Your Questions are Disingenious

[ Follow Ups ] [ Post Followup ] [ Credit ] [ FAQ ]

Posted by Sean on June 14, 1999 at 09:49:14:

In Reply to: I can see I'm going to have to break this down even more for you. posted by Greg Fisher on June 12, 1999 at 09:37:24:

I submit this quote from your site:

creditscoring.com/letters/tu.htm

"Can you, at least, tell me how I can tell how many to add or subtract to avoid the Emperica score giving that as a reason the score is not higher?"

How am I to interpret this question? The word "perfect" is not contained in it at all. It seems merely to be asking how many revolving accounts a person should have to avoid getting the message that you didn't have enough revolving credit accounts.

Ok, if you have 1000 revolving accounts you'll never get dinged for having too few. And if you have 0, you'll never get dinged for having too many. Happy?

Me, personally, I'm aiming for 7. People who have 7 revolving accounts are least likely to be turned down for new credit (see previously quoted research by Fair, Isaac).

To those people who sincerely ask (instead of just grinding their axe), I'd say they should add only one or remove only one because optimizing that portion of your score would result in your score on other sections to decline.

Now in response to your final question: Do I know all the "right" answers to all the questions? No, I do not. But considering the highest score I've ever seen that a person needs to have is 720 (http://www.fairisaac.com/servlet/SiteDriver/Content/718) then I figure that anything over 720 is "good 'nuff."

Here's a cute quote for you (http://www.intelligentanalytics.com/credit.html)

"Improve the quality of your new account/loan decisions with predictive scoring models based on Intelligent Analytics’ advanced Bayesian methods. Whether you are interested in a prescreen model, a new applicant model, or would like specialized front-end tools for determining bankruptcy, delinquency, or write-off risk using credit bureau data and/or user-defined variables, Intelligent Analytics provides the most accurate and flexible systems available."

What are Bayesian methods, you might be wondering? Start by reading http://www.bayes.com/untech.html and maybe it'll interest you, maybe not.

My impression of the scoring model is that it has software with the Bayesian theorm incorporated and that as the credit bureau information changes the model changes too. That means that scoring is a dynamic process and even if they did provide you with the "right answers" today, tomorrow they might not be the right answers anymore based on the heuristic algorithms of the program.

Follow Ups:

Your answers are guesses, and you are losing track of what you said. Greg Fisher 02:33:46 6/19/99 (0)

Post a Followup

Name :
E-Mail :
Subject :

Comments:
: I submit this quote from your site:

: creditscoring.com/letters/tu.htm

: "Can you, at least, tell me how I can tell how many to add or subtract to avoid the Emperica score giving that as a reason the score is not higher?"

:
: How am I to interpret this question? The word "perfect" is not contained in it at all. It seems merely to be asking how many revolving accounts a person should have to avoid getting the message that you didn't have enough revolving credit accounts.

: Ok, if you have 1000 revolving accounts you'll never get dinged for having too few. And if you have 0, you'll never get dinged for having too many. Happy?

: Me, personally, I'm aiming for 7. People who have 7 revolving accounts are least likely to be turned down for new credit (see previously quoted research by Fair, Isaac).

: To those people who sincerely ask (instead of just grinding their axe), I'd say they should add only one or remove only one because optimizing that portion of your score would result in your score on other sections to decline.

: Now in response to your final question: Do I know all the "right" answers to all the questions? No, I do not. But considering the highest score I've ever seen that a person needs to have is 720 (http://www.fairisaac.com/servlet/SiteDriver/Content/718) then I figure that anything over 720 is "good 'nuff."

: Here's a cute quote for you (http://www.intelligentanalytics.com/credit.html)

: "Improve the quality of your new account/loan decisions with predictive scoring models based on Intelligent Analytics’ advanced Bayesian methods. Whether you are interested in a prescreen model, a new applicant model, or would like specialized front-end tools for determining bankruptcy, delinquency, or write-off risk using credit bureau data and/or user-defined variables, Intelligent Analytics provides the most accurate and flexible systems available."

: What are Bayesian methods, you might be wondering? Start by reading http://www.bayes.com/untech.html and maybe it'll interest you, maybe not.

: My impression of the scoring model is that it has software with the Bayesian theorm incorporated and that as the credit bureau information changes the model changes too. That means that scoring is a dynamic process and even if they did provide you with the "right answers" today, tomorrow they might not be the right answers anymore based on the heuristic algorithms of the program.

Optional Link URL:
Link Title:
Optional Image URL:

[ Follow Ups ] [ Post Followup ] [ Credit ] [ FAQ ]