Mr. Sanctions

What did I get, and could I have done better?

I wasn’t sure how to really weight the first 3 factors in my analysis… so I tried it 3 different ways. Here’s the prompt that ran my model:

Run the best fit procedure for all years from 2012 up to and including the current year. Rank the 3 best fits from best fit to worst fit.

Then create, in tabular format, the following details of the matched enforcement action:

  1. Label and  List  the name of the penalized parties
  2. Label and List the date of the enforcement action notice
  3. Label and list the amount of the fine, each on its own line
  4. Finally, explain each factor that matched and how well it matched, and each factor which didn’t match well. List the two sets of factors separately.

Do this 3 times:

  1. Weight  Factors 1, 2 and 3 identically
  2. Weight  Factor 1 higher than Factor 2 and Factor 2 higher than Factor 3
  3. Weight Factor 2 higher than Factor 1 and Factor 1 higher than Factor 3

At the end of each run, explain the models and calculations created and the weights used

and this is what I got back:

Based on the procedure you defined, I have compared the most recent OFAC enforcement action (Target) against the corpus of enforcement actions from 2012 through the present day.

The target action used for this analysis is:

  • Target: Gracetown, Inc.
  • Date: December 4, 2025
  • Characteristics: Violations of Russia-related sanctions (dealing in blocked property of Oleg Deripaska); Egregious; Not Voluntarily Self-Disclosed (Non-VSD); Willful disregard of explicit OFAC notice.

Below are the results of the 3 runs using the specified weighting models, followed by the ranking of the 3 best fits.

Ranking of the 3 Best Fits

  1. Best Fit:An Individual (King Holdings LLC) (2025)
    • Reasoning: This action is a near-perfect statistical twin to the target. Both involve willful dealing in blocked real estate/property connected to Russian oligarchs, both involved ignoring direct notices/subpoenas from OFAC, and both occupy the same “Egregious / Non-VSD” quadrant.
  2. Second Best Fit:Epsilon Electronics, Inc. (2014)
    • Reasoning: Despite the age gap, Epsilon is one of the few historical cases that strongly matches the “Egregious / Non-VSD” classification with a similar pattern of ignoring clear warning signs (red flags) regarding export destinations, mirroring Gracetown’s ignoring of the OFAC notice.
  3. Third Best Fit:Binance Holdings, Ltd. (2023)
    • Reasoning: While the scale differs (crypto/billions), it matches the critical Factor 3 (Egregious / Non-VSD) and Factor 1 (Willful behavior / Systemic disregard). It outscores other recent non-egregious cases (like IPI Partners) because the “Egregious” classification carries significant weight in the Penalty Grid factor.

Run 1: Equal Weighting

Model: Factors 1, 2, and 3 are weighted identically (1.0).

  • Calculation Note: Factor 4 is calculated at 70% of Factor 1 (0.7). Factors 5 and 6 are standard weight.

Match 1: An Individual (King Holdings LLC)

DetailsValue
Penalized PartyAn Individual (acting through King Holdings LLC)
DateNovember 24, 2025
Fine Amount$4,677,552

Factor Analysis

Matching Factors (Strong Match)Non-Matching / Weak Factors
Factor 1 (Behaviors): Score: High.
Perfect match on “Ignoring Warnings” (General Factor A). Both parties received actual notice from OFAC and willfully ignored it to continue dealing in blocked property.		Factor 5 (Period Length): Score: Medium.
Target violation period was ~2 years (Tier 2). Match violation period was ~1 year (Tier 2/3 border). Slight variance in duration tiers.
Factor 2 (Aggravating): Score: High.
Matches on Willfulness (Tier 1) and Harm to Sanctions Program (Tier 1).		Factor 6 (Mitigating): Score: Low.
Both had zero valid mitigating factors (beyond tolling agreements), essentially a “null match” which contributes little positive score.
Factor 3 (Penalty Grid): Score: Maximum.
Both are Egregious and Not Voluntarily Self-Disclosed. Same quadrant.
Factor 4 (Cooperation): Score: High.
Both failed to cooperate substantially (Target delayed reporting; Match submitted false certifications).

Match 2: Epsilon Electronics, Inc.

DetailsValue
Penalized PartyEpsilon Electronics, Inc.
DateJuly 18, 2014
Fine Amount$4,073,000

Factor Analysis

Matching Factors (Strong Match)Non-Matching / Weak Factors
Factor 3 (Penalty Grid): Score: Maximum.
Both classified as Egregious and Not Voluntarily Self-Disclosed.		Factor 1 (Behaviors): Score: Medium.
Match involved “ignoring red flags” (General Factor B) rather than “ignoring actual OFAC notice” (General Factor A). Similar intent, but different mechanism.
Factor 2 (Aggravating): Score: High.
Matches on Recklessness/Willfulness (Tier 1).		Factor 5 (Period Length): Score: Low.
Epsilon’s violations spanned ~4 years (Tier 1), whereas Target was ~2 years (Tier 2). Adjacent tiers receive intermediate score.

Match 3: Binance Holdings, Ltd.

DetailsValue
Penalized PartyBinance Holdings, Ltd.
DateNovember 21, 2023
Fine Amount$968,618,825

Factor Analysis

Matching Factors (Strong Match)Non-Matching / Weak Factors
Factor 3 (Penalty Grid): Score: Maximum.
Both classified as Egregious and Not Voluntarily Self-Disclosed.		Factor 1 (Behaviors): Score: Low.
Target was “dealing in blocked property.” Match was “failure to implement compliance/AML controls.” Different generalized behaviors.
Factor 2 (Aggravating): Score: Medium.
Matches on Willfulness, but Binance had significant “Management Awareness” (General Factor B) which was less central to Target’s specific notice violation.		Factor 4 (Cooperation): Score: Low.
Binance eventually cooperated significantly (settlement). Target cooperation was deficient.

