Forum
|
|
Forum
|
     | Wednesday, March 08, 2000 - 10:14 pm  It is my understanding that the credit bureaus will not give out the methods for calculating credit scores in part because they are too complicated. Do you think if you had the actual formulas in front of you it would be too difficult to actually arrive at your score. Or are they just being cowardly?
|
| | Thursday, March 09, 2000 - 03:58 am Fair, Isaac answers the big question from creditscoring.com: "What horrible thing will happen if you release the consumer's score to him?" MS. TWOHIG: Kirk, I just want to make sure I understood one thing. Is it the case, though, that you don't give consumers their actual score when you explain to them the reasons why they might be denied?http://www.creditscoring.com/pages/forumtranscript.htm#page297
|
| | Thursday, March 09, 2000 - 06:11 am I think Greg has missed the mark in his quote, although it's a nice quote and I've personally read that entire discussion and was glad to have read it. The question is, will Fair Isaac give you enough information about their scoring system to where you could sit down with a calculator and figure out your own score from the raw data. The answer to that is absolutely not -- and not because it is too complicated for someone to understand (although it is complicated). Fair Isaac has devoted a lot of labor hours by people who are paid a lot of money to try to figure out what in a person's credit profile is an accurate predictor of future activity (statistically). In exchange for all this work they recoup their investment by either selling or licensing the computer software that will generate these scores for a lender. Fair Isaac doesn't want to publish their algorithm for fear that people will begin to use it without paying them and for fear that their competitors will be able to take their formulas, improve on them a little and market a new product in competition with theirs cheaper.
|
| | Thursday, March 09, 2000 - 06:58 am So the consumer should suffer because Fair Issac is afraid of losing money? That is not a winning argument, in my book.
|
| | Thursday, March 09, 2000 - 07:01 am Sean, if you read the discussion, why, when someone asked "What is the best? I thought it was 850 but don't recall the source of that memory," did you reply, "I know of only one person that has scored better with an 805." Is 805 "something close to 900," (as Fair, Isaac said) in your view? Sean said, "The question is, will Fair Isaac give you enough information about their scoring system to where you could sit down with a calculator and figure out your own score from the raw data.[?]" No, the questions were: "Do you think if you had the actual formulas in front of you it would be too difficult to actually arrive at your score.[?] Or are they just being cowardly?" Lest I be labeled presumptuous, I found the closest thing to the first real question, and Fair, Isaac's answer to it, as possible. As for the second question, the evidence will have to suffice for a basis on which and individual makes his evaluation of cowardice. Fair, Isaac said, "We don't want you altering your behavior to change the score... We don't want to have consumers trying to alter their behavior in short term ways that will-- regularly has nothing to do with their long term credit risk." Sean said, "Fair Isaac doesn't want to publish their algorithm for fear that people will begin to use it without paying them and for fear that their competitors will be able to take their formulas, improve on them a little and market a new product in competition with theirs cheaper." Sean, I got my information directly from the source. From where did you get your information? Are you just guessing?
|
| | Thursday, March 09, 2000 - 07:08 am Actually, Greg, I heard what Sean said at the FTC meeting. But I agree with techartist that that is not a good reason to keep the scores private. It is an extremely lame reason. I think they could publish the scoring methods and still keep collecting licensing fees, much like they do at the mortgage companies who take the Fair Issac model and customize it.
|
| | Thursday, March 09, 2000 - 07:42 am I was responding to the statement by the original poster: "It is my understanding that the credit bureaus will not give out the methods for calculating credit scores in part because they are too complicated." I do not believe that Fair Isaac refuses to give out the algorithms because they are too complicated. They perceive that giving out the scoring algorithm is not in their best interest. I would tend to agree. Since Fair Isaac created the scoring algorithm they also own it. Part of what makes us a wonderful and free country is that you can do whatever you like with your own property. In the event that the government figures out that they need your property they can take it, but must pay you the fair market value of the property. It says so right in the Bill of Rights, specifically the Fifth Amendment. The Bill of Rights is a big portion of what makes our government the best and most successful on the planet. Even if a majority of people believe the Bill of Rights can be dispensed with in this one case that doesn't matter. If the Bill of Rights can be violated anytime a majority of people say so then soon there will be no freedom of religion or speech and the police will torture confessions out of you because they "just know" you're guilty. If you feel strongly that the FICO algorithm should be in the public domain then I suggest you start a fund and solicit donations to get enough money to buy the algorithm and once you buy it you can do whatever you want with it.
