2024 11 20 Дмитрий Залялетдинов МК

High-level summary

• Presenter: Dmitriy (Dima) Zalyaletdinov.
• Topic: a practical trading playbook for detecting and trading against algorithmic “robots” in illiquid Russian-equity markets using three volume-related indicators in the Vatagi terminal:
  • raw volume,
  • BIS (bid/ask separated prints — “second bid”),
  • communicative Delta.
• Emphasis: detect recurring robot prints, estimate their timing and volume, size entries, manage risk, and exploit micro-inefficiencies. Several live case studies (Bashneft, DME, Tower, etc.) illustrate rules and failure modes.

Frameworks, playbooks, and processes

Robot-detection and execution playbook (step-by-step)

1. Monitor BIS to see per-side prints and repeating timing patterns.
2. Identify a recurring print interval (typical robots ~12–13 seconds) and use that “timer” for predictability.
3. Confirm the pattern with 3–6 sequential prints to rule out imitators or coincidence.
4. Submit successive orders:
   • First order to get exposure.
   • Follow-ups sized by how much you can “swallow” (liquidity tolerance).
5. Reduce exposure after the very first impulse (e.g., keep half) to limit tail risk.
6. Use communicative Delta to compute net long/short accumulation during the robot run and to estimate squeeze risk.
7. After the robot peak, wait for the impulse to pass before covering or re-entering; prefer limit orders for opportunistic fills.

Sizing and pyramid method (risk-managed scaling)

• Start small: 2–4 lots (open 2 lots, take profit on one, hold the other).
• If prints continue in your favor, add increments (e.g., +2–4 lots) as the average price becomes favorable.
• Control the number of pyramids and total exposure based on historical drawdowns and liquidity tolerance.
• Reduce exposure after initial impulses to protect against sudden takeouts.

Using multiple indicators together

• Raw volume: detect relative volume/liquidity.
• BIS: identify side-specific timing and robot-print signatures.
• Communicative Delta: estimate cumulative net position accumulation during impulses/breakouts.

Anti-impostor test

• Check second-level timing consistency: true robots are highly regular; manual imitators show inconsistent intervals.
• Cross-check BIS timing with the order book (glass) and across multiple instruments/correlators.
• Require repeated identical lot sizes and precise timing before assuming a robot.

Key metrics, KPIs, and measurable signals

• Typical robot print interval: ~12–13 seconds.
• Communicative Delta example: 32,000 over 117 seconds (used to estimate net lots on one side).
• Common single-print lot sizes observed as signatures: 500 lots, 600 lots, 25 lots, 20 lots, 5 lots, 2–4 lots (trader entry examples).
• Robot splitting behavior: e.g., robot buys half of an incoming order (if you place 10 lots, robot buys 5).
• Practical cautions: avoid pouring >500 lots into instruments like Tower.
• Example breakout measurement: 6,915 lots identified as a buyback/shoot window.
• Operational KPI: test strategy performance net of commission — small gains can be wiped out by fees.

Concrete examples and case studies

• Bashneft
  • Robot prints visible only on Bashneft; turned on at 78 every 12s, then stopped later.
  • Lesson: single-instrument signals can be false positives; always cross-check.
• Robot that splits orders
  • Observed deterministic behavior (e.g., buys exactly half of an incoming order) — exploitable once identified.
• DME and other ticks
  • Instances of two robots working in opposite directions, allowing simple open-and-close profit cycles.
• Large short-throw example
  • Teammates observed a ~500-lot short throw followed by stealth buy-off from a robot — an inefficiency that produced straightforward gains.
• Communicative Delta use
  • Used to quantify position build-up during robot activation and to select safe entry points on rollbacks.

Actionable recommendations and tactics

• Always display these indicators on your workspace:
  • Raw volume, BIS, and communicative Delta.
• Corroborate BIS signals with the order book (glass): watch for repeated densities and lot sizes.
• Entry sizing:
  • Use small starter lots (2–4) and pyramid only if prints continue in your favor.
  • Size entries by how much you can swallow and by observed robot strength.
• Risk controls:
  • After the first impulse, keep at most half exposure to limit takeout risk.
  • Be prepared to exit quickly if you cannot liquidate in a panic — better to exit early.
  • Always include commission in backtests; commission can erase micro-profits.
• Use cross-instrument correlation:
  • Monitor indices, currencies, and heavyweight papers (Sber, Gazprom) as correlators and drivers.
  • Divergence among correlators (e.g., during special events) can create exploitable inefficiencies.
• Screening vs eyeballing:
  • Screeners may miss seconds/milliseconds-level robot behavior; live quote scanning + BIS visual confirmation is preferred.
• Distinguishing real robots:
  • Look for consistent per-second timing and repeated precise lot sizes.
  • Check multiple instruments and order-book patterns for confirmation.

Operational and organizational notes

• Tools and terminal:
  • Vatagi terminal highlighted for BIS and communicative Delta views.
• Workspace setup:
  • Multi-monitor layout with BIS, minute charts, and order book visible; real-time scanning of quotes and correlators.
• Team workflow:
  • Collaborative detection (team members and interns share screenshots and patterns).
  • Maintain a trading diary and analyze historical liquidity and robot behavior.
• Practical warnings:
  • Illiquid papers often have transparent books (participant names and exact sizes visible) — can be exploited but increases tail risk.
  • As more traders adopt robot-detection, edge compression and crowding risks grow.

Limitations, failure modes, and cautions

• False positives: single-instrument prints can be imitators or lag artifacts — always verify timing and cross-instrument signals.
• Liquidity and panic risk: simultaneous exits can overwhelm robots/market-makers and produce large adverse moves.
• Commission drag: frequent small trades accumulate fees that can nullify gains.
• Market evolution: widespread adoption of the same signals reduces future effectiveness — be conservative when scaling.

Presenters and sources

• Main presenter: Dmitriy (Dima) Zalyaletdinov.
• Contributor referenced: Albert (intern).
• Tool: Vatagi terminal.
• Markets/instruments referenced: Moscow Exchange and multiple Russian tickers (Bashneft, DME, Tower, Sber, Gazprom, Beluga, Segezha).