Consumer Scoring- Insider's Article

That's kind of interesting to say the least.

It should give us all a new respect for worthknowing.com.
Most of those posting about worthknowing who have berated them should come up with a new understanding and a new respect for them.

Here is why.
Most of us, this author included have thought therm quite worthless in the past because of their radically lower numbers compared to Fair Issac and other scoring models, so we pretty muchly thought them to be off the wall and out to lunch.

Now we see that they are using the bell curve for their scoring model which is quite naturally going to yield much lower numbers than almost any other scoring model, but which also makes much more sense once you know what it is and how it works.

For those readers not mathematically inclined, I'll give a greatly simplified and therefore much less accurate description than would be generated if I simply put the mathematical formula for a bell curve and let it go at that.

The bell curve is a mathematical model somewhat like a sine wave or sine curve which is much easier to explain, so I will use that instead.
A sine wave is a mathematical representation of what happens to voltage or current as a generator is turned through 1 360 degree rotation over a given amount of time. It starts at 0 &#186 and slowly turns for &#189 rotation which is the top of the curve or 90 &#186 which equals the maximum amount of voltage generated then starts to degrade back to 0 volts output at 180 &#186 then proceeds to go negative until the maximum negative voltage point is reached at -270 &#186 and then starts back to zero or 360 &#186 making 1 complete revolution in the process.

So what you get is much like what you would get if you cut a circle in &#189 and then twisted the bottom &#189 so that it comes after the top &#189 of the circle instead of below it.

A bell curve is simply a curve that looks like a big bell, having about the same shape as the Liberty Bell. Takes a considerably more complicated mathematical equation to represent than does a sine wave. Then a statistically large number of credit scores taken from another model such as FICO or whatever is used is plotted on this curve, and those who fall in the lower percentiles are going to come up with scores like 5 or 10 or 15 while those with a higher FICO score are going to come up with maybe a 50 or whatever, depending on how they structured the parameters of the bell curve.

In other words, it's how your credit ratings rank compared to the general population while FICO and other models will show scores based on other and completely different criteria.
Those will give or take points based on different parameters such as say 10 points off of a theoretically perfect score for say a negative item such as an inquiry or 100 points off for a bankruptcy. These types of systems are much more linear than are the bell curve models.

The next logical question is going to be, "which is the best type of model to rely on for accuract", and the answer to that is that there is no real answer. It's going to depend upon which model's owner has the most financial and political clout and is perceived to be the best by the majority of lenders over time.

So what is our answer? Are we all going to be turned into statistical mathematicians by Brutus & company? While there are several of us here on credit net who are perfectly capable of going along with the trend, maybe even developing our own models using multiple regression or other such forumlae, most folks are going to be much more likely to use the Popeye model which consists of two toots on the corncob pipe

As far as I am concerned, I yam what I yam and all your fancy scoring models really ain't worth two toots.

 

Well, IMO scoring is a love-hate relationship. If it weren't for "scoring", folks wouldn't get instant credit approvals; they wouldn't be able to walk in a car dealership and drive away a $50,000 car the same day. They wouldn't be instantly considered for a mortgage.

I know you remember the *old* system, Bill. If you wanted a loan, you made an appointment to see the loan officer. Then, you had to haul in all your banking records, pay stubs, utility bills, credit records. You had to provide up to 5 credit references. You had to provide up to 5 personal references. The forms were 2 feet long just to buy some furniture. Every part of your personal life could be scrutinized.

And, heaven forbid if the loan officer didn't LIKE you! You could get turned down for any reason and they didn't even have to tell you why. There are many evil implications here which I won't go in to, but suffice to say many folks were turned down for reasons not having anything to do with credit.

Surely it's not perfect, but it's wayyyyy better than it used to be.

SK

 

Bill:

Thanks for "greatly simplifying" this for us. Sine curves? I haven't used those since college math. I had to ask my kid. What does it sound like when you don't simplify something?

 

Now that's a Funny! I was afraid I was the only one catching MEGO* Disease.

* MEGO= My Eyes Glaze Over


That aside, Bill, I do get your point: WorthKnowing is not as WorthLess as many of us thought. I've come to respect your opinion on most credit related matters, so that's good enough for me.

 

Since Bill brought up worthknowing, I am interested in how the scores of others match up between worthknowing and the CRA's.

I have about 700 on my TU and an even 50 on worthknowing.

 

Re: Consumer Scoring- Insider's Art

Thanks, Bkev, super article.

Doc

 

Re: Consumer Scoring- Insider's Art

Well, be all them fancy numbers aside, as I said at the end of my rant, tirade, whatever anyone want to call it,

I YAM WHAT I YAM AND I YAM WHAT I YAM.

TOOT!!! TOOT!!!

So much for all them facncy numbers.

LOL