@Thomas_Thumb wrote:
@SouthJamaica wrote:
@Thomas_Thumb wrote:
@SouthJamaica wrote:
@Anonymous wrote:
I agree BBS; however common wisdom right now says it only takes into account the highest individual utilization revolver, but other common wisdom has been proven false, so dps are good!In my experience that so called "common wisdom" is utterly false.
I am quite certain that the number of high utilization accounts matters, and FICO does not discount those which exceed a threshold but not as badly as the highest.
I am no longer passing along "common wisdom" which flies in the face of experience. Another term for "common wisdom" is "the madness of crowds".
Just the facts from Fico employees- NOT common wisdom.
Fico looks at highest account utilization as well as aggregate utilization % and aggregate balance amount in $$. This has all been available in published reason statements for many years.[links provided below].
The # of accounts above a certain utilization could be a scoring factor but, that is not explicitly mentioned in the Q&A. We do know #/% of accounts with balances is a factor. I think a change in aggregate $ amount could be a possibility in this situation but, it is unclear to me that all of SJ's revolvers (particularly CU accounts) are being counted.
There are a lot of things that are "not explicitly mentioned in the Q&A". That doesn't mean they can be discounted. (a) Some of the questions asked were never answered, (b) not all questions were asked, and (c) some of the answers were vague and almost none were very "explicit".
You have previously suggested that
onlythe highest of the overutilized accounts affects one's scores. I now know that that is simply incorrect.
The number of accounts with unduly high utilization is undoubtedly, in my personal experience, a significant factor.
I hope that you and Birdman7 never find out.
I found out the hard way.
I previously stated that the highest individual card utilization is scoring factor (#3), which was corroborated in the recent Q & A.
https://ficoforums.myfico.com/t5/Understanding-FICO-Scoring/FICO-AMA-Discussion-Thread/m-p/6137799#
Back in 2014 Fico mentioned having a candidate set of over 500 predictors. Their published reason code list is rather short by comparison. Are the models using some scoring factors not on the published list? Probably. However, if it is not on the list, then more data is needed to validate a hypothesis.
Single data points are interesting but repeatability and reproduceability across multiple profiles is needed. Isolation of a suspected scoring factor from other potential factors is not always possible. In those cases even more datapoints are needed.
And what evidence do you have for your hypothesis that FICO discounts all but the highest above-threshold individual utilization number?
@Thomas_Thumb wrote:
@SouthJamaica wrote:
@Thomas_Thumb wrote:
@SouthJamaica wrote:
@Anonymous wrote:
I agree BBS; however common wisdom right now says it only takes into account the highest individual utilization revolver, but other common wisdom has been proven false, so dps are good!In my experience that so called "common wisdom" is utterly false.
I am quite certain that the number of high utilization accounts matters, and FICO does not discount those which exceed a threshold but not as badly as the highest.
I am no longer passing along "common wisdom" which flies in the face of experience. Another term for "common wisdom" is "the madness of crowds".
Just the facts from Fico employees- NOT common wisdom.
Fico looks at highest account utilization as well as aggregate utilization % and aggregate balance amount in $$. This has all been available in published reason statements for many years.[links provided below].
The # of accounts above a certain utilization could be a scoring factor but, that is not explicitly mentioned in the Q&A. We do know #/% of accounts with balances is a factor. I think a change in aggregate $ amount could be a possibility in this situation but, it is unclear to me that all of SJ's revolvers (particularly CU accounts) are being counted.
There are a lot of things that are "not explicitly mentioned in the Q&A". That doesn't mean they can be discounted. (a) Some of the questions asked were never answered, (b) not all questions were asked, and (c) some of the answers were vague and almost none were very "explicit".
You have previously suggested that
onlythe highest of the overutilized accounts affects one's scores. I now know that that is simply incorrect.
The number of accounts with unduly high utilization is undoubtedly, in my personal experience, a significant factor.
I hope that you and Birdman7 never find out.
I found out the hard way.
I previously stated that the highest individual card utilization is scoring factor (#3), which was corroborated in the recent Q & A.
https://ficoforums.myfico.com/t5/Understanding-FICO-Scoring/FICO-AMA-Discussion-Thread/m-p/6137799#
Back in 2014 Fico mentioned having a candidate set of over 500 predictors. Their published reason code list is rather short by comparison. Are the models using some scoring factors not on the published list? Probably. However, if it is not on the list, then more data is needed to validate a hypothesis.
Single data points are interesting but repeatability and reproduceability across multiple profiles is needed. Isolation of a suspected scoring factor from other potential factors is not always possible. In those cases even more datapoints are needed.
This is awesome data! this is great to know the segmentation is the same across CRA's. So as long as we establish a segmentation point, we know for that version it will be consistent across CRAs.
And I think the predictors are what they use to create the scorecards.
@Anonymous