topic Re: Re-bucketing in Understanding FICO® Scoring

Re-bucketing

Anonymous — Tue, 05 Jun 2012 23:02:57 GMT

Can someone explain this to me? Seems this may be the source of my score drop when a large 30K collection was removed on EX.

Re: Re-bucketing

Anonymous — Tue, 05 Jun 2012 23:43:38 GMT

Rebucketing is the concept that FICO puts folks into "buckets" based on a broad criteria. For example, a bucket could be "people who have a late payment," or "people who have two or more revolving accounts." Moving from one bucket to another - theoretically - affects how some items on your report are weighed.

To my knowledge, these buckets are not technically proved to even exist, but are used to explain large score differences when something big happens (I, for example, saw an ~80 point score increase after my last collection fell off... Now, I expect my score to be more sensitive to inquiries and util than before.)

IN your case, it could be a few factors that dropped your score - AAoA comes to mind. Others might have different info than above, of course

Re: Re-bucketing

llecs — Wed, 06 Jun 2012 00:10:47 GMT

Is the EX score you were looking at a FICO or FAKO?

Re: Re-bucketing

RobertEG — Wed, 06 Jun 2012 01:12:38 GMT

The use of segmentation of the population into sub-categories is not conjecture..... it is a factual part of FICO scoring. FICO has discussed the use of segmentation of scorecards as part of its scoring methodology on numerous occasions. See, for example, the FICO White Paper titled "Using Adaptive Random Trees (ART) for Optimal Scorecard Segmentation," Chris Ralph, Analytical Science Director, Fair Isaac Corp., April, 2006.

The great unknown is exactly what those segments are, and the corresponding shift in scoring parameter weight based on the segment category of the file.

Re: Re-bucketing

Anonymous — Wed, 06 Jun 2012 01:38:57 GMT

It was a new collection from late last year that was deleted and it was for a Private SL.

It dropped my FAKO score from 513 to 490. (Score from credit one pre-qual soft hits)

Re: Re-bucketing

Anonymous — Wed, 06 Jun 2012 01:36:50 GMT

@llecs wrote:
Is the EX score you were looking at a FICO or FAKO?

Didn't even catch that...

Robert, I stand partially corrected But again, no one really knows exactly what the buckets are.

Re: Re-bucketing

marty56 — Wed, 06 Jun 2012 01:42:20 GMT

Forget the FAKO score. Your reall FICO score good have changed in the oposite direction or not at all.

Re: Re-bucketing

Anonymous — Wed, 06 Jun 2012 02:45:56 GMT

Oh ok, well since it's EX, I won't find out until I do another lender pull in a few months.

Re: Re-bucketing

llecs — Wed, 06 Jun 2012 02:56:34 GMT

I don't think FAKOs are that smart. Stupid EX.

Re: Re-bucketing

Tuscani — Wed, 06 Jun 2012 15:37:55 GMT

The purpose of pools (aka scorecards, buckets, ect.) works out more to the benefit of those with derogs or very little to no history than it does them harm.

If there was only one simple bell curve including all credit files then those fortunate, extremely long established folks would have the whole top half of the score numbers scale absolutely locked up tight.

So where would that leave the person who is only beginning to build a credit history at all? Or who has had some derogs in the past but now is trying to work toward a better CR and score? They would be locked down into the basement of the score range for probably ten, twenty years or more. There's no WAY that that new user or rebuilder individual has as much good history yet as that long established individual does.

So what the pools are attempting to do is to help potential lenders figure out which of the new-credit individuals are showing characteristics which usually go on to blossom into a long clean history, and which other ones don't. The same for rebuilding individuals. By assigning relatively better scores to top of each scorecard and lower scores to the least improved, the individuals who are making progress float to the top. They actually are scoring a bit better than they would if being compared head to head with Mr. Jones who has no derogs whatsoever and has 30+ years of history already.

By trying to compare apples to apples and oranges to oranges, that means that amongst the pool of individuals who share certain key history elements, Customer A looks most like someone who will continue to do better and better and is not as likely to default whereas Customer B from the same group is not showing those indicators of steady improvement.

If it weren't for scorecards, IMO, it would be very difficult to get decent mortgages and accounts and loans without thirty years of history already established. That would mean that it would be like climbing an ice mountain barefoot to buy a house or a car or open a credit card before the age of at least forty or fifty.

The pools tend to be split out according to:

-Age of file (length of credit history)

-Thickness of file (number of trade lines)

-Presence of a new account (opened within past X months)

-Presence of seriously negative payment history (90+ days late, charge off, etc.)