|
| | Thursday, March 09, 2000 - 07:45 am My personal opinion is in the middle ground. Fair Isaac spent a lot of time and money on this and should not have provide its competitors with anything for free. It should be treated like a patent or a copyright. If decisions based on that score affect me: 1) I have the right to know what that score was and how it was reached. 2) I have the right to challenge the algorithm in court to contest its validity. Not just to say 'well if it wasn't accurate the credit grantors wouldn't be using it'. It just means that without access to it, no one can disprove it. And it is the closest model available with some semblance of validity.
|
| | Thursday, March 09, 2000 - 08:27 am Sean: You didn't answer my questions. You're making it up, right? What about that 805 vs. "something close to 900" business? While you're acting as spokesman for the industry, here's another question: If the credit bureaus gave Fair, Isaac access to the data, will they allow access by anybody to create a new formula? And, why didn't they answer that question when I asked it, in person, to their faces (again, since you are speaking for the industry)? Kristy: What did you mean when you said, "Actually, Greg, I heard what Sean said at the FTC meeting." Do you mean McCorkell?
|
| | Thursday, March 09, 2000 - 09:01 am Greg: I've reviewed this thread carefully for the text "805" or "close to 900" and I have not found it in this thread. Please paste a link to where I can find it. Greg & Kristy: I was not at the FTC meeting and did not speak there.
|
| | Thursday, March 09, 2000 - 09:23 am Sean: Yes, I know you weren't there. We were. At the top of this thread, you said: "Fair Isaac doesn't want to publish their algorithm for fear that people will begin to use it without paying them and for fear that their competitors will be able to take their formulas, improve on them a little and market a new product in competition with theirs cheaper." I was just telling you I heard the Fair Issac rep say something along those lines at the FTC meeting. (Greg asked where you heard this.)
|
| | Thursday, March 09, 2000 - 09:44 am http://www.bayhouse.com/discus/messages/4/371.html?952629693#POST2377 You still didn't answer my questions. You're making it up, right? I got my information directly from the source. From where did you get your information? Are you just guessing? What about that 805 vs. "something close to 900" business? While you're acting as spokesman for the industry, here's another question: If the credit bureaus gave Fair, Isaac access to the data, will they allow access by anybody to create a new formula? And, why didn't they answer that question when I asked it, in person, to their faces (again, since you are speaking for the industry)?
|
| | Thursday, March 09, 2000 - 09:53 am No one is quite sure what the upper limit is on FICO scores. A 799 is very, very good. I know of only one person that has scored better with an 805. No, I'm not making it up that I personally know someone who scored 805. Nor am I making it up that no one seems to know what the exact FICO range is. If you're interested in hearing various suggested ranges, why not try here? I'm surprised that you're not aware of this information since it's on your own website. Peter McCorkle said the number range is "...the range is from somewhere in the low 300's to something close to 900." Is 850 "something close" to 900? What about 860? 870? 880? 891? Is 349 in the "low 300's"?
|
| | Thursday, March 09, 2000 - 10:11 am This is a hoot. As any idiot can see, in the thread I quoted (http://www.bayhouse.com/discus/messages/4/371.html?952629693#POST2377), you were responding to the question, "What is the best?" If you knew that Fair, Isaac said "something close to 900," you should have said it. You could be playing games, or could really not know what you're talking about. 805 is not "close to 900." It is just over 800. We can let the audience decide. Given a range of about 550, 95 points is significant (17%). It is amusing to me that given the bean counters, their computers, and their statistics, that even they can't say what the highest (or lowest) score, ever, is.
|
| | Thursday, March 09, 2000 - 10:31 am As anyone can see from reading Voigtkamff's Post he said, "...I hate to broadcast my ignorance, but how good is [a] 799?" My answer "...A 799 is very, very good. I know of only one person that has scored better with an 805. That person had 8 credit cards, 4 apartment building mortgages, one personal residence mortgage and two car loans, no inquiries and had never been late on anything. She works for Washington Mutual." I believe this makes it clear thatthat I am speaking of someone I know personally.
