Forum
|
|
Forum
|
     | Sunday, January 16, 2000 - 03:24 pm  Sean posted in the FICO Fraud topic: "The reason there was a change in the scores has to do with the changeover from the one month to the other. It should come as no surprise to anyone that a person's score that is taken on 6/29 will be different from a person's score that is taken on 7/2. Past delinquencies are a month further in the past on 7/2 than they were on 6/29. On the other hand the scoring model is also looking at the revolving account information and saying, "Hmmm...he was current on 6/1 but now it's 7/2 and there's no update as to how his account was doing on 7/1. He has a lack of revolving account information and a lack of recent installment loan information. It's been 30+ days since this account was updated -- that's bad." I think if you pull ANYONE's information on the 30th of the month and pull their info on the very next day you'll find different scores. Maybe those scores will go up and maybe they'll go down. It's a crap shoot." Thank you again for agreeing that Scores are pure gambling, as I've said for a long time. Is it true that your Scores would be significantly lower because creditors haven't reported by the 1st of the month? Would a Midol help?
|
| | Sunday, January 16, 2000 - 04:03 pm I don't think Midol is needed at all. But a brain would help. As I carefully reviewed the reports posted at http://www.bayhouse.com/FICOisFRAUD.html I concluded that the evidence is right there. On 8/1/97 three accounts still show the 6/97 reporting date. Yet, none of the Scores are down from 7/31/97. Sean: I'm very disappointed. Just recently we went through this discussion over FNMA due-on-sale clauses pertaining to leases. After endless irrelevant link posts and backtracking you finally took the "I don't care anyway" position, after your own link proved you wrong. And now you really stepped in it. At least you could have looked at the actual reports before going off on your bla bla bla meaningless made up as you go rambling. Too bad, really.
|
| | Monday, January 17, 2000 - 05:56 am Note how carefully I worded my response and how carelessly you overlooked it. Read my quote again: "It's been 30+ days since this account was updated -- that's bad." is what I said. Are you denying that on 7/31 that it may very well have been 30 or more days since the account was updated, assuming that the information went in on 7/1 we would find that 7/31 is exactly 30 days. And I did look at the reports, very carefully in fact. I noticed, among other things that your report is incomplete. We know that inquiries affect a person's score if they are within the past 12 months despite this we find that the credit profile you have given us claims "Credit Inquiries in the past 90 days: *** NO INQUIRIES ***" Does that mean there was, or very well could have been inquiries greater than 90 days that did not show on your report? I dare say there might have been. Could the primary borrower have had an inquiry made on, let's say, 7/17/96 which was considered to be 12+ months old on 8/1/97 and, therefore, no longer affecting the score? I dare say that's a possibility. I can make a number of guesses as to why the score might be different and can prove or disprove none of them without seeing a coded report instead of the EZ-read format you have displayed. If you don't want to hear my guesses, don't ask for them.
|
| | Monday, January 17, 2000 - 02:04 pm Sean, Obviously, it could NOT have been that an inquiry on 7/16/96 would affect the Score on 7/30/97 and drop off on 8/1/97. I assume that the scoring software at least knows how to calculate a year. What EVIDENCE do you have to make you state that an account that is not updated in the previous 30 days will negatively affect the Credit Score? I have actually NEVER seen the full date of the update. However, I'm very used to seeing reports that obviously have not been updated in the last 30 days. 2 months lags are quite common. I don't even see how the Scoring Software would count 30 days, since the date of the update isn't there. You seem to say that it looks at the MONTH reported and that Scores will then vary dramatically from the last day of the month to the first. Is that it? Specifically I refer to the 8/1 report. How much higher do you think the Scores would be if all accounts showed a 7/97 update? If you are right, people should be able to sue creditors for not updating in a timely manner.
|
| | Tuesday, January 18, 2000 - 09:19 am Well, I understand what you're saying when you figure that it can't be that hard to have a computer figure out when something is one year old. But let's take a look around and see what we can see by looking at my credit profile. As you can see near the top of the credit profile on the left hand side is the day the credit profile was run: 1-14-2000 at 8:24:56. Now go all the way to the end of the profile and see the last inquiry they show, which is FIRST CONSUMERS NATL B on 12-11-97. Now you and I both know that this inquiry is more than 2 years (24 months) old. In fact it's more than 25 months old, yet it still is on my credit profile. How hard can it be for Experian to write a program that removes these inquiries at the appropriate time? It doesn't seem that hard -- yet it's not happening. Next let's go to the Southland Civic FCU entry, which is directly above the inquiries. As you can see it shows as opened in 9-99 with an original amount of $500.00 and the last time it was updated was 10-31-99. A little math will tell you that it's been 75 days since this account was updated. Now check out the Experian Adverse Reason Codes and you'll see that item (32) is lack of recent installment loan information. It seems pretty likely to me that if my credit had been pulled that day I'd have lost points for the information on that account being 75 days old. How many? Who knows? But you have to figure people think "one point won't hurt me" until they get a 659 score and they needed a 660.
|
| | Tuesday, January 18, 2000 - 05:04 pm Sean: 1) Are ALL inquiries on the report included in any Scores? I still think that the Scoring Software is looking at today's date and then goes back one year. And so far I haven't had any evidence that this is not so. I didn't think that the inquiries *listed* could lower my credit worthiness. True or false? 2) Regarding the reporting dates, I would be concerned about that Southland Civic FCU account. Apparently their NOT updating is lowering your Scores. And that doesn't surprise me. However, I didn't think that anything under 2 months would be a considered negative. Most reports contain 30 to 60 day reporting lags. If that impacts on Scores NEGATIVELY, I think we should know. How do we find out?
|
| | Wednesday, January 19, 2000 - 02:32 pm As far as we know inquiries are counted for 12 months. The only point I'm making is that let's say you're the top brass at Fair, Isaac and you need to make a scoring model and you've decided that it's going to have 10 scorecards and you also know that the model being constructed is going to have a limited life, so you need to get it to market as quickly as possible. If it were me I'd assign 11 statisticians to the program and 11 computer programmers. One statistician to design each scorecard and a computer programmer to create a computer algorithm for that scorecard (that's 10 teams) and a final team to create the "front end" which will determine which scorecard the person's credit profile goes on to. To make things more complicated your programmers are going to be borrowing code from your previous generation of modeling that was written by entirely different statisticians and programmers. Under these circumstances, isn't it possible that different scorecards may calculate time differently because they were written by different teams? Perhaps one scorecard lops off all inquiries that were made more than 365 days ago (that means years with leapyears may be a little off, but who cares, right?) Perhaps another scorecard just ignores the day of the inquiry and just goes by the month and year, so that an inquiry made on 12/1/98 might still be counted on an inquiry made on 12/31/99. Perhaps another algorithm might even take the time of the inquiry into account, so that an inquiry made at 15:30 in a day might still count on the score until 15:31 a year later. From there each scorecard generates codes that describe why the person didn't score higher and then it's handed off to the lawyers. The lawyers sit down with the regulators and the regulator says, "I think saying 'Too many inquiries in the last 12 months' is least confusing to a potential consumer." So that's what goes in and it's certified ECOA-compliant even if that isn't the 100 percent truth it's close enough 'for government work' right? You want a theory as to why the scoring algorithm sometimes seems to be geeked, well that's my theory. It could be totally wrong and I'm willing to listen to anyone else's theory and give theirs the same consideration that someone might give mine.
|
Credit Forum CreditCourt Forum 2003 Credit Suit CreditFactors Order Credit Reports |