Re: Re-bucketing

Anonymous — Fri, 08 Jun 2012 00:11:15 GMT

Your credit score: 482
Date: 06/07/2012
Scores range from a low of 350 to a high of 850

Does this mean it is a FICO, got it from Credit One pre-qual.

If so, that's a drop of 31points.

Re: Re-bucketing

llecs — Fri, 08 Jun 2012 03:11:59 GMT

It's not a FICO. Could be an internal score maybe?

Re: Re-bucketing

Revelate — Fri, 08 Jun 2012 22:25:28 GMT

@llecs wrote:
It's not a FICO. Could be an internal score maybe?

Now I'm confused Isn't 350-850 the traditional FICO score ranges?

Granted there's a real max which isn't 850 but I didn't think that had to be stated on a score report? Also, with a Beacon 5.0 pull of sub 450 in April, might not a 480 be accurate?

Re: Re-bucketing

llecs — Sat, 09 Jun 2012 01:38:13 GMT

@Revelate wrote:
@llecs wrote:
It's not a FICO. Could be an internal score maybe?
Now I'm confused Isn't 350-850 the traditional FICO score ranges?

Granted there's a real max which isn't 850 but I didn't think that had to be stated on a score report? Also, with a Beacon 5.0 pull of sub 450 in April, might not a 480 be accurate?

FICO is 300-850 vs. 350-850. And yes, FICO isn't a perfect 300-850, but I wouldn't think the bank, when they published the score range, would round the 300 up to 350.

Re: Re-bucketing

Revelate — Sat, 09 Jun 2012 02:21:41 GMT

@llecs wrote:
@Revelate wrote:
@llecs wrote:
It's not a FICO. Could be an internal score maybe?
Now I'm confused Isn't 350-850 the traditional FICO score ranges?

Granted there's a real max which isn't 850 but I didn't think that had to be stated on a score report? Also, with a Beacon 5.0 pull of sub 450 in April, might not a 480 be accurate?

FICO is 300-850 vs. 350-850. And yes, FICO isn't a perfect 300-850, but I wouldn't think the bank, when they published the score range, would round the 300 up to 350.

Right you are, I was indeed confused and my quick GoogleFu failed me when I went looking. Thanks illecs for the smack upside the head on my apparently mixing it up with the CE score or some other FAKO.

Re: Re-bucketing

llecs — Sat, 09 Jun 2012 12:40:02 GMT

I get mixed up every now and then too. It seems like there's a FAKO with 350-850. Transrisk maybe? Def. not saying OP's was that though. Probably an internal score.

Re: Re-bucketing

HiLine — Mon, 09 Feb 2015 19:29:32 GMT

@Tuscani wrote:
The purpose of pools (aka scorecards, buckets, ect.) works out more to the benefit of those with derogs or very little to no history than it does them harm.

If there was only one simple bell curve including all credit files then those fortunate, extremely long established folks would have the whole top half of the score numbers scale absolutely locked up tight.

So where would that leave the person who is only beginning to build a credit history at all? Or who has had some derogs in the past but now is trying to work toward a better CR and score? They would be locked down into the basement of the score range for probably ten, twenty years or more. There's no WAY that that new user or rebuilder individual has as much good history yet as that long established individual does.

So what the pools are attempting to do is to help potential lenders figure out which of the new-credit individuals are showing characteristics which usually go on to blossom into a long clean history, and which other ones don't. The same for rebuilding individuals. By assigning relatively better scores to top of each scorecard and lower scores to the least improved, the individuals who are making progress float to the top. They actually are scoring a bit better than they would if being compared head to head with Mr. Jones who has no derogs whatsoever and has 30+ years of history already.

By trying to compare apples to apples and oranges to oranges, that means that amongst the pool of individuals who share certain key history elements, Customer A looks most like someone who will continue to do better and better and is not as likely to default whereas Customer B from the same group is not showing those indicators of steady improvement.

If it weren't for scorecards, IMO, it would be very difficult to get decent mortgages and accounts and loans without thirty years of history already established. That would mean that it would be like climbing an ice mountain barefoot to buy a house or a car or open a credit card before the age of at least forty or fifty.

The pools tend to be split out according to:

-Age of file (length of credit history)
-Thickness of file (number of trade lines)
-Presence of a new account (opened within past X months)
-Presence of seriously negative payment history (90+ days late, charge off, etc.)

Thanks for the detailed explanation!