Age of Revolving Accounts is a scoring factor- under study

D. AAoOLA- Average Age of Open Loan Accounts

E. AAoLA-Average Age of Loan Accounts is a scoring factor- under study

F. AoORA-Age of Oldest Revolving Account is a scoring factor- under study

G. AoOOLA-Age of Oldest Open Loan Account is a scoring factor- under study

H. AoOLA-Age of Oldest Loan Account is a scoring factor- under study

I. AoYA-Age of Youngest Account is a scoring factor- under study


4. NEW CREDIT CATEGORY (No New Account/New Account)   

A. AoYRA - Age of Youngest Revolving account (Segmentation Factor)

B. Inquiries

              i. Buffering/De-duplication of Installment Inquiries

       C. Spree Penalty - "Too many accounts recently opened" reason code





A. Number of Accounts (Segmentation Factor)

B. Mix Diversity

      i.  Revolving

     ii.  Installment Loans

     iii. Open-ended

     iv. Mortgages

     v.  Retail Accounts

     vi. CFAs

C. Number of Bankcards (Revolvers)

D. Revolver:Loan Ratio

E. Account Exclusion/Removal Timeframes


6. DISPUTES                                                                

A. Direct disputes.

B. CRA disputes

C. MOV Request

D. CFPB/Judicial remedies 



A. Locking

B. Freezing

C. Fraud Alerts

       D. Identity Theft Exclusion

8. SEARCHING THE FORUM (Search secrets from CassieCard)





 With Link to Cassie's Score Factors Thread (Has Negative Reason Code Charts)



      A. EQ5

      B. TU4

      C. EX2








13. References/Permalinks



14. TOC/Index                                                                             




Post 10- Additional helpful info!


Return to beginning of thread. 





A very special thanks to @CassieCard , my Technical Advisor, for assisting with technical issues, tables, images, research, attribution, adding great links and information, the presentation of this thread, encouragement and so much more. Would not have been anywhere near as nice, comprehensive, and robust without her help. Thank you! And thank you for your continued help in keeping this updated and a wonderful resource for all members. 


The above posts are from what I have learned from this forum and from my own testing. They represent the best knowledge we have, but that doesn't mean there may not be errors. There is quite a bit we still don't know and probably never will. Nevertheless, I have done my best to present the best information possible IMHO.


If you find errors, please LMK. If you think something should be added, LMK. If you think something is unclear or needs clarification, LMK.


Thank you to all our members for helping each other and keeping this forum strong!


Also, I would like to give proper attribution for theories and discoveries. Please LMK if one of the findings is yours. If so, LMK and provide the link and I will add it.


To be added:


1. Any requests?




Last edited: 2.23.21 3:21pm. © 2021, Birdman7. All rights reserved.

-Our Community’s updated scoring wisdom: Link to Scoring Primer.
-For Negative Reason Codes see: CassieCard’s Score Factors thread.
-ccquest’s workbook to calculate metrics for you: Link to Workbook.

Correct Ag.Util. under 5% all times. (Oldest/avg varies. Estimates above.)
Real world mortgage maxes are: EQ5-818, TU4-839, EX2-844.


(Everything said is JMHO and is not endorsed by FICO or MF. I have no affiliation with either, just a grateful member.)
Message 8 of 342
Moderator Emeritus

Re: General Scoring Primer and Version 8 Master Thread rev.5.17.20

Fantastic and thank you and also everyone else that has helped put it together or contributed.


Couple of things:


You define some acronyms inline but others like TL aren't; actually it might make sense to just make an acronym list as well, maybe the important ones and then link to the forum acronym list if it can be found again haha.  Probably in helpful threads pinned somewhere.


Speaking of information organization, might I suggest a table of contents and then break things into sections in individual posts akin to the Helpful Threads posts when we consolidated stickies?  Also collect like information, aging and associated metrics metrics on one page, derogatories on another, derogatory remediation separate... they are all kind of jumbled together now or at least that's how it read to me on first glance.  Apologies I only pick on such things because I am having to write a pseudo ISO compliant document right now for one of my gigs and structure matters.


Actually on second read it isn't as muddled as I thought but it's written in essay format, you might want to switch some of the order up like put the optimal values front in center near the top instead of at the end where it lands today.  That might allow to put more detailed information which would lose the casual reader further down thread.


I think we know some additional details for scorecard segmentation, vis a vis lates segment different than public records seemingly and actually a 30D late is a minor delinquency and sorts top 8 in every FICO model I have seen.


Did I miss a reference on FICO 8 patterns?  Namely: "With too many 30 day lates within a period of time, you’ll go to a dirty card. It’s unknown how many it takes, though." ?


I will try to read through more carefully later, in my own documentation hell right now and I need a nap too.

Message 9 of 342
Senior Contributor

Re: General Scoring Primer and Version 8 Master Thread rev.5.17.20

Score Factors / Reason Statements with explanations are in a separate Score Factors topic.


There are explanations for all reason codes and associated statements on EX 8, TU 8, EX 2, TU 4, and EQ 5.

28 FICO Scores + FICO RESilience Index + 3 VS3. MTG (Mortgage), AUT (Auto), and BKC (Bankcard) are scores 5,4, and 2 from the top.

Keeper of the Scrolls of Resilience: EQ FICO Resilience Index Score Leaderboard

55 forum members have shared this new credit score. Share yours today!

Message 10 of 342
Advertiser Disclosure: The offers that appear on this site are from third party advertisers from whom FICO receives compensation.