Run 2: Behavior Dominant

Model: Factor 1 (Behaviors) > Factor 2 > Factor 3.

  • Weights: F1 (3.0), F2 (2.0), F3 (1.0).
  • Effect: This model emphasizes what they did (ignoring notices, dealing in property) over the final penalty classification.

Match 1: An Individual (King Holdings LLC)

DetailsValue
Penalized PartyAn Individual (acting through King Holdings LLC)
DateNovember 24, 2025
Fine Amount$4,677,552

Factor Analysis

Matching Factors (Very Strong Match)Non-Matching / Weak Factors
Factor 1 (Behaviors): Score: Extremely High (Weighted 3x).
The specific behavior of ignoring a direct OFAC notice is unique to these two cases. This tripled weight pushes this match far ahead of all others.		Factor 5 (Period Length): Score: Medium.
Minor mismatch in duration (1 year vs 2 years) is negligible compared to the massive behavior score.
Factor 4 (Cooperation): Score: High.
Weighted at 70% of Factor 1 (so 2.1x). Both demonstrated poor cooperation/bad faith.

Match 2: Epsilon Electronics, Inc.

DetailsValue
Penalized PartyEpsilon Electronics, Inc.
DateJuly 18, 2014
Fine Amount$4,073,000

Factor Analysis

Matching Factors (Good Match)Non-Matching / Weak Factors
Factor 2 (Aggravating):Score: High (Weighted 2x).
Reckless conduct matches well.		Factor 1 (Behaviors): Score: Moderate (Weighted 3x).
Epsilon ignored general red flags, not a specific OFAC cease-and-desist/notice. The 3x weight on this factor widens the gap between Epsilon and the “Individual” match above.

Match 3: Binance Holdings, Ltd.

DetailsValue
Penalized PartyBinance Holdings, Ltd.
DateNovember 21, 2023
Fine Amount$968,618,825

Factor Analysis

Matching Factors (Moderate Match)Non-Matching / Weak Factors
Factor 3 (Penalty Grid): Score: Low Impact (Weighted 1x).
The perfect grid match helps less here.		Factor 1 (Behaviors): Score: Low (Weighted 3x).
Systemic compliance failure (Binance) vs. Specific transactional violation (Target). The heavy weighting on behavior punishes this match significantly.

Run 3: Aggravating Factor Dominant

Model: Factor 2 (Aggravating) > Factor 1 > Factor 3.

  • Weights: F2 (3.0), F1 (2.0), F3 (1.0).
  • Effect: This model emphasizes the nature of the violation (Willfulness, Harm to Sanctions Program, Awareness) over the specific actions.

Match 1: An Individual (King Holdings LLC)

DetailsValue
Penalized PartyAn Individual (acting through King Holdings LLC)
DateNovember 24, 2025
Fine Amount$4,677,552

Factor Analysis

Matching Factors (Strong Match)Non-Matching / Weak Factors
Factor 2 (Aggravating): Score: Very High (Weighted 3x).
Both cases hit Tier 1 General Factors (A, B, C) for Willfulness and Awareness. The 3x weight cements this as the top rank.		(No significant weaknesses relative to other candidates)

Match 2: Epsilon Electronics, Inc.

DetailsValue
Penalized PartyEpsilon Electronics, Inc.
DateJuly 18, 2014
Fine Amount$4,073,000

Factor Analysis

Matching Factors (Strong Match)Non-Matching / Weak Factors
Factor 2 (Aggravating): Score: High (Weighted 3x).
Epsilon’s “Reckless disregard” scores similarly to Gracetown’s “Willful disregard” in Tier 1. This model favors Epsilon more than Run 2 did.		Factor 3 (Penalty Grid): Score: Low Impact (Weighted 1x).
The perfect grid match contributes less to the total score.

Match 3: Binance Holdings, Ltd.

DetailsValue
Penalized PartyBinance Holdings, Ltd.
DateNovember 21, 2023
Fine Amount$968,618,825

Factor Analysis

Matching Factors (Moderate Match)Non-Matching / Weak Factors
Factor 2 (Aggravating): Score: High (Weighted 3x).
Binance had massive aggravating factors regarding “Management Awareness” (General Factor B). This weights heavily here, potentially bringing it closer to Epsilon in score.		Factor 1 (Behaviors):Score: Moderate (Weighted 2x).
Still a mismatch on the specific behavioral mechanics.

Summary of Models and Weights

  • Model 1 (Equal Weight): Used a baseline of 1.0 for Factors 1, 2, and 3. Factor 4 was calculated as 0.7 * Factor 1 Score. This model provided a balanced view where the “Penalty Grid” (Egregious/Non-VSD) played a significant role in elevating Epsilon and Binance.
  • Model 2 (Behavior Weighted): Used weights F1=3.0, F2=2.0, F3=1.0. This model strongly favored An Individual (King Holdings) because it was the only case that shared the specific behavior of “Ignoring an OFAC Notice.” It penalized Binance heavily for having different underlying behaviors.
  • Model 3 (Aggravating Weighted): Used weights F2=3.0, F1=2.0, F3=1.0. This model focused on “Willfulness” and “Awareness.” It brought Epsilon and Binance closer to the top because all three cases involved high-level willfulness (Tier 1 General Factors), regardless of the specific transaction type.

At the end of the day, I think there’s value in all 3 approaches, as they answer different questions.

What do you think?

Leave a Reply

Discover more from Mr. Sanctions

Subscribe now to keep reading and get access to the full archive.

Continue reading