On Dec 2, 11:21�pm, joetaxpayer <joetaxpa...[at]nospam.com> wrote:
- quote -

You can pay your bills on time and still not have a FICO score of
700. If you will check the way that FICO scores are tallied, you will
find that a large part is based on debt to credit limit ratio, amount
of credit applied for, types of credit used, etc. Therefore, if you
have 5 cards and all 5 are near the credit limit, and you have just so
happened to shop around for cards with lower interest rates, your
score is going to be less than 700. It's a catch 22!

On Dec 4, 3:08 pm, "Elle" <honda.lion...[at]nospam.earthlink.net> wrote:
- quote -

No, I'm not. I don't know how I can be any more explicit so I'm going
to stop trying. I had hoped that my simple example would drive the
point home, but I guess not. As you say, we'll just have to agree to
disagree.

- quote -

Again, no I'm not. A trend is a trend and it cannot be denied.
However, is a given trend globally applicable to EVERYONE in EVERY
PART of the parameter space? Without explicit knowledge of the FICO
function, that's a very difficult question to answer. Those are the
sorts of issues I'm raising. The conventional wisdom is based on
experimental evidence, NOT on explicit knowledge of the FICO function.

- quote -

Approximating the FICO function, specifically how well the
conventional wisdom surrounding the FICO score approximates the FICO
function.. That has been my topic all along. I've been very explicit
about it so I don't know how you've lost sight of that.

--Bill

"joetaxpayer" <joetaxpayer[at]nospam.com> wrote
- quote -

Among other things, I think he is saying that, even freezing
all variables but one, a person cannot know the effects,
even in general so as to be at least a little useful, of
changing that one variable.

- quote -

Seriously, I wonder whether the bigger lesson here is that
the two of you discuss your finances in detail and work
together to stay financially comfortable. :-)

"Bill Woessner" <woessner[at]gmail.com> wrote
- quote -

IMO, you are confusing the uncertainty surrounding inputs to
the FICO algorithm with alleged irregularities in output.
Until this point is cleared up, your question above appears
to me to be loaded. You have to remember that someone
somewhere says, "Oh, closing a line of credit with Lender X
is weighted abcyada, because of yada. Closing a line with
Lender Y is weighted defyada, because of yada." One cannot
say with precision what will happen to Jane Doe when she
closes a line of credit with Lender Z. Maybe you want to
hear people rephrase the advice above to "doing x will tend
to raise/lower.... " If so, many sites do have
qualifications like this, and your objection seems to me to
be more along the lines of "the writing is poor."

But you so far dismiss giving advice even about trends. You
cringe even when a person is advised to pay off all balances
to improve his/her credit score. I'd say you're in a tiny
minority who mistrust what FICO scores are said to be.

- quote -

Which problem? I cannot tell at this point.

I think we're just going to disagree and so continue to test
the poorly paid moderators. I am sure enough people with
reputable sources will continue to post to threads like
this. Your voice is heard. So are theirs.

Apply for the right one then once you get it merge the two together into
the right one.

You will take a few points hit for opening the new one and will recover
the points when you merge them. This is my experience and YMMV

phrankbooth[at]hotmail.com wrote:
- quote -

joetaxpayer wrote on [Tue, 4 Dec 2007 11:25:03 -0600]:
- quote -

Assuming she got the card under her own name, your score isn't affacted
at all. If she is signing you up for the account as well, the question
is why?

There is no "our FICO". There's one for you and one for her. Only joint
accounts and Authorised User (for now) accounts will count for both of
you. Which is why it's a good idea for a husband and a wife to have
seperate credit cards to maintain a good credit profile for each.

Bill Woessner wrote on [Tue, 4 Dec 2007 08:58:34 -0600]:
- quote -

There are multiple factors at play in closing a card. You are closing a
tradeline, but it will still be on your report for 7 years. The
shrinkage of your total available credit and assuming the card was at a
zero balance and there are other cards that are reporting a balance
(Most cards report the balance on statement closing date, so very few
paid in full report as zero) the growth of overall utlisation across all
lines.

On Dec 4, 12:10 pm, BreadWithS...[at]fractious.net wrote:
- quote -

Show me where Fair Isaac says "doing X will raise/lower your credit
score" and I'll believe it. All I've ever seen from Fair Isaac is
some approximate weighting of how the various categories impact your
credit score. That's a far cry from identifying, even at a very high
level, how the FICO algorithm works.

