🚨
$OPEN — POSSIBLE ALGORITHMIC MARKET MANIPULATION.
@nejatian — I am addressing this directly to you as CEO of Opendoor.
After weeks of forensic work on
$OPEN trading, I believe our findings justify review by Opendoor, Nasdaq, FINRA and potentially the SEC.
As an
$OPEN shareholder, I can provide the complete dataset, order IDs, timestamps, code and methodology.
We reconstructed ~2.23 MILLION individual orders across Nasdaq, NYSE Arca, Cboe BZX, EDGX and MEMX, covering ~86% of observed volume: MBO/L3 orders, executions, cancellations, off-exchange trades, auctions, options and short-sale data.
Thousands of quantitative tests: regressions, event studies, cross-venue synchronization, market-impact models, VPIN and order-lifecycle reconstruction.
The findings raise serious questions about potential algorithmic market manipulation.
From Sep 9–17 alone, 14:00–15:00 ET produced −4.94M shares of aggressive net flow, negative in 6/7 sessions.
But the key evidence is inside the book.
📍 SEP 9 — 11,200
An 11,200-share SELL executes @ $3.05.
Then 335 SELL orders of exactly 11,200 shares appear.
326/335 — 97.3% — disappear within ONE SECOND.
After
$OPEN trades lower:
🟢 BUY 11,200 @ $3.00 → EXECUTED.
SELL → REPEATED SUPPLY → CANCEL → LOWER PRICE → BUY
If CAT links these actions to the same beneficial owner/coordinated accounts, regulators can determine whether displayed supply influenced price discovery before a lower execution.
📍 SEP 17 — THE $2.57 WALL
172 SELL orders of exactly 14,500 shares.
167/172 — 97.1% — cancelled within one second.
Around 108 identical 14,500-share SELL orders repeatedly appear around $2.57 while genuine selling hits
$OPEN.
Median lifetime:
~78 MILLISECONDS.
Same size. Same price. Appear. Disappear. Repeat.
Genuine liquidity — or the appearance of a larger sell wall than actually intended for execution?
📍 SEP 24 — THE CLOSE
Nasdaq SELL imbalance jumps:
354K → 970K → 1.17M shares in ~20 seconds.
At virtually the same time, selling appears across ALL FIVE exchanges.
Around 15:54:40 ET, in ~ONE SECOND:
🔴 ~134,060 aggressively SOLD
🟢 ~810 aggressively BOUGHT
~165 : 1
Nasdaq. Arca. EDGX. BZX. MEMX.
The five venues finish at approximately −822K shares of identified aggressive net flow.
⸻
THESE ARE ONLY EXAMPLES.
Across the five-exchange sample:
🔴 38,076 large fleeting SELL orders
🔴 ~353M shares of cumulatively displayed sell liquidity
A stricter test — 5,000+ share orders appearing before an opposite-side execution and cancelled within 50ms — identified:
68,524 SELL-side candidate sequences
vs 49,769 BUY-side
= 18,755 more SELL occurrences.
We also tested thresholds down to 5 milliseconds.
These totals alone don’t prove manipulation; high-frequency activity exists on both sides. But the specific fingerprints come from tens of thousands of order/cancellation sequences and deserve account-level investigation.
WE DOCUMENTED:
• persistent aggressive selling
• identical non-round SELL orders
• 97%+ cancellation rates in specific fingerprints
• lifetimes measured in milliseconds
• apparent sell walls
• auction imbalances
• simultaneous cross-exchange selling
These patterns overlap with spoofing/layering when non-bona-fide liquidity creates misleading supply/demand while genuine transactions obtain an economic advantage.
The decisive question: WHO controlled these orders?
Public data cannot answer it. The Consolidated Audit Trail (CAT) can.
@nejatian , I hope you and
@Opendoor take this seriously. I can provide the entire analysis and dataset immediately.
If Opendoor does not act, on Monday, as an
@Opendoor shareholder, I intend through legal counsel to submit the prepared documentation directly to the SEC and appropriate regulators for independent review and to protect Opendoor shareholders’ interests.
The data, order IDs, timestamps and anomalies exist. Now they deserve to be investigated.
@rabois @ericjackson @mikealfred