Case File 12
The $5,500 Kalshi Pattern Is Real. Here’s What It Actually Shows.
Roughly half of the ETH-perpetual notional in four highlighted windows came from executions clustered near $5,500. The deeper pattern was fixed-dollar sizing—not literally identical trades—and the public tape does not prove wash trading.
This investigation began with Beni’s public post highlighting the repeated executions ↗.
The concentration is real and persistent. The evidence supports a highly standardized, likely automated execution pattern. It does not identify its operator or establish that the same beneficial owner controlled both sides.
Contents
What we found
MSR Decode reconstructed the four chart windows highlighted by Beni. In each window, executions within $5 of $5,500 represented roughly half of ETH-perpetual notional.
| Highlighted window | Share near $5,500 |
|---|---|
| September 16 | 47.0% |
| September 18 | 58.4% |
| September 19–20 | 51.3% |
| September 20 | 47.6% |
These are overlapping chart windows, not four separate calendar days. The first window also depends on a sub-minute boundary inside the opening minute shown on the chart. That boundary reconciles the chart’s displayed total, but the original chart does not disclose its exact inclusion rule.
The broader record is stronger. Across 35 complete UTC days from August 17 through September 20, the $5,500 band accounted for $5.25 billion of $11.37 billion in measured notional. On September 9 alone, its share reached 68.7%.
They were not literally identical $5,500 trades
The contract quantities changed. The resulting dollar size barely did.
When ETH’s quoted price moved higher, the number of contracts generally became smaller. When the price moved lower, the number of contracts became larger. The executions clustered near the same dollar target even though their quantities were different.
The simplest mechanical description
Fixed-dollar sizing around a target of approximately $5,500.
That is consistent with automated execution or market-making logic. It does not show whether one system generated every trade, whether one participant or several participants were involved, or whether the same beneficial owner stood on both sides.
The target itself changed
Standardized sizing appeared before $5,500 became dominant. The apparent target moved over time:
The interesting phenomenon is therefore not simply that someone kept trading exactly $5,500. It looks more like a standardized execution process whose dollar target changed. The public tape does not reveal why.
What else we tested
We looked for public signatures that might turn repetition into stronger evidence of self-trading. None was decisive.
- Trade direction: the $5,500-class notional was almost evenly split between public
askandbidtaker labels—50.14% versus 49.86%. Those labels show which side took liquidity, not who owned either account. - Open interest: it moved more during periods of heavy $5,500-class activity, but not consistently in one direction. The flow did not simply create positions or simply close them.
- Timing: median gaps in the highlighted windows ranged from roughly 0.7 to 2.2 seconds. We did not find one exact repeating clock interval.
- External price movement: Kalshi and the broader ETH market moved closely together, but the link between clustered volume and external price changes was modest and did not establish cause or direction.
These tests support a description of highly repetitive automated activity. They do not establish self-trading.
What about Kalshi’s rebate programs?
Kalshi’s August regulatory filing ↗ described conditional exchange-level economics under which eligible makers could receive approximately 0.3 basis points while eligible takers paid approximately 0.3 basis points.
Those terms could make very large gross turnover inexpensive at the exchange-fee level. But the public evidence reviewed here does not prove that the program was activated for these executions, that the participants behind the clustered flow were eligible, or that one participant captured economics on both sides.
The timeline also resists a simple explanation: standardized sizing was visible before the specific $5,500 target emerged.
What the public data can—and cannot—show
Kalshi’s public tape provides prices, quantities, timestamps and taker-side labels. It does not provide the participant and beneficial-ownership information needed to connect both sides of a trade.
The same visible pattern could arise from automated market making, execution algorithms, inventory management, hedging, arbitrage, several participants following similar sizing rules—or abusive activity. The public record alone cannot distinguish among them.
Establishing wash trading would require evidence not present in the public tape: account linkage, beneficial-ownership information, exchange surveillance records or other records connecting both sides of the transactions.
Bottom line
Beni found something real. Around half of the ETH-perpetual notional in the highlighted windows was concentrated near $5,500, and the wider reconstruction shows that the phenomenon persisted far beyond those screenshots.
But these were not literally identical trades. They were a highly concentrated fixed-dollar sizing pattern, with quantities changing as prices moved and the apparent target shifting from roughly $4,000 to $4,500 and then $5,500.
That is unusual enough to deserve an explanation. It is not, on the public evidence currently available, proof of wash trading.
What process generated $5.25 billion of ETH-perpetual turnover—46.2% of the 35-day sample—inside one narrow dollar-sizing band, and why did its target change?
The public tape cannot answer that. Kalshi can.
Figures are rounded for readability. “Notional” in this report means execution-price-weighted trading activity. This report does not identify any trader, beneficial owner or account.
Questions for Kalshi
- What type of participant or execution process produced the concentrated sizing pattern?
- Why did the apparent target move from roughly $4,000 to $4,500 and then $5,500?
- Did Kalshi’s surveillance systems review the activity for common ownership, self-matching or pre-arranged trading?
- Which fee or rebate terms actually applied to these executions, and when were those terms activated?
- Can Kalshi confirm whether the pattern came from one participant, several participants or exchange-supported liquidity programs?