All zero except one with a store card

[folks19]

BTW, does this also apply to a store card which is a Visa or Mastercard (and can be used anywhere)?

[Anonymous]

---

@folks19 wrote:

---

@Anonymous wrote:

---

@folks19 wrote:

I now have 3 credit cards, but 1 of them is a store card (Amazon). Would the AZEO technique work to the same extent if one card is a store card? Does it matter which card has the small balance?

---

@folks19 excellent question. At Equifax there is a score loss for not having a balance on a national bankcard. There is also apparently a penalty for having a balance on a retail card. 

I think Experian also tracks retail Balances. 

So the effect at Equifax is you would lose points for not having a balance on a national bankcard and you would lose points for having a balance on a retail card, so it wouldn't be a good idea there. 

May I suggest you peruse the Scoring Primer linked at the top of my thread?

---

Wait... but I thought the score is generated by FICO not by Experian and Equifax?

---

@folks19 CRAs license a "black box" that is customized for each credit bureau. When the credit bureau needs a score, it puts its data in the box and the Score and Negative Reason Codes come out the other side, to make it simple.  

so in other words, fico creates the algorithm and then they license it in a proprietary format to the bureaus so that the bureaus can use that algorithm to create scores. 

[Anonymous]

---

@folks19 wrote:

---

@Anonymous wrote:

---

@folks19 wrote:

---

---

Wait... but I thought the score is generated by FICO not by Experian and Equifax?

---

[Fico] scores are generated using the Fico algorithm and provided by whatever source you get them from.  Those scores are drawn from whichever bureau data (EX, EQ, TU) is used. 

---

But I thought @Anonymous was saying that Equifax and Experian themselves penalties the use of a store card. I thought that the credit bureaus just provide the data, and FICO (or Vantage or any other scoring model) then creates a score based on its algorithm. I didn't think that EX and EQ (and TU or any other credit bureau) are actually involved in the algorithm which is used to generate the score.

---

@folks19 you really need to read the Scoring Primer. here's what happens: fico creates the scoring algorithm for each CRA. They do that based on that CRA's datasets.

Therefore the resulting algorithms are somewhat different at each CRA due to the differences in the datasets among the bureaus. That's why yes Equifax appears to penalize for a retail balance and there is a loss for the lack of a balance on a national bankcard and for the lack of a balance on a revolver, where other CRAs may not.

So at EQ8, by just having a balance on a revolver, you would get the national bankcard loss, but a balance on a retail card would net a retail penalty and a national bankcard loss. 

that's why identical data across bureaus can result in different scores, because the Version of the algorithm is customized for each bureau.

we haven't confirmed Experian, but it seems that it penalizes for retail balances as well, otherwise why is it separately tracking retail balances? There's not a loss for a lack of a retail balance.

We actually have a recent thread discussing all this I think it was entitled inactivity and we got into that as well.

[Anonymous]

---

@folks19 wrote:

---

I didn't think that EX and EQ (and TU or any other credit bureau) are actually involved in the algorithm which is used to generate the score.

---

Basically correct, although I believe EQ does have their own scoring model/score you can get directly/only from them.  It's quite irrelevant though, as no real lenders are ever going to look at it.

[Anonymous]

---

@folks19 wrote:

BTW, does this also apply to a store card which is a Visa or Mastercard (and can be used anywhere)?

---

@folks19 if it's got a MasterCard Visa American Express or Discover label on it, it's a bankcard. 

[folks19]

@Anonymous But I thought you were saying that Equifax has a way of treating store cards. That would imply that Equifax also has a "say" in the algorithm used for generate the score. I didn't think that's the case.

[folks19]

---

@Anonymous wrote:

---

@folks19 wrote:

---

@Anonymous wrote:

---

@folks19 wrote:

---

---

Wait... but I thought the score is generated by FICO not by Experian and Equifax?

---

[Fico] scores are generated using the Fico algorithm and provided by whatever source you get them from.  Those scores are drawn from whichever bureau data (EX, EQ, TU) is used. 

---

But I thought @Anonymous was saying that Equifax and Experian themselves penalties the use of a store card. I thought that the credit bureaus just provide the data, and FICO (or Vantage or any other scoring model) then creates a score based on its algorithm. I didn't think that EX and EQ (and TU or any other credit bureau) are actually involved in the algorithm which is used to generate the score.

---

@folks19 you really need to read the Scoring Primer. here's what happens: fico creates the scoring algorithm for each CRA. They do that based on that CRA's datasets.

Therefore the resulting algorithms are somewhat different at each CRA due to the differences in the datasets among the bureaus. That's why yes Equifax appears to penalize for a retail balance and there is a loss for the lack of a balance on a national bankcard and for the lack of a balance on a revolver, where other CRAs may not.

So at EQ8, by just having a balance on a revolver, you would get the national bankcard loss, but a balance on a retail card would net a retail penalty and a national bankcard loss. 

that's why identical data across bureaus can result in different scores, because the Version of the algorithm is customized for each bureau.

we haven't confirmed Experian, but it seems that it penalizes for retail balances as well, otherwise why is it separately tracking retail balances? There's not a loss for a lack of a retail balance.

We actually have a recent thread discussing all this I think it was entitled inactivity and we got into that as well.

---

@Anonymous Yes, I understand that each CRA has different data about me. For example, not every bank uses the the same CRA for inquiries, and some banks don't report to all 3 major credit bureaus. But I don't think the credit bureaus "treat" data differently. For example, the affect of a hard inquiry on EX is the same affect on TU, assuming that both reports are completely identical (though that's usually not the case). Furthermore, the credit bureaus don't even have a "say" in the algorithm used to generate a FICO score, and I would assume that FICO uses the same algorithm for any profile that's inputted,  regardless of the credit bureau providing that profile. So how would one credit bureau "treat" retail credit differently in the FICO algorithm?