|
| | Thursday, March 09, 2000 - 11:32 am The only thing it makes clear is that you answered the question with a half-truth. I have been to Detroit, Phoenix, Boston, and St. Petersburg. But that doesn't mean I haven't been to Omaha and Seattle. You still didn't answer the other questions, above. Stop trying to change the subject. You're losing.
|
| | Thursday, March 09, 2000 - 12:52 pm To label it clearly, the following is my theory. If anyone here was in the business of extending credit, we would reasonably want some way to predict which individuals would pay us back our money. There was a demand for a means of predicting. The supply was met by FICO. FICO built a mousetrap. They advanced money out of pocket in R&D to develop that mousetrap in hopes that there would be a return on their investment. But instead of a mousetrap, they developed a series of observations. I know little about intellectual property law, but it seems difficult to claim exclusivity to the observation that a person with too many inquiries may be a bad risk. Even when you combine the statistical observations and develop a mathematical formula (for example E = MC2) how do you stop people from using it without paying? Let's pick an easier formula. Suppose we needed to calculate the area of triangles on a regular basis. FICO comes up with computer software to do so. But if they let the formula slip, we could just calculate it ourselves. Or some competitor could develop a tweaked software to improve on the earlier mousetrap. And we would buy their software instead. The software would be cheaper because the copycat company spent no money on R&D.
|
| | Thursday, March 09, 2000 - 02:22 pm Greg: If you'll carefully read the opening post that started this whole thread and break it down you'll see that the person had three concerns: 1) He wanted to understand if the reason the algorithm wasn't given out was if it was too complicated; 2) He wanted to know if a simple person with long and detailed algorithm instructions in front of him and a calculator could do it; and 3) He wanted to know if Fair Isaac was just being cowardly. You abruptly changed the conversation to a discussion "805 vs. close to 900" and you finally had to paste the topic in from another thread in order to get your completely irrelevant point across. Then you say: "You still didn't answer the other questions, above. Stop trying to change the subject. You're losing." Here's my recommendations: 1) If you have specific questions for me, lay them out in a numbered format. 2) Don't accuse others of doing things (such as changing the subject) when it's quite clear that you are doing the exact same thing. 3) If you wish to turn this into a win-lose competition then don't you think a referee should be appointed? And don't you think it would be a bad idea to appoint yourself as the referee? It just doesn't look right. Please refrain from pointless and irrelevant comments like "You're losing" as that would be a matter to be decided by a neutral third party. 4) Unlike you, I'm not here to "win" some pointless argument competition. I'm here to help and to learn.
|
| | Thursday, March 09, 2000 - 02:41 pm Sean: You're losing. And you're here to push your agenda: that the consumer can't get their score (in writing-- ha, ha, ha, ha, ha, ha, ho, ho, he, he). I won't be numbering my questions-- you'll just term them, in your infinite wisdon, irrelevant. If you really want to make an effort to answer them (which I'm sure you don't-- causing you to go long winded in the last post about things other than the issues), you can identify the quetions as sentences with "?" at their ends. The top two questions on creditscoring.com (asked to their faces-- no answers given): "What percentage of loans evaluated by Fannie Mae and Freddie Mac automated mortgage underwriting use scores created by Fair, Isaac?" "Will the credit reporting agencies release data to anyone who wants to create a new scoring system?" They thought they got away with something, but not answering only bought them proposed legislation less than two months later to release the scores-- and a LOT of press. They can't develop a national system, used by two companies who have a unique, favorable status with the federal government (and use it to control the mortgage banking industry) and not expect people to ask questions. The two companies themselves are under fire now, with congress breathing down their necks, threatening to take away subsidies, including the $2.25 billion line of credit from the Treasury Department for each enterprise. http://www.house.gov/banking/gsebsbs.htm. Questions like, "OK, now that you have dominance and control, and you are the arbiters of what is good and bad credit, what is good credit and how do I get it?" Duuuhhhhh. They say that bankers use the score as a tool. But they can't even say what the highest and lowest scores are. If you don't know the limits (like 0-100%), then how do you know how to grade somebody? A score with no scale. Isn't that a riot? Keystone Cops. On second thought, Sean, here's a spoon-fed question to you which you didn't answer above. Let's do this one... at... a... time. 1. While you're acting as spokesman for the industry, here's another question: If the credit bureaus gave Fair, Isaac access to the data, will they allow access by anybody to create a new formula?
