Testing Credit Card Utilization Thresholds

[Thomas_Thumb]

---

@mowglidude wrote:

---

@Thomas_Thumb wrote:

Yes, in the range of 5 to 10 points. Commonly accepted thresholds for % of cards reporting are:

1) crossing above 30% (or 33%) depending on who you talk to

2) crossing above 50% (some people think hitting 50% is the milestone - not in my book)

 - I saw no change going from 2 of 6 to 3 of 6 cards with balances. However, I did see a drop on EQ Fico going to 4 of 6 (going from 50% to above 50%)

Other thresholds not as widely accepted but with some supporting data include:

1) Going above 20% of cards reporting balances

2) Going above 80% of cards reporting balances

3) Having 100% of cards report balances (profiles with only one card should report a balance on it). The all zero penalty is far greater than 1 of 1 cards reporting a balance.

Side note: Fico also looks at NUMBER of accounts with balances. Accounts are all inclusive with the possible exception of AU accounts. Too many accounts with balances can result in a score penalty independent of % of cards with balances.

---

So it looks like I went below 50% reporting. I can even go down to about 23% reporting and see if that does anything.

---

I think paying off the 3 cards with under $100 balances is a good test. That will have minimal impact on aggregate utilization but takes you below the 33% (and 30%) levels.

[Thomas_Thumb]

---

@Anonymous wrote:

@Thomas_Thumb those could likely be the thresholds you may very well be correct. 

---

@Anonymous 

After looking at my data again, no change 2 or 6 (33%) to 3 of 6 (50%) kind of rules out 40% for me. Also the OP's score change going from 7/13 (54%) to 6/13 (46%) re-enforces 50% as a boundry.

Now I do see score drops on mortgage ficos going from 4 of 6 (67%) to 5 of 6 (83%) and again going from 5 of 6 to 6 of 6 (100%). I do believe 100% cards reporting is a threshold as all my mortgage Ficos took a significant hit going from 5 of 6 to 6 of 6.

[Anonymous]

I definitely think 20%, 50% and 100%; the only question is what’s in between there to me, varying by version of course. And it probably varies by scorecard. 30%, 33%? Would be nice to know.

And the 80%? Cool dp,! I recall a reference to 75% or 88% somewhere I think.

[Anonymous]

---

@mowglidude wrote:

---

---

I'm going to go below 8% aggregate UT next billing cycle for my card. I was thinking maybe try first under 8.499% and see if it rounds down to 8%, maybe like 8.35%. What do you think? I have one more card that posts the next day, if all goes according to schedule, that I can bring that I can pay down, to arrive below 8% aggregate UT.

---

Based on what the Principal Scientist at FICO told us in the AMA, anything less than 9.50% should keep util from the 10% threshold.

He didn't directly say that the algorithm uses the typical midpoint rounding, but he did say that my example 9.02% would round to the nearest integer, or 9% in that case.

(Midpoint rounding is also called round half away from zero or commercial rounding. Anything (x).5 to less than (x+1).5 will be (x+1).0 , where x is an integer or whole number.

I'm setting up my own test for this on 2 cards reporting on the 2nd and 5th. First one will be at 9.450%, and the second at 9.492%.

Unless I got trolled by FICO's principal scientist, that won't cause any score change.

My total balances will only change around +$50 from the highest I've ever let report - $1688 - which didn't affect my score a few months ago.

[Anonymous]

---

@Anonymous wrote:

---

@mowglidude wrote:

---

---

I'm going to go below 8% aggregate UT next billing cycle for my card. I was thinking maybe try first under 8.499% and see if it rounds down to 8%, maybe like 8.35%. What do you think? I have one more card that posts the next day, if all goes according to schedule, that I can bring that I can pay down, to arrive below 8% aggregate UT.

---

Based on what the Principal Scientist at FICO told us in the AMA, anything less than 9.50% should keep util from the 10% threshold.

He didn't directly say that the algorithm uses the typical midpoint rounding, but he did say that my example 9.02% would round to the nearest integer, or 9% in that case.

(Midpoint rounding is also called round half away from zero or commercial rounding. Anything (x).5 to less than (x+1).5 will be (x+1).0 , where x is an integer or whole number.

I'm setting up my own test for this on 2 cards reporting on the 2nd and 5th. First one will be at 9.450%, and the second at 9.492%.

Unless I got trolled by FICO's principal scientist, that won't cause any score change.

My total balances will only change around +$50 from the highest I've ever let report - $1688 - which didn't affect my score a few months ago.

---

@Anonymous OP did not have a score change when he crossed 9%. He is testing 8, you are testing 9. Unless there's some other reason he did not receive an award. 

[Anonymous]

Either the thresholds are at 10 or 9 but you’ve got a hit 8 to be under 9 and when he hit 9 he didn’t get anything so is the threshold really at 9 and you have to hit 8 to remain under it?

You would think this is one of the most tested thresholds and that it would have to be 10 because everybody’s been doing it at 8.99% which is 9-according to that information.

The only other thing I can imagine is that he is on a dirty card and maybe maxed out and therefore not receiving a point gain for crossing the threshold.

And anywhere btw 7.5-<8.5 should =8.

[Anonymous]

---

@Anonymous wrote:
Either the thresholds are at 10 or 9 but you’ve got a hit 8 to be under 9 and when he hit 9 he didn’t get anything so is the threshold really at 9 and you have to hit 8 to remain under it?

The thresholds in the util chart are the ones FICO uses - internal to the algorithm, and the only ones that matter. Those are post-rounding thresholds.

For that 10%-19% threshold they list, we would have to have utilization less than 9.5% to avoid that interval, to stay in the [0,9%] rounded interval. 9.50% would be 10% and most likely cause a relatively small loss of points.

I'm not seeing much score change in the entire 0-9% interval (post-rounding) on a highly sensitive new credit file, so I don't think anyone else will either. Unless it's due to balances or number of cards reporting. 100% of cards reporting is a significant penalty for my scorecard on EQ/TU 8 (-19/-14) - but nothing on EX 8.

[Anonymous]

That’s exactly my thoughts on the matter but he crossed the threshold and did not receive any award.

[Anonymous]

So my only thought was the threshold might possibly be 9 instead of 10 because if not, he must be maxed out on the dirty card.

True enough utilization is weighted less on a dirty card.

[mowglidude]

---

@Anonymous wrote:
So my only thought was the threshold might possibly be 9 instead of 10 because if not, he must be maxed out on the dirty card.

True enough utilization is weighted less on a dirty card.

---

From what I understand, I have a clean card. I posted my stats earlier and I have no negs, etc...