[Thomas_Thumb]

---

@folks19 wrote:

---

@Anonymous wrote:

---

@folks19 wrote:

I now have 3 credit cards, but 1 of them is a store card (Amazon). Would the AZEO technique work to the same extent if one card is a store card? Does it matter which card has the small balance?

---

@folks19 excellent question. At Equifax there is a score loss for not having a balance on a national bankcard. There is also apparently a penalty for having a balance on a retail card. 

I think Experian also tracks retail Balances. 

So the effect at Equifax is you would lose points for not having a balance on a national bankcard and you would lose points for having a balance on a retail card, so it wouldn't be a good idea there. 

May I suggest you peruse the Scoring Primer linked at the top of my thread?

---

Wait... but I thought the score is generated by FICO not by Experian and Equifax?

---

FICO sells a suite of products. The base algorithms are customized to meet primary customer "needs" (customers being the CRAs). It's like buying a certain model car where options are selectable. The allowed level of customization has been reduced from Fico98/Fico04 to Fico 8 and even less CRA customization is allowed with Fico 9.

Perhaps no customization in Fico 10T. (like a Ford model T - you can have any color as long as it's black)

Side Note: The CRAs do not know the design details of the algorithms but, they know some basics ... so not truely a black box. Regardless, as mentioned the CRAs plug in the data and the algorithm (licensed for use from Fico) outputs a score with reason codes/statements.

Regarding store cards, merchants often offer pure play cards and co-branded Visa/Mastercard or AMEX cards. My understanding has been that co-branded "store" cards are viewed like bank cards by the algorithms. However, not 100% sure.

[Anonymous]

@folks19 yes the algorithm at each CRA is different. And everything @Thomas_Thumb  said is correct, including the fact that co-branded cards are treated as bankcards.

If you take score 8 and you have identical data across all three bureaus you can have three different scores because for instance, Equifax's algorithm can have a penalty that transunion's does not.

This is not necessarily by picking and choosing at the request of the bureau, but based on an analysis of their datasets. They have about 500 characteristics and they run high-level math on the dataset in question and they find the ~40 most predictive characteristics for that dataset and that becomes the basis of the scorecard metrics for the algorithm's version at that bureau. 

it's called logistic regression. 

[Anonymous]

---

@folks19 wrote:

---

@Anonymous wrote:

---

@folks19 wrote:

---

@Anonymous wrote:

---

@folks19 wrote:

---

---

Wait... but I thought the score is generated by FICO not by Experian and Equifax?

---

[Fico] scores are generated using the Fico algorithm and provided by whatever source you get them from.  Those scores are drawn from whichever bureau data (EX, EQ, TU) is used. 

---

But I thought @Anonymous was saying that Equifax and Experian themselves penalties the use of a store card. I thought that the credit bureaus just provide the data, and FICO (or Vantage or any other scoring model) then creates a score based on its algorithm. I didn't think that EX and EQ (and TU or any other credit bureau) are actually involved in the algorithm which is used to generate the score.

---

@folks19 you really need to read the Scoring Primer. here's what happens: fico creates the scoring algorithm for each CRA. They do that based on that CRA's datasets.

Therefore the resulting algorithms are somewhat different at each CRA due to the differences in the datasets among the bureaus. That's why yes Equifax appears to penalize for a retail balance and there is a loss for the lack of a balance on a national bankcard and for the lack of a balance on a revolver, where other CRAs may not.

So at EQ8, by just having a balance on a revolver, you would get the national bankcard loss, but a balance on a retail card would net a retail penalty and a national bankcard loss. 

that's why identical data across bureaus can result in different scores, because the Version of the algorithm is customized for each bureau.

we haven't confirmed Experian, but it seems that it penalizes for retail balances as well, otherwise why is it separately tracking retail balances? There's not a loss for a lack of a retail balance.

We actually have a recent thread discussing all this I think it was entitled inactivity and we got into that as well.

---

@Anonymous Yes, I understand that each CRA has different data about me. For example, not every bank uses the the same CRA for inquiries, and some banks don't report to all 3 major credit bureaus. But I don't think the credit bureaus "treat" data differently. For example, the affect of a hard inquiry on EX is the same affect on TU, assuming that both reports are completely identical (though that's usually not the case). Furthermore, the credit bureaus don't even have a "say" in the algorithm used to generate a FICO score, and I would assume that FICO uses the same algorithm for any profile that's inputted,  regardless of the credit bureau providing that profile. So how would one credit bureau "treat" retail credit differently in the FICO algorithm?

---

@folks19 no, the score change for an inquiry is not necessarily going to be the same across CRAs with the same data.

I don't know how much say the CRA has in customization, but I would imagine they probably have some say, considering they're the one paying for it.

no, actually there are 12 different algorithms for score 8 at TransUnion; 12 different algorithms for score 8 at Equifax; and 12 different algorithms for score 8 at Experian.

Actually there's more than that if you wanna include the industry options. Each scorecard is its own algorithm. The preliminary algorithm is the same across bureaus, that determines which scorecard you go to, and then that scorecard is a unique algorithm with characteristics and weightings that vary based on analyses of the respective bureau's datasets with that versions framework. 

So a clean/thick/mature/new accounts Scorecard at Experian may have different scoring metrics and weightings than the corresponding Scorecard for that version at Equifax or TransUnion.

respectfully that's why we do all the studying and testing. if they were all the same, this would be pretty easy.

Just like the mortgage algorithms, each one of those are different at each bureau. Each version has its own set of scorecards (unique algorithms) at each bureau, based on that bureau's unique dataset that it was created from.