Skip to Academy content

Quantitative Trading · Level III · Professional

Microstructure, HFT and Production

Market microstructure, the limit order book, market making and high-frequency trading in full detail, then trading system architecture, production pipelines, real-time risk controls, live monitoring, capacity and the failures that define the field. Ends with the autonomous strategy project.

“How do professional trading operations really run?”

Outcomes

What you will be able to do.

  • Explain how prices form and spreads arise from the main microstructure models.
  • Model the limit order book and build market-making strategies.
  • Describe HFT strategies and the infrastructure that runs them.
  • Design a production trading system and its pipelines.
  • Control risk in real time and decide when to switch a strategy off.
  • Run an autonomous strategy and report its results honestly.

Modules

12 modules, plus a free orientation.

  1. 01

    Kyle, Glosten-Milgrom and Roll, inventory models, adverse selection, price discovery and order-flow toxicity.

    The Roll spreadGlosten-MilgromKyle's lambdaInventory modelsPrice discoveryOrder-flow toxicity
  2. 02

    Matching engines, price-time and pro-rata priority, queue dynamics, order flow imbalance, the microprice and order book data.

    Matching enginesPriority rulesQueue dynamicsOrder flow imbalanceThe micropriceOrder book data
  3. 03

    Avellaneda-Stoikov, inventory risk, quoting, adverse selection and rebates.

    The market maker's problemAvellaneda-StoikovInventory controlAdverse selectionFees and rebatesMarket making in practice
  4. 04

    Passive market making, latency arbitrage, cross-venue and ETF arbitrage, news latency and order anticipation, and their economics.

    What HFT isLatency arbitrageCross-venue and ETF arbitrageStatistical HFTNews and event latencyThe economics of speed
  5. 05

    Colocation, microwave links, kernel bypass, FPGAs, clock synchronization, tick-to-trade latency and low-latency software design.

    ColocationMicrowave and laser linksKernel bypass networkingFPGAsClock synchronizationLow-latency software
  6. 06

    Queue position models, latency simulation, fill probability and market replay, and why HFT backtests lie.

    Why HFT backtests lieMarket replayQueue position modelsLatency simulationFill probabilityValidating a simulator
  7. 07

    Market data handlers, order and execution management, the FIX protocol, event-driven design, state and recovery.

    System componentsMarket data handlersOMS and EMSThe FIX protocolEvent-driven designState and recovery
  8. 08

    Ingestion, scheduling, retraining, deployment, monitoring, alerting, reconciliation and disaster recovery.

    Data ingestionSchedulingRetraining and deploymentMonitoring and alertingReconciliationDisaster recovery
  9. 09

    Pre-trade checks, kill switches, position and loss limits, SEC Rule 15c3-5, MiFID II RTS 6 and market abuse.

    Pre-trade checksKill switchesLimits and throttlesSEC Rule 15c3-5MiFID II RTS 6Market abuse
  10. 10

    Live versus backtest divergence, decay detection, slippage attribution and when to switch a strategy off.

    Live versus backtestSlippage attributionDecay detectionRegime changeSwitching a strategy offThe strategy lifecycle
  11. 11

    Alpha decay, capacity estimation, crowding, the August 2007 quant quake and multi-strategy platforms.

    Alpha decayEstimating capacityCrowdingThe 2007 quant quakeMulti-strategy platformsScaling a book
  12. 12

    Knight Capital, the 2010 Flash Crash, spoofing, LTCM, Amaranth, Volmageddon and Archegos: what failed and which control would have stopped it.

    Knight CapitalThe Flash CrashSpoofingLTCMAmaranthVolmageddon and Archegos

Practice and assessment

Prepare, practise, get assessed.

Final project

An autonomous trading strategy: walk-forward stability, autonomous operation, a reconciled ledger.

Electives

Optional courses for this level.

Python for Trading III

Production Python: services, scheduling, databases for tick data, monitoring and testing.

Services and schedulingTick data storageMonitoringTesting and CIPerformance profilingDeployment

Available

Low-Latency Programming

The engineering behind high-frequency trading, in C++ and Rust.

Memory and cachesLock-free queuesNetworking and kernel bypassMeasuring latencyC++ for tradingRust for trading

Available