Long reads
These are longer pieces on how manipulation works in practice: case teardowns, the forensics of detection, the distinctions people get wrong, and analysis of this site's own enforcement data. Each links out to the technique pages and case records behind it.
-
Six regulators compared
What the library's records look like for the SEC, CFTC, ASIC, the Ontario Securities Commission, the FCA and the AMF in France: technique mix, money, outcome, bans and criminal cases, with penalties kept in each regulator's own currency and no conversion. A comparison of the records collected, not of the regulators.
-
Civil cases with a criminal parallel
About three in ten of the library's 2,637 records carry a corrected "criminal parallel" flag. This post measures the corrected flag by regulator, technique and outcome, how far the criminal cases went, what it still cannot show, and the matters that span more than one regulator.
-
Enforcement by year and agency
Records per filing year by agency and by technique, read after measuring where each source starts and stops. Most apparent waves are the library's own coverage; a few are not.
-
How cases begin
A phrase-based reading of 2,333 regulator documents for any statement about where a case came from or who helped. Three documents in a hundred say how a case began, about four in ten thank another body, and the rest are silent. Silence means the document does not say, not that nobody referred the matter.
-
How long enforcement cases take
After the 4 October 2026 status rounds, 622 of 2,637 records can be timed from filing to a recorded end, with a median of about 1.6 years. Of the 733 once marked "filed", 543 are now resolved and 190 remain filed, an upper bound on the truly pending.
-
Landmark matter, Control-Finance 2021
A $429 million CFTC penalty in a bitcoin fraud case that no defendant contested. What the complaint alleged, what the 2021 default judgment found, why the penalty is three times the loss, and why the money figure and the Ponzi tag both need caveats.
-
Landmark matter, Glencore 2022
The CFTC's May 2022 order against three Glencore companies: what it finds about oil benchmark windows, bribery and stolen information, how the $1.186 billion splits, why up to $852.8 million is offset by Justice Department payments, and what the order leaves unproven.
-
JPMorgan 2020: the metals and Treasuries spoofing matter
What the CFTC and SEC orders of 29 September 2020 against JPMorgan entities found, who was a respondent, how the $920.2 million CFTC total divides, how it overlaps with the SEC and Justice Department resolutions, and what the matter does not establish.
-
The 2013 and 2014 benchmark rigging orders
Fifteen records cover the CFTC and UK regulator orders of 2013 and 2014 over LIBOR, Euribor and foreign exchange benchmarks. This is what the orders and final notices say, who the respondents were, which sums are in which currency, and what the matters leave open.
-
The 29 January 2018 anti-spoofing announcement
Eight CFTC spoofing actions announced on one day in January 2018, against three banks and six individuals, read against the releases and later orders: what was alleged, what was found, what each respondent paid, and what the day does not show.
-
The Stanford judgment, the library's largest penalty
One SEC litigation release records the largest civil penalty in the library, $5.9 billion on R. Allen Stanford. This is what the 2025 judgments say, who they bind, which numbers are penalties and which are not, and why the figure bends every total it touches.
-
What sanctions the records show
Civil penalties, disgorgement, prison terms and bars across 2,637 enforcement records, by regulator, technique, outcome and respondent. Medians, quartiles, the concentration of dollars in a handful of records, and every result with and without the outlier.
-
Who appears more than once
Of 4,219 names in the library's defendant lists, 669 appear in two or more records, but only 46 appear in separate enforcement episodes. Most of those 46 are banks, securities firms and other market infrastructure, and a repeat name is not a finding of repeat misconduct. This is how the names were counted, what the repeats are, and what the counts cannot say.
-
What reading every record found
On 2 and 3 October 2026 the library's case records were read against the regulators' own documents, by Claude AI agents under written instructions and with no lawyer involved. One record in five lost a technique tag, and money, status or names changed on many more. This is the method, the numbers, the recurring causes and the limits.
-
The binary options "trades" that were coin flips the house controlled
Four CFTC matters against binary-options platforms — Option Mint, BeeOptions, Blue Bit Banc and Yukom Communications — did not just allege bad odds or misrepresented risk. Each alleged or found that the platform's own software, or an employee acting for it, set the result of a customer's trade directly, regardless of where the real market finished.
-
Eleven records were tagged chat group pumps. One of them is.
This library tagged eleven enforcement records chat-group-pumps until its 2 October 2026 audit. Reading each primary document against the technique's own definition — organisers accumulate, announce in a group, sell into the response — only one matter fits. The other ten involve a chat platform for an unrelated reason, and the audit removed the tag from all ten; it now sits only on that one matter's three ASIC records.