On Dec 4, 12:20 pm, "Elle" <honda.lion...[at]nospam.earthlink.net> wrote:
- quote -

In that case, I'm ESPECIALLY puzzled that you're not more supsect of
the conventional wisdom surrounding the FICO score. How can you hear
a statement like "doing X will raise/lower your credit score" and not
cringe? We're talking about (I believe) highly dynamic function. Do
you really think it's even possible to isolate a single variable and
say that the function varies monotonically with it? Forget about how
MUCH the function varies with a given variable, I'm not even convinced
the sign is constant.

For the discussion at hand, the purpose of the system is irrelevent.
The purpose of the APPROXIMATION (i.e. conventional wisdom) is to help
people raise their credit score. If you did system identification,
I'm sure you know George Box's famous quote, "All models are false but
some models are useful." The approximations are probably locally
accurate (there's Taylor's theorem at work again). They might even be
locally useful. I'm simply not convinced the approximations are good
enough to be globally accepted as accurate, even just to point of the
the sign being correct.

Just to make a concrete case, consider the function:

f(x, y) = x - 2xy + y

If you only sample the function in the set 0 < x < .5, no matter how
many samples you take, you'll conclude that f increases with y.
That's locally true and perhaps even useful. But if you try and
generalize your result to, say, .5 < x < 1, you'll be extremely
disappointed. This is a simple function of 2 variables. The FICO
function has dozens (hundreds?) of variables and is (probably) far
more complicated. Take the approximations on faith if you like. I'll
wait for a more rigorous treatment of the problem. And I won't hold
my breath.

--Bill

Elle wrote:
- quote -

I may be splitting hairs here, but I think you and Bill are talking
about two different things. I believe Bill is commenting on the
equations used, not on FICO correlating 100% to credit worthiness.
Bill suggests that even knowing the variables, we do not know the
coefficients for each. As his post states, there are some factors that
may contradict. You gain points for doing one thing, but lose on a
different criteria.

I agree with your overall comment, Elle. My wife goes to buy some big
ticket item (which, she deserves, of course, she works, and I am not
looking for veto power) but she pages me from the store that by opening
a store card she will get 20% off. $400 saving. I say to do it. If I
were FICO OCD'd I'd worry that the new credit inquiry, combined with a
store card that may only authorize $2000, which she'd use 100% would
somehow ding our FICO score. I only give it this thought now, and not
knowing or caring how FICO saw that one bill and if my score went down 5
pts or 10. I know how long it takes to clear $400, and the effort to
have her get the card, and me as I handle the bills, was worth that
savings. Did I become less creditworthy for that 3 or 4 weeks period?
JOE

"Bill Woessner" <woessner[at]gmail.com> wrote
- quote -

To /any/ degree of accuracy? We disagree. I will continue to
advocate to that which myfico.com, many reputable web
sites, as well as countless souls testify. Namely and
generally speaking, focusing in proportion to improve on the
five FICO categories will improve one's credit scores.

I think it's important to remember that FICO operates at no
small risk to liability, since Congress inter alia would
nail it if it was capricious in assigning credit scores,
which are so critical to people's fortunes.

- quote -

System identification (the correct term to use in place of
"reverse engineer") is the sort of analysis I did for
several years. Unfortunately those who focus strictly on the
numbers without considering the value of the input itself
can be distracted from the purpose of the system (=
function) in the first place. Approximations, without
nailing the function precisely, are perfectly useful. This
is particularly so when, as you should know, the output of a
model (= the algorithm) is only as good as the input. The
input rests on assumptions, so the ostensibly rigorous
output is subject to some interpretation. Thus the caveats
from FICO's authors and others, about granting the score
some leeway, are entirely appropriate. At the same time, the
big picture a FICO score provides remains the most reliable
we have at present.

Besides, we'd sure hear about it if someone did clearly
follow the general FICO guides for improving his/her credit
score and it failed to improve. We simply do not hear such
stories.

Bill Woessner <woessner[at]gmail.com> writes:

- quote -

But you'll ignore what Fair Isaac *themselves* have publicly
said about it? I figure not much (maybe a little) salt needed
for that.

Back to the original point - don't obsess over it, pay your
bills on time, etc. A few points here or there are not worth
paying any attention to. It's the big things - late and
missing payments - that are worth paying attention to.

Arguing over the minutia of the secret formula is a waste
of time, especially given the publicly available documents
from FIC themselves. I posted two links last time. They
don't give the details of the formula, but they say the
same things I and others have said here regarding this.

