Buckets

[ficonightmare]

Robert and my-own-fico, thanks for the detailed explanation and graph. This makes sense.

My TU is now clean. Score jumped up 46 points and the only thing hurting it now on TU is "too many cards with balances". A few weeks ago it was "Short Account History and "Serious delinquency". I think I was re-bucketed on TU and short account history isn't making enough of a impact to count in this new "bucket".

Then again, two TL's were deleted. Aside from them being negatvie TL's, one was 7 years old, the other 5 years. So even though my score jumped, I'm wondering what would have happend if they stayed but no negs were associated with them. Ahh well...

[RobertEG]

Congrats!  You are apparently now hanging from branch 3 or 4!

[guydiver]

RobertEG: That is just plain awesome.  I have been hanging out on these forums for several years now, reading thousands of posts and learning the FICO process, trying to rebuild my credit, and in some small way occasionally and possibly helping another along a path that I myself have already been down. This post just boggles my mind and reminds me how little I do know,(or think I know). Anyway, thanks for a great post. Now I gotta get to that other tree limb over there

[Anonymous]

---

@RobertEG wrote:

However, there is a published document by a major US creditor,  First USA Bank, that presents a detailed disclosure of their risk scoring of consumer based on various categories of consideration, .  It is U.S. Patent No. 6,202,053.  I call this the "Bucket Patent." 

---

The elephant in the room, of course, is that this patent is for a fako scoring algorithm and may or may not operate in the same way as the FICO algorithm's buckets.

My favorite description of bucketing is given by John R. Ulzheimer, who worked with FICO a while back and is familiar with their particular flavor of bucketing.  In his book, he devotes a chapter (about four pages) to "Scorecards", which we often refer to as 'buckets'.

In my mind, it's like trying to determine what's in the Pepsi Formula.  Pepsi has a patent on that formula and it's not published for obvious reasons.  If I go to work for Coke, and learn one of their ingredients; it doesn't mean that I can whip up a batch of Pepsi.  Pepsi may use the ingredient but in a different proportion, or they may not use the ingredient at all.  Or they may add the ingredient during a different point of the process.  Ten year's later, Coke itself may use the ingredient in a different way, a different proportion, or not at all; based on their ongoing research.

While you're looking at the "bucket" process; another factor is the "bins" which are within the "buckets".  Some experts believe that a bin change can sometimes yield a more significant FICO score impact than a bucket change.

As we keep our focus on this link for actions that will build our FICO's, we'll find ourself in a good spot, even if we're not able to definitively label our current "bucket" or scorecard; or the potentially significant impact of the bins within those buckets.

Whew!  It's definitely time to find some breakfast - what on earth am I thinking?  Posting before oatmeal (or a donut ), or for that matter - a Pepsi -  in the AM?

[guydiver]

It's sort of funny that when someone happens to mention "my bucket list" in the topic of conversation, my mind isn't going in the same direction as theirs. Guess I've been hangin' out on these forums too long. 

[vanillabean]

---

@Anonymous wrote:

My favorite description of bucketing is given by John R. Ulzheimer, who worked with FICO a while back and is familiar with their particular flavor of bucketing.  In his book, he devotes a chapter (about four pages) to "Scorecards", which we often refer to as 'buckets'.

---

Thank you! A major headache reliever for me. DW will be so excited when she finds this literary work under the tree.

Just kidding. It would be a sure death sentence to do so. But I did purchase it for myself; has been shipped already.

[RobertEG]

BeamMeUp, I totally agree with your caution of using the bucket patent I have referred to as any specific indication of how Fair Isaac does it!!!!

Fair Isaac does it own thing, but it appears to me to be remarkably similar.

It is just, in my opinion, the only truly published document by a major creditor that really discusses this issue in deptt.  And your are also correct in observing that that  patent does not even go into actual credit scoring at all once bucket placement is determined.  It is just a bucketing algorithm. not a scoring algorthm that in any way deals, once you are categorized, or bucketed, how that will actually affect the actual wieghting of information in your credit file. 

It is very rare to ever see anyone who does credit evaluation to ever apply for a U.S. Patent.  I was an examiner, supervisor, and senior executive with the U.S. Patent Office for 31 years, and now am a registered patent agent with the Patent Office.  I can tell you that there is no such thing, period, as a "secret patent."  To obtain a patent requires a full, published disclosure of how to make and use the invention at the time of application for the patent, and it will be published 18-months hereafter, .  Neither Pepsi of Coke have ever applied for, or received a U.S. Patent on their formulations.  They have chosen to forgo patent protection, and rely on protection in the form of maintaining it as a trade secret.  Once they choose to market their product, and choose not to secure a patent, they loose any and all rights to thereafter apply for and obtain a patent once one year has passed from the date they put it into public use or sale..  So if someone does not apply for a patent, andt thus chooses not, put it into the public domain, and also has a invention that is not subject to easy  reverse-enginnering, then their invention could thus be valuable far after the dates of expiration of a U.S. patent, which is only 20 years.  Coke has won that gamble for over 100 years.  But it is not because they, or even Pepsi, has any "secret patent."

They have NO patent protection, and cannot now even apply for it.

Thus, that is why you see such a pausity of any patents or voluntary, detailed discloure of both bucketing and credit scoring algorithms.  Most companies, such as Fair Isaac, see the benefit of NOT making public disclosure in exchange for a U.S Patent, and thus just rely on their abiltiy to keep their algorithms secret, and to try to prevent their reverse engineering.  That is the primary reason why Fair Isaac wont provide details on eiher their initiallly bucketing algorithms, or their final scoring algorithms.  Also, for example, why their scoring simulator is so vague in letting you see multiple impacts on multple categories at the same time, or very short term scoring imppacts.  .  They are protecting their intellectual property right, which is the entire key to their business.

I doubt seriously that Mr. Ultheimer's book discloses anything more about Fair Isaac bucketing than we already know, even if he once worked for them.  Fair Isaac protects confidentiality of its trade secrets by non-disc;osure agreements with all who they permit access to it.

In my humbled opinion, the amount of infromation willingly put into the public domain by Fair Isaac as a means to assist consumers in understanding credit scoring surpasses anything I have witnessed in my professional career by a company whose business is prevent reverse engineering of its trade secret product.