Points gained/lost for utilization%

[Anonymous]

Hi I came across this in my research

According to the Fair Isaac Scorecard(slide 8 0f 49), it goes as follows

 >50% = -18 ;
41 - 50 % = -10;
31 - 40% = -3;
16 - 30%= 5;
and 0 - 15% = 15

one ? is this considering only the collective total? meaning that you lose only 18 pts if your total util is over 50%

or you losing any points individually as well? like a double wammy...

one card is 51%  minus 18

and one is 52% minus 18

So, would you be losing 18 points times 3?

Also, has anyone seen data to support what type of points you are losing based on 61% usage and 71% etc? I am worried as I have one card that is at 71% that I really can't comfortably pay down at this point.

I was told that you could lose b/w 70-100 points being maxed out.

Message Edited by takemehigher_1 on 11-12-2007 05:12 PM

[inquring_minds]

Interesting! - where did you find this scorecard?

[Anonymous]

I believe that is an old slide-
I believe that you are getting hit for EACH one over 30% and would be consistent with my experience-

I have pulled enough scores to show a consistent 12-15 points for each 10% of utl and the same 15-30 points for each acct over 30% UTL.

Mine was about a 60 point difference between 40% utl and 19 utl. 9 Revolving cards in that mix.

---

@Anonymous wrote:

Hi I came across this in my research

According to the Fair Isaac Scorecard(slide 8 0f 49), it goes as follows

>50% = -18 ;
41 - 50 % = -10;
31 - 40% = -3;
16 - 30%= 5;
and 0 - 15% = 15

one ? is this considering only the collective total? meaning that you lose only 18 pts if your total util is over 50%

or you losing any points individually as well? like a double wammy...

one card is 51% minus 18

and one is 52% minus 18

So, would you be losing 18 points times 3?

Also, has anyone seen data to support what type of points you are losing based on 61% usage and 71% etc? I am worried as I have one card that is at 71% that I really can't comfortably pay down at this point.

I was told that you could lose b/w 70-100 points being maxed out.

Message Edited by takemehigher_1 on 11-12-2007 05:12 PM

---

[Anonymous]

Now is this point boost on a monthly basis...

something tells me that would be too aggressive to the whole scoring system.

how often would the points fluctuate?

30, 60, 90, 120, 150? every 365???

[Anonymous]

I don't think you would get a points boost on a periodic basis - only when the situation for each card changes.  So if one card is over 50%, then your score is adjusted down by 18 pts.  If it stays over 50%, no change to your score.  But if it drops to 1%, then instead of -18, you get +15 for that card.

There's got to be a limit on how many cards factor into this.  Say if you had 3 cards that went from >50% (-54pts) to 1% (+15) - that would cause a normally 700 score to fluctuate from 646 to 715.  But if you had 10 cards, would that mean your score would fluctuate from 520 to 750?   If so, then more cards is better (as long as you keep them under 10%)?  I would think there's a limit to the boost that you would get.  (BTW - I know utilization only contributes 30% of the score, so the actual figures wouldn't match the above, but trying to keep it simple here )

Message Edited by WAFICOwatcher on 11-13-2007 09:19 AM

Message Edited by WAFICOwatcher on 11-13-2007 09:27 AM

[RobertEG]

FICO score calcuation is a pure mathematical algorithm once data is entered, so a clear mathematical view of this calculation is required.  And yes, the slide referred to is a very old slide, and does not discuss what baseline a "gain" or "loss" is calculated from. It thus has no disclosed baseline for the "loss" or "gain" approximations given, and accordingly has no clear mathematical basis for its intepretation, which is probably why FairIsaac no longer uses this in its discussions.   It makes mathematical sense (and FICO is a mathematical algorithm) only when compared to some defined baseline from which any increase or decrease is compared.  In reality, total FICO %util points are 30% of 850, or 255 points max.  So the real baseline is 255 points, which is a max, and thus points are never gained for %util if total points are the baseline, which they are....., they are only lost from the 255 max.  For absolute scoring changes, it is the amount of loss from max that is the true mathematical issue, not any gain from some other baseline.  While the exact algorithm used by FICO is a trade secret, if one assumes a simple linear scale, then points at each level would be:

10% util     255-25 = 230 positive, and of course, about a 10% loss from max of 255

20% util     255 - 50 = 205 positive, or about 10% lower score than 10% util of 230

30% util    255 - 76 = 179 positive, or about 12.1% lower score than 20% util score of 205

and so on...