--
Plain Bread alone for e-mail, thanks. The rest gets trashed.
No HTML in E-Mail! -- http://www.expita.com/nomime.html
Are you posting responses that are easy for others to follow?
http://www.greenend.org.uk/rjk/2000/06/14/quoting

On Dec 4, 10:34 am, "Elle" <honda.lion...[at]nospam.earthlink.net> wrote:
- quote -

No, I don't think it's an exact science. But that's neither here nor
there. The important thing is that the FICO score IS exact (at
least... I HOPE it is; it would be really creepy if the algorithm were
nondeterministic). You give it a bunch of inputs and it spits out an
output. It's a function, plain and simple. It may not be analytic,
it may not be smooth, it may not even be continuous, but it is a
function. I believe, without any direct evidence, that it's a very
complicated function and, as such, it's extremely difficult to reverse
engineer to any degree of accuracy.

Just take a moment to think about how complicated it would be to try
and reverse engineer the FICO function. Do we know all the inputs?
Do we even know how many inputs there are? How many points within the
parameter space are we allowed to sample? To develop even a slightly
nonlinear approximation to the function, we need twice as many samples
as there are inputs. How many inputs are there, again?

I'm not saying it's impossible. I'm not even saying it hasn't been
done. But this is the sort of analysis I do every day and I know what
it takes to get it right. Until I see some sort of rigorous treatment
of the problem, I will continue to take conventional wisdom regarding
FICO scores with a grain of salt.

--Bill

"Bill Woessner" <woessner[at]gmail.com> wrote
- quote -

Just my opinion, but I think the problem is you are
insisting that evaluating a person's credit riskworthiness
is an exact science. The myfico.com site sheds some light on
why it is not. IMO, it's common sense. Any useful credit
algorithm has to rely on assumptions. E.g. it could assume
credit from one company is the same as any other. But
realistically, is this wise, supported by statistics about
credit reliability, etc.?

The FICO score's authors, from my reading, among other
things indicate that credit from one company may be treated
differently from another. Now of course this means some
subjectivity is applied when, say, a person shuts down
credit from Lender X as oppposed to shutting it down from
Lender Y. Then there's the question of quantifying how much
the difference should be! See how non-black-and-white this
is?

ISTM this is why it is said (IIRC by FICO authors and
lenders alike) that being in a certain range is what
determines whether one gets a certain loan, and that a 720
will frequently be read as no different from a 760.

It's not a perfect system, and it's some times annoyingly
expensive, but I can see that there is a need for such a
system. I have yet to read of a better one for evaluating a
person's risk and ensuring the freeing up of capital to
build business, homes, etc. and so improve the general human
condition.

On Dec 3, 2:24 pm, BreadWithS...[at]fractious.net wrote:
- quote -

I'm even doubtful of that. That assumes that the response to these
events are monotonic through the entire parameter space. Even
conventional wisdom says this isn't the case. You have to have open
lines of credit, but not too many. You have to USE your open lines of
credit, but not too much. See where I'm going with this?

In fact, there are even contradictions in the conventional wisdom.
Consider this exceprt from the Wikipedia page on the credit score:

"Length of credit history" is also a murky concept; it consists of
multiple factors -- two being the oldest account open and the average
length of time an account has been open.

Let's apply that statement to the OP's original question. The OP has
a credit card that's less than one year old. What effect will closing
it have on his credit score? According to the statement above (and
assuming he has an account older than the credit card in question),
his score should go UP. The oldest account open is unaffected and the
average age of accounts will go up. Yet the consensus here, which I
believe to be correct, is that his credit score will go down.

So I stand by my original statement. The FICO algorithm is very
complex. The parameter space is enormous and I doubt the function is
monotonic in ANY of the variables. Claiming to have reverse
engineered the FICO algorithm to any degree of accuracy is sort of
like cold fusion. It may sound good; people may even buy in to it.
But until I see a more rigorous analysis (or I land a job at Fair
Isaac), I will continue to view the FICO algorithm as a black box.

--Bill

On Sun, 2 Dec 2007 22:21:44 -0600, joetaxpayer
<joetaxpayer[at]nospam.com> wrote:

- quote -

Mine had a 50 point difference between 2 companies. Keep in mind that
not all lenders report to all 3 agencies.
Thumper

Elle wrote:
- quote -

Once again, you are a valued source of information. I'd not seen this
brochure before now.