|
| | Thursday, March 09, 2000 - 02:46 pm 1. I am not acting for spokesman for the industry. I am not affiliated with Fannie Mae, Freddie Mac, Fair Isaac, Equifax, Experian or Trans Union in anyway. I cannot answer your question. And I hope that both Fannie Mae and Freddie Mac lose their federal "preferred" status. They are both privately held corporations.
|
| | Thursday, March 09, 2000 - 05:44 pm 2. Fair, Isaac said, "We don't want you altering your behavior to change the score... We don't want to have consumers trying to alter their behavior in short term ways that will-- regularly has nothing to do with their long term credit risk." http://www.creditscoring.com/pages/forumtranscript.htm#page297 Sean said, "Fair Isaac doesn't want to publish their algorithm for fear that people will begin to use it without paying them and for fear that their competitors will be able to take their formulas, improve on them a little and market a new product in competition with theirs cheaper." Sean, I got my information directly from the source. From where did you get your information? Are you just guessing? Or did you read their minds? Or did they say that to you, but not to the public?
|
| | Friday, March 10, 2000 - 04:14 am First of all we're talking about two separate topics. 1. Should consumers be allowed to see their score? 2. Should everyone be allowed to see the algorithms? In my opinion: 1. Yes, they should be allowed to see their score; provided, however, that they don't get anything in writing; and 2. No way should the algorithm be public because Fair Isaac owns it and everyone would just begin copying it. Also I got my opinion direct from the source -- ME!
|
| | Friday, March 10, 2000 - 08:02 am Whenever somebody says "First of all...," I get the feeling that a weak argument is coming. Then, when they don't follow it with "Second...," I am sure of it. I was right again. You were making it all up. I love that "not in writing" business. 3. Do you think if you had the actual formulas in front of you it would be too difficult to actually arrive at your score[?]
|
| | Friday, March 10, 2000 - 08:50 am 3. Since each item is simply given a point value I would consider that scoring a credit profile would be deceptively simple once the proper scorecard was selected. The greatest danger would be losing count or overconfidence leading to a careless mistake. As far as I understand scoring a credit profile is a matter of simple addition and subtraction. On a personal note, Greg, I find you to be verbally abusive. I recall reading a story of a gentleman whose wife filed for divorce and the judge ordered him to sell all property and give the wife half. Accordingly he sold his vehicle, house and many other possessions all for $2 and gave the wife $1. This is a somewhat extreme example of someone living an enemy-centered life. His primary thoughts, goals, movations and actions are to hurt others and he is not shy about extreme actions that also harm himself and act other than in his own best interest. For another example of enemy-centered behavior I suggest you hit up this website. The author obviously has some serious issues with the credit scoring and reporting industry and has spent a good deal of time building this website as a vehicle from which he can hurl his anger and hurt against those whom he perceives has harmed him. I think it would be interesting to know what happened that caused this person to expend so many hours and days on a project of this magnitude. I also think it would help us to assess the objectivity of the information and presentation on that website. One example can be found where he taunts his perceived enemy with the comment, 'You misspelled the words "whether" and "Cadillac." If you are the data entry operator working my file, that error could affect my credit score.' Yet the author frequently misspells items himself, such as the name of his own website, unless of course he has changed it to "creitscoring.com" Perhaps if the author spent as much time fixing his own faults and flaws as he does trying to belittle other people for their real or imagined flaws he'd pay his bills on time, have a decent credit score and not need to devote an entire website to how "unfair" the system is for accurately reporting on his failings.
|
| | Friday, March 10, 2000 - 05:52 pm You don't know my credit score or payment habits any more than you know what I had for breakfast. Stop trying to evade the issue with corny tactics. That answer in the first paragraph: it took three sentences to answer a yes-or-no question. But was that a "Yes" or a "No"? 4. Is 805 "something close to 900," (as Fair, Isaac said) in your view? I know it will hurt, but bring yourself to answer the question.
|
| | Saturday, March 11, 2000 - 03:14 am Are you trying to tell me, Greg, that you constructed that whole website and you have no personal issue with the credit reporting agencies? Why are you so reluctant to say what your score is? 3. I feel confident that I would have no problem calculating a score accurately with an Excel spreadsheet and the scorecard in front of me. (a) Fair Isaac uses 10 different scorecards and I don't know what is used to determine which scorecard and I don't know how complicated it is. (b) Considering the general lack of quality education provided by America's socialized education system I am not going to make any declaration about whether the average American can or cannot add and subtract. 4. I won't answer this question until you answer my 6 questions I asked you about the range Peter McCorkle outlined.
