Market Anomaly: Marking the Close (Fix/Settlement)

• Deribit, 31 Oct 2019 (index/settlement window): An error in the index calculation led to an abrupt price drop, stops/liquidations; reimbursement of ~150 BTC (~USD 1.3m) and a public explanation. Demonstrates sensitivity of index/settlement windows.
• OKEx BCH-Futures, 11/2018 (early settlement): After violent moves the contract was settled at the last traded price; accusations of manipulation followed and the exchange responded. Regardless of outcome, the case highlights the criticality of the settlement point.

Context: Crypto spot markets have no traditional exchange close. "Marking the Close" appears here as "Marking the Fix/Settlement" - influencing reference, mark or settlement prices around expiry, fixing or snapshot windows.

Short & understandable: Moving reference/settlement prices inside narrow time windows (e.g. 30-min TWAP, closing print) to influence PnL, payoffs, funding or valuations.

How this works in practice on CEXs

1. Time windows: Defined observation windows (e.g. 30-min TWAP, 4s snapshots) determine settlement price; mark price is based on an index plus basis adjustment.
2. Leverage points: Moves close to the fix/settlement in spot sources or the spot-perp basis shape TWAP/final-print – especially with thin books or few constituents.
3. Tactics: Concentrated buys/sells just before/within the window, often in waves aligned to snapshot times.
4. Outcome: PnL advantages (e.g. options settling ITM), more favourable funding/settlement terms, smoother reporting values at cutoffs.

• Price spikes in index sources only during the settlement/TWAP window; multiple micro-bursts at snapshot times; reversal shortly after window end.
• Spot-perp basis jump in the final minutes before expiry/funding snapshot, followed by rapid normalization.
• Asymmetric impact: small execution sizes cause disproportionate TWAP shifts while the intraday profile is otherwise quiet.
• Concentration on few index constituents (high methodological weight).

• Centralised index/mark prices with few sources → less outlier filtering; single outliers have stronger effect.
• 24/7 trading + narrow windows → clearly schedulable leverage points.
• No CAT-equivalent (cross-venue audit trail) → reconstructing intent/actors is harder.

• Clear definitions & surveillance: "Marking the close" is explicitly addressed; prohibited price-securing behaviour is monitored. Surveillance compares close windows vs intraday baseline.
• Closing auctions & time-sync: Auction mechanisms, MiFID II/RTS 25 UTC precision and cross-market surveillance complicate window tactics and increase evidentiary quality.

• Market quality & fairness: Settlement deviations impact PnL, liquidations and margin directly – small distortions can have large consequences.
• MiCA (EU) – transition: Largely applicable since 30 Dec 2024; transitional arrangements until 01 Jul 2026 possible. Expect data-driven monitoring of fix/settlement windows, outlier filters and STOR processes.

Adaptive, quantile-based, per symbol/venue against a 30-day baseline for the same time-of-day. Required: L1/L2 of index sources, perp/spot basis, TWAP snapshot schedule, trades.

E3 (high) – "Fix/Settlement clearly shifted"

• ΔTWAP ≥ ±0.75% in the official observation window, and
• ≥ 60% of the TWAP displacement originates from ≤ 2 index sources (source concentration), and
• Spot-perp-basis jump ≥ 0.50% in-window with reversal ≥ 50% of the move ≤ 15 min after window end, and
• Burst clustering at snapshot times (e.g. every 4s) demonstrable.

E2 (medium)

• ΔTWAP 0.30–0.75% and asymmetry (≥ 40% notional from 1–2 sources) and short reversal after the window.

E1 (low)

• Repeated micro-bursts on 2–3 snapshots with small basis spikes → watchlist until concentration/reversal is confirmed.

• Separate genuine roll/expiry flows (hedging, desk rolls) - legitimate window activity exists.
• Know index methodology & outlier filters (e.g. 30-min TWAP, 4s sampling, multi-source index) - only flag patterns that exceed expected robustness.
• Cross-venue check: a signal on only 1–2 sources is more suspicious than a broad market move.
• Liquidity & weighting: monitor high-weight constituents closely; document governance weights.

• For traders: Recognise fix/settlement signatures, avoid incorrect settlement levels, reduce slippage/liquidation risk.
• For operators/compliance: Well-documented window anomalies (timestamps, source concentration, snapshot bursts, reversal proof) support reviews, client communication and STOR filings (MiCA/MAR).

• Deribit – "Deribit Flash Crash: A Letter from the Founders", 1 Nov 2019; reports on ~USD 1.3m reimbursement; founder interview on index error cause.
• OKEx BCH-Futures (11/2018) – industry coverage on settlement and accusations; exchange statement.
• DFSA Rulebook – "CMC 2-2 Market Manipulation" (examples incl. marking the open/close).
• ESMA / MAR implementing acts – indicators for "false or misleading signals" / "price securing".
• IOSCO – "Investigating and Prosecuting Market Manipulation" (marking-the-close as a classic pattern).
• Deribit Docs – "Settlement" & "Index Prices": 30-min TWAP, 4-sec snapshots, multi-source index.
• Trading Technologies – "Marking the Close: Surveillance Model" (working definition & patterns).
• Cube.Exchange – "What is Mark Price?" (mark/index price for perps/liquidations).
• CME – "Crypto Futures FAQs" (formal close/reference definitions; best-practice benchmark).
• CoinDesk Research – "TWAP Settlement Rate Methodology" (2025) (general description of TWAP fixings).