@Thomas_Thumb wrote:It can be insightful to graph the scores using Excel. Given your data I would put together 3 graphs: one each with EQ scores, TU scores and EX scores by model with date on the x axis.
I'll graph it and post tomorrow. Thanks for the suggestion.
For me 10 my lowest non industry score across the board, with 9 and 10T being my highest (9 highest on EQ and EX, 10T highest on TU). I'm intrigued as to why your 10T was hit worse than 10. 10T seems to be a bit more forgiving to dirty profiles than 10 if you keep super low utilization, but since you have clean profile I don't know if the same applies.
@Thomas_Thumb wrote:It can be insightful to graph the scores using Excel. Given your data I would put together 3 graphs: one each with EQ scores, TU scores and EX scores by model with date on the x axis.
@Thomas_Thumb I did this really quick so there's only the Standard (No Auto or BankCard) scores graphed. I can add them/replace something later.
Recap:
New Inquiry on EX only April 6th, reporting April 8th.
Closed 2 Sync accouts, reporting on April 13th.
No revolvers reporting balances, April 15th (No AZEO)
New Chase account reported, May 3rd.
AZEO (1 account @Anonymous%), May 3rd.
No other changes.. yet.
My main focus is how FICO 10T is manipulated. I've been throwing everything at them to see how they react. My gut says the 10T's respond more like FICO Score 2, 4 and 5 with the excception of Transunuion FICO 4 and FICO 10T looking like outliers. I don't know enough about the FICO 4 model to understand how that model sees risk differently from EX FICO 2 and EQ FICO 5.
@Thomas_Thumb wrote:* It looks like a trendline is missing on the TU graph.
Interesting on how EX Fico 10T went up 11 points on 4/26 and then nose dived 28 points on 5/3. EQ 10T had no spike but does show a nose dive on 5/5 (2 day delay). Clearly 5/3 was a trigger event. An account reporting delay on EQ vs EX?
Scores appear to be more erratic on EX. Anything different in your EX file outside of the seeking new credit inquiry? In general EX is less sensitive to # of accounts with balances.
EX score 2 (which is a Fico 98 model) was the only other score to drop in a manner similar to 10T. It looks at AMEX charge card B/HB as a scoring metric - Fico 04 (EQ score 5 and TU score 4), Fico 8 and Fico 9 do not. These scores spiked up when the others dropped. Why the spike? New account reach 90 days?
Did you payoff revolving credit which spiked those scores but also post a sizeable balance on an account like an AMEX charge card?
It appears to me that 10T has some metric(s) trends # of open accounts and # of new account openings as metrics not considered in the other models. It looks like 10T is also tweaked by CRA ... Or not all CRAs are capturing the desired account data for trending.
Hard to see much on my phone. May take a closer look on my laptop later.
@Thomas_Thumb I completely dissected Experian since it has had the most action. This from all models of Experian ONLY.
I did not have an AMEX Charge report during this time period. I went back and re-called all of my Experian reports to trace back every account. The timing was off on the 26th. It was actually the 23rd when the spike accurred for EX10T (which also affected EX AUTO 8 by a point and EQ 10T by 2 points). The culprit was a personal loan had reached the 76% paid threshold. I listed every variable reported to EX for the 36 days I have daily data for below.
Just had a bounce on these scores ONLY. One authorized user account reported a $25 balance today. There are no other changes to any other models, no changes to length of credit, no changes to number of accounts reporting balances and no changes to the amount owed vs credit limit. The AU account is the only effective change and is only reported to Experian. Equifax and Transunion do not show this update, yet. I was under the impression AU accounts should have no affect on FICO 8 models. I certainly did not have any reason to expect FICO 10T to be affected since FICO 10T is based on trended data.
EX FICO SCORE 8 - 839 > 847
EX FICO AUTO SCORE8 - 873 > 882
EX FICO BANKCARD SCORE8 - 882 > 891
EX FICO 10T - 818 > 825
@Thomas_Thumb thanks for those links. I'm going to read through them tomorrow evening and Sunday. I'm also going to crawl back through my reports to see if there were any reported AU accounts with balances. My gut says none were reported but I will confirm. I'm also waiting for EQ and TU to update their reporting of the same AU account. I'll post when those updates similarly.
What piqued my interest was the seeing FICO 10T increase with an AU account reported balance. I'm trying to understand how an AU account could affect trended data. Do AU accounts have any weight for FICO 10T? Any ideas?
Also, if you have any suggestions as to what I should do to test FICO 10T, let me know.
Some interesting data.
It's safe to say that recently closed accounts have no affect on anything including FICO 10T as the two recently closed accounts did not move any scores on any models. I am starting to see how FICO 10T differs from Equifax, Experian and Transunion, even with identical data being provided. My Experian FICO 10T scores are the most sensitive when processing new account reporting and AZEO out of the three bureaus. All FICO 10T's reacted negatively to AZEO combined with the new account reporting whereas nearly all other models reacted positively. Transunion was the most lenient to anything negative. Equifax seems to be the most sesnitive to number of accounts reporting balances and does not appear to follow the 1% UTI to register threshold.
Transunion and Equifax numbers are now coming together.
- May 3rd TU reported a balance on a revolver (Chase Amazon, $158/$8,000 = .19% UTI)that made no impact for AZEO.
- May 5th EQ reported a balance on a revolver (Chase Amazon, $158/$8,000 = .019% UTI) that did register as AZEO increasing FICO 8 to 850 (max). Same day, EQ then reported the new Chase CSP which dropped EQ FICO 8 back to 845 for not being AZEO then (oddly) later that same day, Chase Amazon updated to $0 reviving AZEO for EQ and bumping FICO 8 back to 850 (max).
- May 5th both TU & EQ reported AZEO and also reported the new Chase account that bumped every score except FICO 10T which lost 8 points on TU and 15 points on EQ.
- TU stopped reporting a CFA (still reporting to EQ only) that closed in 2015 on May 7th that provided a nice score bump from 1 to 6 points accross all TU scores.
- May 9th TU reports a second revolver (AMEX $224/$30,000 = .74% UTI) with no effect on AZEO. All scores on TU maintain.
- May 11th EQ also reports a second revolver (AMEX $224/$30,000 = .74% UTI) with dropping FICO 5 and 8 (all verisons).
- May 13th TU Reports a balance for an AU Acct ($25/$20,000 = .125% UTI) and a 3rd revolver $633 / $20,000 = 3.16%) increasing ALL scores accross TU including FICO 10T by 6 points however FICO AUTO 4 and BANKCARD 4 are not influenced.
- May 13th EQ Reports a balance for an AU Acct ($25/$20,000 = .125% UTI) and a 3rd revolver $633 / $20,000 = 3.16%) dropping FICO 5 (all 3 versions) plus FICO AUTO 9.. but increases FICO 10T by 6 points, (not sure yet how this helps a trended score).