|
| | Saturday, March 11, 2000 - 08:08 am Oh, please. You're the only one asking what my score is because you're on the ropes and going for broke. My motivation doesn't change the facts: they give my score to potential creditors, but not to me-- which makes no sense, given the widespread use and standardization of the credit bureau score. I could say that I'm reluctant to say what my score is for the same reason they don't to give it to me: Idonwanna, and Idon'thafta-- the law says so, and I'm going to hide behind it. So there. The score will do you no good (as they say). Too bad. Stop wasting my time. Television network correspondents don't have to reveal their political party memberships in order to have credibility, newspaper reporters don't have to reveal their religion, and I don't have to give my score to comment, make points about, and shape their goofy industry. Drop it; you are being ridiculous and redundant. I told you two days ago that I wouldn't disclose it. What's wrong with you? Didn't you understand that? The answers to your six questions: "Is 850 'something close' to 900? What about 860? 870? 880? 891? "Is 349 in the 'low 300's'?" No, no, no, no, no, and no.
|
| | Saturday, March 11, 2000 - 09:08 am I wouldn't consider 805 to be close to 900. I'm a +/- 5% kinda guy, which would put "close" range from 855 to 945. Why do you say 891 isn't close to 900? Why isn't 880? Where do you draw the line at what is and isn't close? Why didn't you ask Peter McCorkle what he meant by close? You were at that meeting ... I wasn't.
|
| | Saturday, March 11, 2000 - 10:11 am If it's 301 to 945, that's a range of 644. 45/644= 6.99%, not "5%." If anybody takes the time to read the posts above, they'll catch your baloney in a minute. Given those gaffes, you aren't in any position to discuss numbers with me and split hairs while doing it. My idea of "something close to 900" would be 899.9999, in this case, because of the dorky games they're playing. I didn't ask the question because I had the chance to ask five others, and did. See if you can find the answers to the five. It's funny. You're the one who responded to the question about what is best with "805." I'm questioning you, not them.
|
| | Saturday, March 11, 2000 - 11:02 am That's the stupidest thing I've ever heard. You really think Peter McCorkle meant 899.99999 when he said "something close to 900" as the upper end of the scoring algorithm?
|
| | Sunday, March 12, 2000 - 11:34 pm Joe +/- 5%: I have no idea what he meant. http://www.bayhouse.com/discus/messages/4/397.html?952713631#POST2486
|
| | Wednesday, March 22, 2000 - 02:59 pm Greg and Sean Don't know if I can help either of you, but in my own research I have found a few things: 1. In one of the links in the long list at creditscoring.com, I found a link to a mortgage company site where the author said he had seen a woman's report with an 824. If I find it again, I'll post it. 2. Having done a lot of work on scores in the last year, I can say that even if you knew the score cards, its not terribly simple. First, time is a weighting factor on most of the others. And it appears to go in discrete increments, not smooth transitions. For example, when a recent delinquency or inquiry hits the 6 month age mark, (given no other changes) its impact on the score drops, dramatically if it is a delinquency, less so, obviously, if it is an inquiry. 3. Having said that, there is a site I found that has a free little credit model that has, on the 6 reports of mine I tried it with, always come up within 10 points of what the score actually was on the report (in one case 3 points). Now, it uses income, which, if the bureau reports had that data, they didn't show it, and length of employment, which, two of my bureau reports don't show if they do have it. But it would give someone a good idea if all they had was a consumer report. Obviously, I have only my own experienc and can't promise anyone else it woud be so predictive. 4. In running some VERY limited statistical analysis on some of my own old reports and make my own scoring model (nothing as complex as the FICO ones) I found that since the score I was generating represented a probablility, it appeared was theoretically no limit to the score. It did, however, appear that it might display an asymptoptic behavior since I was using finite numbers. In other words, at some point, if someone has a 1 in 100,000 chance of default, that score will be pretty much the same as a 1 in a 100,000,000 chance if the data set was limited to say 10,000,000 (my made up numbers) and simply couldn't predict that big a number. I obviously have no idea if the FICO scores would exihibit the same kind of behavior, but it wouldn't surprise me. Hope some of this helps
|
| | Wednesday, March 22, 2000 - 04:50 pm Didn't we already have a discussion like this where I said credit scores are higher at the end of the month and they take a noticable, sudden drop at the beginning of the next month? I wanted to post the thread but I can't find it.