If the scale is not linear, then the differences will also be non-linear.  The point is, you only "gain" over a prior baseline, and not in the absolute sense.  Also, there seems to be some feeling on here that FICO deducts more points if you are at 0% than if in the 0-9% range, and thus showing some utilization.  Who knows?  Only Fair Isaac!   And also, 255 is for all util, which incudes both installment and revolving.  Your guess is as good as mine as to how much is weighted for revolving, and how much for installment.  Clearly, revolving is weigthed much higher, but the weighting is a FICO secret.

Message Edited by RobertEG on 11-13-2007 03:44 PM

Message Edited by RobertEG on 11-13-2007 03:50 PM

[Anonymous]

---

RobertEG wrote:

FICO score calcuation is a pure mathematical algorithm once data is entered, so a clear mathematical view of this calculation is required.  And yes, the slide referred to is a very old slide, and does not discuss what baseline a "gain" or "loss" is calculated from. It thus has no disclosed baseline for the "loss" or "gain" approximations given, and accordingly has no clear mathematical basis for its intepretation, which is probably why FairIsaac no longer uses this in its discussions.   It makes mathematical sense (and FICO is a mathematical algorithm) only when compared to some defined baseline from which any increase or decrease is compared.  In reality, total FICO %util points are 30% of 850, or 255 points max.  So the real baseline is 255 points, which is a max, and thus points are never gained for %util if total points are the baseline, which they are....., they are only lost from the 255 max.  For absolute scoring changes, it is the amount of loss from max that is the true mathematical issue, not any gain from some other baseline.  While the exact algorithm used by FICO is a trade secret, if one assumes a simple linear scale, then points at each level would be:

10% util     255-25 = 230 positive, and of course, about a 10% loss from max of 255

20% util     255 - 50 = 205 positive, or about 10% lower score than 10% util of 230

30% util    255 - 76 = 179 positive, or about 12.1% lower score than 20% util score of 205

and so on...

If the scale is not linear, then the differences will also be non-linear.  The point is, you only "gain" over a prior baseline, and not in the absolute sense.  Also, there seems to be some feeling on here that FICO deducts more points if you are at 0% than if in the 0-9% range, and thus showing some utilization.  Who knows?  Only Fair Isaac!   And also, 255 is for all util, which incudes both installment and revolving.  Your guess is as good as mine as to how much is weighted for revolving, and how much for installment.  Clearly, revolving is weigthed much higher, but the weighting is a FICO secret.

Message Edited by RobertEG on 11-13-2007 03:44 PM

Message Edited by RobertEG on 11-13-2007 03:50 PM

---

Very thorough research.

So basically we're a slave to these formulas and we just have to cross our fingers and pray we're doing the right thing.

[fused]

Interesting stuff folks but I disagree a bit on the specific numbers. Although FICO scores range from 850-350, few have a score over 800 and VERY few have scores below the 425-450 range. It might be better if we look at the group which represent 95% of all scores, the 450-815 group. In this there is a difference of 365 points to be gained or had. Here's the FICO scoring breakdown rounded up:

Payment HX:                                                                           35%: 128 points to be had

Amounts Owed (Util%):                                                          30%:  110

Length of History (including average-age of accounts):         15%:    55

Types of Credit (credit mix):                                                   10%:    37

New Credit (accounts):                                                           10%:   37

Total:                                                                                                  367 (rounded)

Just my sense of how it works.

Message Edited by fused on 11-13-2007 02:32 PM

[haulingthescoreup]

---

@Anonymous wrote:

Now is this point boost on a monthly basis...

something tells me that would be too aggressive to the whole scoring system.

how often would the points fluctuate?

30, 60, 90, 120, 150? every 365???

---

Scores don't exist until they are pulled. They are just potential. Whatever shows on your CRA's at that moment determines your score (as crunched by the mystery FICO formula.)

If you win the lottery and pull all three every day, they won't change much, because your reports don't change much. If you pull once a year, there's no telling what they might have been on the other 364 days. It just is what it is, when you take a look.

[fused]

---

@haulingthescoreup wrote:

---

@Anonymous wrote:

Now is this point boost on a monthly basis...

something tells me that would be too aggressive to the whole scoring system.

how often would the points fluctuate?

30, 60, 90, 120, 150? every 365???

---

Scores don't exist until they are pulled. They are just potential. Whatever shows on your CRA's at that moment determines your score (as crunched by the mystery FICO formula.)

If you win the lottery and pull all three every day, they won't change much, because your reports don't change much. If you pull once a year, there's no telling what they might have been on the other 364 days. It just is what it is, when you take a look.

---

A snapshot!...just like a still photograph