@Revelate wrote:
@SouthJamaica wrote:
@Revelate wrote:I think we have to agree to throw out the CMS alerts: those aren't reference; the lone exception to that being the reason codes TU provides in their FICO 8 monitoring here on some of those alerts: those are gold.
The base report data is, I'd assume that national bankcard absolutely counted and the CU may or may not have and now you're looking at possible 2 inquiry shift not one.
I need to go document my last two inquiry score shifts and then see what happens when they start coming off as it should match 1:1 assuming I'm on the same scorecard which I should be at that point.
Is there any chance that the inquiries that were close in time were grouped together? I.e., is it possible that the 11/27/18 and 11/29/18 inquiries, the 10/7/18 and 10/10/18 inquiries, and the 5/30/18 and 5/31/18 inquiries were each counted as one rather than two, even though they were neither auto loan nor mortgage inquiries? If so that would explain why EX is treating 8 inquiries as 6, and 7 inquiries as 5.
I've never seen "installment" loans being deduped in any literature I've seen, only Mortgage, Auto, Student.
Or credit union ones either but I'd think it's possible; I just don't think we quite know how these work which unfortunately throws off the data analysis. My hope is we'll get enough datapoints from enough people on inquiries that we can tie something together but it's a messy analysis when some of the assumptions that we've had in the past are being debunked.
Yeah, this thread from 2016 shows how controversial it is. The OP started out thinking, and ended up thinking, that credit card inquiries are de-duped in some way.
Perhaps, now that we have access to daily updates with EX, we will be able to make some progress on pinning this down.
In my case, 8 inquiries were for credit card applications, so there is no question of rate shopping for auto, mortgage, or student loans.
The EX site, although identifying each of the 8, counted them as 6, and is now counting the 7 remaining as 5. And what stands out is that there were 3 pairs of very close-in-time inquiries, from 1 to 3 days apart, and one of those pairs just dropped out.
So maybe, just maybe, multiple credit card inquiries which are from 1 to 3 days apart are de-duped on EX.
It certainly bears watching.
I just revisited my last 3-bureau report. First I looked at the inquiries listed in the data section. Then I looked at the number of inquiries counted in the scoring section. There was a huge difference, suggesting that there is definitely de-duping of inquiries other than auto, mortgage, and student loan inquiries, because none of these were for such loan applications. And most of these were not extremely close in time to each other.
Actual in 12 mos Counted
EQ 12 6
TU 2 1
EX 9 7
SJ: the CMS stuff isn't reference. Just because they say you have six it could well be wrong as they absolutely don't know the algorithm implementation.
Score changes and reason codes really the only way to track this.
