Length of Credit History - Negative Factor

[Anonymous]

---

RobertEG wrote:

The best statistical answer that I can offer, without pure speculation, is based on the report that FairIsaac made before the FTC back in 1999, which gave some data on how their model considered amount of time credit has been in use.  In general, they reported the following risk factors for the length of time from the oldest revolving trade line. Pts are assumed as deducted from the 127.5 max for this category.  However, this category also considers avg time for revolv accounts, which they did not rreport to the FTC,  so the 127.5 is probably not the real baseline for this sub-category in thier algorithm, but it at least reflect the upper limit, and the percentages of whatver that upper limit are should still apply.  It is probably lower, because avg time is also a factor. Yet the linear progression is obvious.

0-2 yrs,   around 90% risk, and thus hit of 90% of the 127.5  in this category, which would be -120

2-4 yrs    around 80% risk, and thus a hit of around 100 pts in this category

4-6 yrs    around 56% risk, and thus a hit of around 70 pts in this category

6-10 yrs  around 33% risk, and thus a hit of around 40 pts in this category

10+ yrs   around 10% risk, and thus a hit of around 10 pts for this category.

The percentages are interpreted from the FairIsaac graphs presented to the FTC, and the neg pts are extrapolated by me from that data, based on 127.5 pts for this overall category.

This data is old, but the best that I think FairIsaac has publicly reported.

Message Edited by RobertEG on 12-05-2007 07:17 PM

---

You may want to rethink your point values here.  You seem to have arrived at 127.5 by using 15% of 850 (the highest score).

Remember, however, that the lowest score is not 0.  It's 300.  That means there are only 550 points "to be had" overall.  And 15% of 550 is 82.5, not 127.5.

Also, I would be shocked if the algorithm turned out to be this straightforward.  Besides being extremely old, I would venture that the data Fair Isaac presented to the FTC are also extremely simplified.

[Anonymous]

One more thing I forgot to mention:

What you have spelled out here assumes that all of the points in the "Credit History" category are determined by the age of one's oldest TL.  In reality, that is just one factor in that category.  Here are the factors that myFico says are included in that category:

Length of Credit History

• Time since accounts opened
• Time since accounts opened, by specific type of account
• Time since account activity

I understand that your calculations are just attempting to generate a "rough guess," but frankly, IMO it's a futile undertaking.  You are doing the same thing that generators of FAKO scores do - attempting to extrapolate how FICO might score a certain set of data - and we all know how useless those are.  Even a rough guess is useless if it can be off by more than 40 or so points in either direction.

I'm not trying to be overly critical here.   I guess I just don't understand the usefulness of attempting to replicate a complex algorithm without even accounting for the most basic data points.

[fused]

---

RobertEG wrote:

Remember that each FICO category has a fixed percentage, and thus a fixed max..  You only lose less from the 127.5 for lenght of credit as time increases, compared with  prior negatives.  So you never become "positive."  The postive is only the increase over the prior negative.  FICO is a negative algoritthm.  So, accentuate the postive, eliminate the negative, and dont mess with mister in between!   Dat be da name of da song!

---

Yes 127.5 is 15% of 340 but this number is very misleading. By using this number it's assumed that many can and will achieve a FICO score of 850. This is simply not the case. IMO, I thinks it better we use the number 51 (15% of 815) instead of 127.5 and here is why. 95% of all consumers who have a FICO score will fall between 475-815. In reality the vast majority of us cannot earn more than 51 points with respect to credit history length and average-age of accounts combined. This same thinking applies to FICO's other scoring factors: payment history, util, new credit and credit mix.

Message Edited by fused on 12-09-2007 03:44 PM

[haulingthescoreup]

---

@fused wrote:
IMO, I think it better we use the number 51 (15% of 815) instead of 127.5 and here is why.

---

You meant to say "15% of 340", right? (815-475=340) Sorry, not to be quibbling, it's just been a loooong day, and I got myself tangled up on this the other day.