-
Nine of the library's 25 churning records aren't churning
When it was audited on 2026-09-29 the library tagged 25 enforcement records churning; it now tags 14. Reading the primary document behind each of the 25 showed nine describe something else — three are reverse churning, the mirror-image conduct of charging a fee on an account that barely trades, and six mention churning only in a defendant's background or a firm's compliance program, never in the conduct actually charged. The sixteen that held up on that date show the exact turnover and cost-to-equity math the technique page describes.
-
Twenty-Five Reverse Merger Records, and Fourteen Name the Gatekeeper
Of the 25 case records this library now tags reverse merger schemes, fourteen name a transfer agent, broker-dealer, attorney or accountant as a defendant rather than only a promoter or issuer — and two of those orders describe the same shell mill from opposite sides of the same trade. Read on 2026-09-29, when 59 records carried the tag, the same check gave nine.
-
Fourteen front-running records, and four are front running
When this check was made on 2026-09-29 the library tagged fourteen enforcement records as front running. Reading the primary document behind each, only four describe someone trading ahead of an order they knew about: Motazedi, Gibson, Billimek and Williams, and Polevikov. The other ten were tagged from a word in a policy description, a guidance list or a footnote, from a different case, or from a release that charges misuse of customer order information without trading ahead. The first version of this piece counted seven.
-
Paid to make a market that was not there: the crypto market-maker cases
In October 2024 the SEC and the Department of Justice charged four crypto market makers with selling the same thing to many token issuers — trading with themselves to make a token look busy. Three of them were charged over offers to wash trade a token the FBI itself created. What the record shows about how common that was, and what it does not.
-
More than half of this library is not market manipulation
A count of what the library's 2,637 enforcement records are actually about, by regulator and by technique family — and why the answer changes what the headline number means.
-
Hack to trade: press releases stolen before they were published
In hack-to-trade cases nothing false is said to the market, so they are not manipulation. But the missing lie is why a court had to decide whether stealing information counts as deception under the securities laws, and the answer has limits.
-
Mismarking: when the price is a number someone types
Mismarking does not move any market price. It corrupts the value a fund, a bank or a lender sees. The library's sixteen records show two different kinds of case, and the best-known related matter, JPMorgan's London Whale trading, turns out not to be about marks at all.
-
Why the SEC fined eleven celebrities for a tweet
Floyd Mayweather, Kim Kardashian, Paul Pierce and eight others were not charged with fraud or manipulation. They were charged with a 1933 disclosure statute for not saying who paid them and how much — and the civil penalty on top of the disgorgement grew sharply between the 2018 cases and the 2022–23 ones.
-
The takeover bid that was never real
A takeover bid is the most price-sensitive thing a stranger can say about a company, and the SEC's records show it being counterfeited through EDGAR filings, wire press releases and, once, a New York Times advertisement.
-
How a stranger's brokerage account becomes someone else's pump
Four SEC matters, from 2016 to 2025, allege that hijacked retail brokerage accounts were used to buy stocks so the perpetrators could sell into the rise. The record supports a two-sided harm, and a more mixed story about lone actors turning into networks than the tidy version.
-
The anti-manipulation rule you can break without intending anything
SEC orders under Rule 105 of Regulation M say the rule applies irrespective of intent, so an order is not a finding that anyone meant to manipulate a price. This library holds 85 records under it, mostly from two 2013 and 2014 sweeps, and the record shows what the rule catches and what it does not.
-
How spoofing actually gets caught
The surveillance forensics behind order-book manipulation cases — which measures matter, which produce false positives, and why the strongest evidence is a conditional probability.
-
Anatomy of a spoofing case: what the enforcement record actually contains
A close reading of how spoofing prosecutions are actually built, using the enforcement records in this library rather than the narrative versions.
-
Is short selling market manipulation? The honest answer
Short selling is lawful, useful and constantly described as manipulation. Here is where the line actually falls, and what the enforcement record shows about how often it is crossed.
-
The Hunt brothers and the silver corner, explained properly
The most famous attempted corner in modern markets, and what it teaches about why corners are easy to build and nearly impossible to exit.
-
Pump and dump has moved to Telegram
The mechanic has not changed since the 1920s stock pools. The distribution channel has, and the change made the fraud cheaper to run and far easier to prove.
-
What LIBOR rigging teaches us about benchmark design
The benchmark scandals were not a failure of imagination. The vulnerability was described before it was exploited, and the fix was structural rather than punitive.
-
Spoofing enforcement since 2015, charted
What this library's own records show about how spoofing and layering enforcement has actually developed, and what the numbers cannot tell you.
-
Manipulation, insider trading and fraud: the distinctions that matter
Three things that get used interchangeably, are not the same, and attract different statutes, different burdens of proof and different victims.