|
| | Thursday, March 23, 2000 - 01:46 am Richard: Thank you. Freddie Mac's site had a graphic of score distribution-- a bell curve-- that centered around 620 as I remember. That's as significant as the limits, to me, and that nature of the numbers would explain why they can't come right out and say what the highs and lows are: theoretically, there are none. Practically, there are. The model is only, perhaps, a million consumers. http://www.shodor.org/interactivate/lessons/sm1a.html The comedy comes in when they try to answer the question about what the limits are. If they can't give a direct answer, the next question would be: "You said somewhere from about ____ to ____ ; when was that recorded? And how often do you review the model for the mean, median, mode, standard deviation and limits?" Blah, blah, blah. The problem, of course, is that this sociological model-- used by academics for eons-- is now close to everyday life, and they're having a hard time explaining it, and a hard time explaining why they don't want to explain it. The notion that credit scoring "really does sound like the SATs" (FTC)(I wonder who planted that seed) could be closer to the mark than I thought, legally and ethically, because actual people have to write the SATs, and actual people are making decisions about what goes on within the models, too. If their methods were revealed, it may show how biased and off their assumptions were to begin with-- but, of course, it will never see the light of day. These magicians would rather drown than reveal their trick. They say, "... At Fair, Isaac, we have always believed that fairness is essential to the products we make and the way we make them. We acknowledge there is no universal standard... " So, we're stuck with theirs: some lawyerly, upper middle-class, west coast, Bay area, psuedo-intellectual, academic, la-de-da, created-to-sell-to-the bureaus, view of life. What kind of work have you done on scores?
|
| | Thursday, March 23, 2000 - 06:58 am Sean I think your comments about the end of the month thing is in the thread about Christines FICO is Fraud page. I agree with you there. Greg Yep, you said what I did in a lot fewer words. It is because the sample is a finite number that there are practical limits. I doubt anyone at Fair Isaac DOES know, and if they do, I actually agree that the upper limit probably doesn't matter. Given that most lenders use scores, all most consumers really care about is getting their scores high enough to get what they want. If a lender is using the score as it was intended, they are saying "Well, for those with a score above X, 1 in A will default, and if I charge M%, I'll make money. AND, I can even accept those with a score below X but above Y, because 1 in B will default, and at that rate, if I charge N%, I'll still make money". Given that the score is supposed to represent likelihood of default, what we as consumers who are stuck with this crap want to do is make sure we get a high enough score that the odds of us defaulting (according to the model) is low enough to get the best rate (or just accepted if you have had scores as low as I have). Of course, in my 6 years of experience dealing with scores, only 1 in about 12-15 of the loan officers I dealt with even knew what they were supposed to mean. I agree, they sound pretty stupid trying to explain it, and say "Why would a consumer want a score". If they can't figure that out, they are even dumber than they sounded at first. As far as the work I have done with scores, (being an engineer and physicist and having a sick interest in numbers), I started a few years ago and decided I would increase my score. Over the past twelve months, I managed to bring it up about 155 points and, as you can imagine, have a stack of credit reports about 2 feet thick. Even with that much data, I couldn't place a number on the different factors, so I started gathering other peoples reports (friends or relatives who, after listening to me gripe for a year about scores, would be about to buy a car or house and say, "Is there anything I should know", and I'd say I'd have to see their report, if they didn't mind). I should probably organize and get as many people as I can to send me reports and see if I can put together enough to fit some curves to it. I actually think I have a good idea what the behavior of the curves are, its the exact numbers that I still don't have enough data to figure out. For example, I don't know exactly what increments age of things goes by. Is it 1 month for the first 6, then 3, then 6, then 12, then 24. I have had things change over time, but didn't have score changes until something important hit a 6 month age increment. Stuff like that. Everytime I think I have nailed down a number, I find something different. For example, my bet (i could be way wrong) is that on Christines Fraud page, the reason for the drop from day 1 to day two (same month) is the previous days inquiry (I know it doesn't show, but it is a merged file, so you can't be sure). I have never seen an inquiry cause a 24 point drop, but then, I have never seen a report with just one tradeline either. And, for files with no kind of derogatory whatsoever, they still want to score them, so other things like inquiries, age and types of accounts count more. It's interesting you say Freddie Mac shows a bell curve. I suppose that represents the distribution of scores vs percentage of people. The curves I have seen look to be very lopsided, I think they are the score vs probability of default curves. Where on the FreddieMac site do they show that curve, I couldn't find it. Richard
|
| | Thursday, March 23, 2000 - 09:10 am Where is the site with the credit score model on it?
