Score Simulator details?

[poet]

I have a quibble with the Score Simulator. When you ask it about paying down all outstanding debt, it only gives one very general response based on 24 months. I'd like to know one simple thing:

Assuming that all accounts are paid on time for 24 months... will one's score go up faster if debts paid off all at once, THEN maintained for 24 months of on-time usage, OR (assuming 0-10% UTIL over that same period), would it be better/faster to pay down the existing debt over 24 months?

If anyone knows the best-case scenario (or a more finely-grained calculator), please speak up!

Thanks!

[Anonymous]

I prefer the FICO Estimator (link in my sig line).  In this way you get to "what if" by including or excluding certain derogs (if you expect to drop or GW or PFD) and you can see result in combinations by increasing number of inquiries, age of new accounts, number of accounts, balances, etc.

The Simulator only allows for you to take 1 item into consideration, and does not allow for the inclusion of derogs dropping or being disputed/deleted.

[MattH]

As TxJohn says, the Estimator is somewhat more flexible, because it allows a wider range of scenarios.  On the other hand, the Simulator starts from an actual score.  I like to use both.  With either tool, they intentionally give fairly broad ranges of predicted scores because if it gave exact predictions it could be used to reverse-engineer their algorithms, which are trade secrets.  Anybody who is familiar with statistical data mining knows in general terms the sorts of algorithms FICO uses -- I use very similar tools in my work analyzing pharmaceutical research data -- but FICO has invested a lot of their R&D in refining their methods over the years and would hardly want to reveal all that any more than my employers would want to release all our data.

In fact, much of my interest in the world of credit scoring comes from intellectual fascination with the data mining aspects.  In my world there are some public risk scoring algorithms, for example google Framingham Risk Score for an algorithm that estimates a person's risk of a heart attack.  Physicians use the Framingham Risk Score to help them determine which patients are at high risk of heart attack.

Message Edited by MattH on 01-23-2010 05:47 PM

[poet]

I'm less interested in reverse-engineering (I *wish* my math were that good!) than I am in being able to run more than one scenario simultaneously. For example, what if I paid off X amount of total outstading debt, paid off one of X collections, then paid everything on time for 3 mos/6mos/1yr/2yrs?

That would be far more useful, to me, in planning and budgeting than the current "just wait it out" conventional wisdom. 

I also tried Suze Orman's iPhone app, but the numbers were very screwed up. First, it warned me not to put the wrong numbers in, or else it wouldn't work right. So I put in the right numbers, to two decimal places. That made the app throw up. I went back and only put in the APRs to one decimal place. The app still returned completely illogical numbers (e.g., You have $24 worth of debt and your minimum payments come to $12 a month, when it's more like a couple of thousand in debt and maybe less than $100 in minimums). I used her online FICO kit and got the right figures, though. Maybe I need to reinstall the app and try again. At this point, though, it's purely academic, as I paid about 99% of my CC debt this weekend.

Now to PFD the sudden spate of collections...