Market Anomaly: Propping / Capping

• 2013–2014 – Mt. Gox (the “Willy/Markus” bots): Public Gox data shows day-to-day price support via recurring buy sequences with limited counter-volatility - consistent with propping. Status: academically analysed.
• 29 Aug 2024 – order-book depth / spot analyses: Large concentrated sell blocks capped short-lived breakouts (capping); after blocks were removed spread and depth normalized and VWAP converged back to mid. Status: completed analysis.
• 2017 (published 2018/2020) – stablecoin inflows & BTC spot: Time-targeted spot buys after pullbacks show statistical support effects (propping mechanism). Debate on causality continues.

Note: Public enforcement specifically for propping/capping in crypto spot markets is rare; microstructure signatures are nonetheless modellable (replenishment, OTR, half-life, VWAP drift).

1. Set the wall: Large visible limits placed close to the touch (bid = propping, ask = capping) stabilise or constrain price movement.
2. Partial fill → refill: After a small fill the order is immediately replenished (same or adjacent tick).
3. Queue dominance: Persistent re-quoting dominates the queue; aggressor flows bounce off the wall.
4. Differentiation: Not every wall is abusive (TWAP/POV, inventory management, block execution). Key factors are intent, persistence and cross-venue consistency.

• Replenishment signature: repeated refills of the same level within ≤ 300–600 ms after a partial fill; sequences persist for minutes.
• Best-quote half-life ↓: Best bid/ask reverts unusually quickly (P1–P5 vs a 30-day baseline).
• OTR spikes at top-5 depth: many updates per fill; quote-fade just before touch engagement.
• VWAP drift vs mid: propping: VWAP < mid; capping: VWAP > mid - despite small spread changes.
• Round-size / Benford: unusually round sizes (10k / 25k / 50k) and elevated Benford distance (L1) in the sequence.

• Order toolkit & fees: maker/taker models and VIP rebates make aggressive re-quoting cheaper - the “wall game” can be economically attractive.
• Limited visible depth: L2 feeds & hidden/iceberg orders make verification of remaining size difficult.
• Heterogeneous surveillance & 24/7 trading: no market-wide audit trail; global retail activity amplifies round-number / threshold effects.
• Venue specifics: differing tick/lot rules and queue policies allow queue-gaming at key ticks.

Rules & obligations: misleading quoting/reporting is sanctioned (e.g. FINRA Rule 5210). Audit trail: the US CAT (SEC Rule 613) captures order life-cycles - sequences (quote→trade→cancel) are analysable and evidential. EU-MAR: forbids false or misleading signals - detection and enforcement risk is higher.

• MiCA timeline: in force since 29 June 2023; parts effective from 30 June 2024 (ART/EMT) and CASP obligations from 30 Dec 2024. Transitional arrangements until 01 July 2026 may apply per member state.
• Market integrity (IOSCO): “Same activity, same risk, same outcome” - focus on spot market integrity, monitoring and enforcement.
• Practical benefit: early warning reduces slippage, supports best-execution evidence and builds audit-ready dossiers (E-level flags, UTC timestamps, venues, metrics).

Calibration: per symbol × venue; baseline = 30 days at the same time of day (UTC); hourly bins; tolerance ±15 minutes. Thresholds expressed as percentiles vs baseline.

Sample metrics: OTR, aggressor quote, round-size ratio, Benford distance (L1), spread jump (bps), VWAP drift vs mid (bps), best-quote half-life (ms), replenishment latency (ms), lead/lag (1-s bins, ≥ 3 venues).

E3 (high) - “persistent wall + refill engine”

• Replenishment latency ≤ P1 and best-quote half-life ≤ P1 across ≥ 4 cycles within ≤ 3 min,
• OTR ≥ P99.5 at top-5 depth and round-size ratio ≥ 2× baseline,
• VWAP drift ≥ P99 while spread jump ≤ P50 (damped spread despite pressure),
• Cross-venue consensus: ≥ 3 core venues trigger E2–E3 within ≤ 3 s.

E2 (medium)

• Replenishment latency ≤ P5 on ≥ 2 consecutive levels, or
• best-quote half-life ≤ P5 or OTR ≥ P99,
• Benford distance (L1) ≥ P95 or round-size ratio ≥ 1.5× baseline.

E1 (low)

• Single wall sequence (≥ 2 replenishments in ≤ 2 min) or OTR ≥ P97,
• VWAP drift ≥ P90 (propping negative / capping positive) lasting ≥ 60 s.

• Legitimate alternatives: TWAP/POV resting, inventory management or block executions can look similar; true blocks typically do not produce persistent refill patterns.
• Event filters (UTC): exclude listings/re-listings, maintenance windows, large on-chain transfers, MM withdrawals; exclude macro slots (e.g. CPI / FOMC).
• Cross-venue check: consider E3 only when patterns are consistent across ≥ 3 large venues (single-venue artefacts often false positives).
• Data quality: discard windows with L2 gaps/errors ≥ 1%; compute in bps / ms and store UTC timestamps and notionals.

• Execution: recognise persistent walls - don’t run into them; reduce slippage and optimise limit placement.
• Alpha & risk: exploit reversion after wall removal; avoid unnecessary market impact.
• Governance & supervision: E-level logs with sequence evidence (replenishment, half-life, VWAP drift, OTR) are audit-grade - support best-execution and surveillance under MiCA/MAR.

• Gandal, Hamrick, Moore, Oberman – Price Manipulation in the Bitcoin Ecosystem (2018).
• Griffin & Shams – Is Bitcoin Really Un-Tethered? (Journal of Finance, Aug 2020; working paper 2018).
• Kaiko – Moving markets: liquidity and large sell orders (29 Aug 2024).
• FINRA – Rule 5210 Publication of Transactions and Quotations (ongoing).
• SEC – Rule 613 Consolidated Audit Trail (CAT) (ongoing).
• ESMA – Statement on MiCA Transitional Measures (17 Dec 2024).
• IOSCO – Policy Recommendations for Crypto- and Digital-Asset Markets (16 Nov 2023).