|
| | Thursday, March 23, 2000 - 09:33 am Sorry, if I remember correctly it is www.freecreditanalyzer.com
|
| | Thursday, March 23, 2000 - 09:43 am I've been missing it for a long time. Here's some, though: http://www.ftc.gov/bcp/creditscoring/present/sld016.htm http://www.ftc.gov/bcp/creditscoring/present/sld048.htm http://www.ftc.gov/bcp/creditscoring/present/sld041.htm
|
| | Thursday, March 23, 2000 - 11:31 am Thanks, Greg! Now I can throw out all of the paper I collected from the meeting. I'm surprised they posted this.... Kristy
|
| | Thursday, March 23, 2000 - 01:41 pm Another note, they just published this on March 3. 2000. Some of the most interesting screens are slide 8 (points for various aspects of your profile) and slides 44-49, where they go into scoring based on the fact that you live in a HMA (High minority Area).
|
| | Thursday, March 23, 2000 - 02:07 pm Kristy (In reference to slide 8) I downloaded the minutes of the meeting several months ago, and in it I remember McCorkell saying something like "...now don't go home and try to calculate your score with this...". I doubt the numbers in the table are even valid for the factors they are against.
|
| | Thursday, March 23, 2000 - 03:55 pm Yes, of course, but you can get a general feel for what they value, don't you think. Yes, Peter McCorkle did say that the info presented was not the whole picture. Did you go? Which room did you sit in?
|
| | Friday, March 24, 2000 - 06:38 am Kristy Unfortunately I didn't get to go. And no, based on what I have seen, I doubt the values on page 8 are right. The only thing about those last few lines I might believe is the shapes of the curves those numbers would generate, and even those would be questionable at best. For example, I haven't seen enough to know if 0 inquiries is the same as 2, if 1 is best, and if no record gets you a zero. And of course the age of the inquiries makes a difference. And is the inquiry curve really that shape, is there a hump or does it just go down from zero. So I wouldn't put much faith in that page. Greg Of those charts, I'd say 46 is the only one that has a chance of representing real data, and even then it looks squished (if 650 is in the middle,does the whole x axis truly cover the entire range). I think the rest are just there for illustration purposes. On a side note, if the graph on page 46 WERE from real data, (see if you can figure why I think it might) there are some interesting features to it, nes pas?
|
| | Friday, March 24, 2000 - 11:02 am oH, yEaH. There isn't any reason to believe it isn't the real thing since it is titled "Credit Bureau Score." At that point in the presentation, he was flipping though and skipping though the slides, running out of time-- though, that one is one of the most significant. This is the page where he talks about slide 46: http://www.creditscoring.com/pages/forumtranscript.htm#page56 The graph is sufficiently vague... "proprietary" (oh, brother), as they say. You think it is the real thing... because it has sharp angles? On another note, see the new links on the creditscoring.com home page.
|
| | Friday, March 24, 2000 - 01:52 pm Greg Just to be sure (not to nit), are we talking about page 8, it is titled "Scorecard"? Obviously things like the inquiry line, even if you believed those numbers, don't show any age weighting. Are those how inquiries count if they are <6 months old, 6-12, 12-24? Aside from that, the only reason I doubt it is that when I try to fit the data I have, I don't get numbers quite like those. It is a bit of a moot point though, because obviously there isn't enough info there anyway. Maybe they did give us the exact numbers. Who knows. I re-read Kristy's post, I might have misinterpreted what she said. If she meant, by "values", that the line items represent what they count, at least for the bottom of the sheet I agree. On the other hand, it is just my opinion. As for page 46, the thing I found interesting is that both lines have peaks in the same places. I have seen this in my attempts to do curve fits. It seems to result from things like score card changes. Coincidence? Maybe, maybe not.
|
| | Friday, March 24, 2000 - 06:41 pm My comments only referred to slide 46.
|
Credit Forum CreditCourt Forum 2003 Credit Suit CreditFactors Order Credit Reports |