This leaves me more curious as to how the financial institutions (e.g.
lenders, insurance providers, employers) treat these numbers. If 750+ is
treated as prime, and points within that range make no difference, then
40% of people have nothing to be concerned about, except, perhaps
dropping out of that level. If FICO is looked at closely so that every
20 points makes a real difference, that would surprise me. Suze Orman
refers to FICO constantly as if it were an SAT score, 600, a state
school, 750, IVY league. And yet, my experience mirrors John's, it never
seemed to be an issue. When I was in over my head with rental property,
the next bank said I had too much available credit, a condition of the
loan was to bring in the excess cards, with letters authorizing
cancellation. I brought that in, and walked away with a mortgage
commitment.
I have one card with my wife that we use regularly. Pay in full every
month. I now understand that if I wish to bump my FICO score a few
points, I should pay the bill before the invoice is cut, i.e. have the
bill itself always reflect zero or close to it. Or I can wait for the
bill and take the extra couple week's float. Whatever.
Since I am not in the market for any new mortgage, and even if I were,
I'd think the credit report would tell me enough.
JOE

"Bill Woessner" <woessner[at]gmail.com> wrote
- quote -

It depends on what you mean by "details." Numerous,
reputable, consumer-oriented sites give approximate
percentage breakdowns of the credit score like those
appearing at
http://en.wikipedia.org/wiki/FICO_sc...e_credit_score
and
http://banking.about.com/od/loans/a/ficocreditscore.htm .

The company that owns the FICO formula itself provides this
information at no charge via
http://www.myfico.com/Downloads/Brochures.aspx

Bill Woessner <woessner[at]gmail.com> writes:

- quote -

The details are guarded. The broader themes, however, are
discussed pretty openly by Fair Isaac.

Take a look at http://www.myfico.com/CreditEducation
and the forums that Fair Isaac maintains and moderates
on their site. http://ficofurms.myfico.com
Look at the first post under "Credit Scoring 101"

- quote -

It's hard for anyone to actually do that reverse engineering
because folks rarely can isolate specific credit events.
Nevertheless, while the *magnitude* of a specific event's
impact may not be easy to figure out, it's easy to know
that certain events are positive or negative, and in
some cases that two similar events may have different
impacts (ie. closing down a credit card you don't use -
if you have two with identical limits, but you've had
one - with a good payment history - for 15 yrs and you've
had the other - with similar payment history - for only
1 year - closing the newer one will have less impact. How
much less, nobody who knows is permitted to tell. But
that there is a difference is well publicized by FI.)

The bottom line, though, is that the *big* events
are the ones worth worrying about - make damned sure
you pay your bills on time, that you don't exceed
your credit limits, etc.

Agonizing over whether some small thing is going to
knock your scores around by a few points, however,
is effort and energy which could be used far more
productively.

Pay your bills on time, don't overextend, and stop
worrying about your credit score - unless your score
really stinks *and* you need to borrow money soon -
in which case, chances are that your problems are
the big things, not the little ones and your credit
score is probably not your biggest concern.

--
Plain Bread alone for e-mail, thanks. The rest gets trashed.
No HTML in E-Mail! -- http://www.expita.com/nomime.html
Are you posting responses that are easy for others to follow?
http://www.greenend.org.uk/rjk/2000/06/14/quoting

On Dec 3, 11:19 am, "Elle" <honda.lion...[at]nospam.earthlink.net> wrote:
- quote -

To be fair, I don't. But, like I said in my original post, the
details of the FICO algorithm are very closely guarded. I highly
doubt anyone who's actually privy to those details are going to post
them on Usenet. Fair Isaac would sue them faster than you can say
"intellectual property".

- quote -

Maybe, maybe not. My point is you just don't know. Yes, there's
plenty of conventional widsom involving the FICO algorithm. But
conventional wisdom is not necessarily correct, no matter how many
people believe it. There have been

It goes back to my basic point about the complexity of the algorithm.
Anyone who claims that "doing x will raise/lower your FICO score" is
essentially claiming to have reverse engineered the formula to some
degree of accuracy. As a mathematician, I'm suspicious of anyone
making such a claim.

--Bill

"Bill Woessner" <woessner[at]gmail.com> wrote
- quote -

How do you know this?

One can google for "credit score" and find a multitude of
sites that fairly consistently say the same thing about how
to maximize it. Asking here is fine, but the original poster
would do well to see the extensive discussion on the net on
credit scores.

- quote -

I think it's more accurate to say that the three credit
scores computed for an individual (by the three consumer
credit reporting agencies, Transunion, Experian and Equifax)
all will be significantly affected by certain actions. E.g.
keeping a high, not paid off, balance each month that
denotes a significant fraction of one's available credit.

Lastly, perhaps somewhat like Weeks, I am tired of the
deluded consumer who thinks he/she is commendable for having
a high credit score. It does not mean that a person is a
saver. It can mean simply that he/she likes taking on a lot
of debt. It says nothing about whether this debt is being
wisely spent.