EXANTE Algorithmic Trading: Practical Aspects

Click here to load reader

Embed Size (px)

description

Slides for speech of EXANTE Managing Partners Vladimir Maslyakov and Anatoliy Knyaze , entitled "Practical aspects of algorithmic trading and high-frequency trading", on TradeTech Russia 2011 Presentation highlights the problems associated with the development of a model (pre-trade analysis), the launch of the strategy (trading) and the post-trade analysis, as well as an overview of the algorithmic trading in general, and a small glimpse into the future.

Transcript of EXANTE Algorithmic Trading: Practical Aspects

  • 1. Algorithmic trading: practical aspectsEXANTE Ltd. exante.com.mtinfo@exante.com.mtMoscow 2011

2. I. Algorithmic tradingII. Develop the modelIII. Launch the strategyIV. Analyze the resultsV. Trends 3. Algorithmic tradingAutomated tradingHFT 4. Arbitrage Pricing Automated trading Buy-sideSell-sideTrendfollowing Smart order routingStat arbitrage Market Making / HFT VWAP 5. I. Algorithmic tradingII. Develop the modelIII. Launch the strategyIV. Analyze the resultsV. Trends 6. Data Hypothesis Model Testing 7. Historical Data WidthDepthCorrectnessInstrumentsPast periodSplits and divsVenuesResolutionGapsCorp. actionsOrder bookTimestamps News CounterpartiesValidation 8. Data Hypothesis Model Testing 9. A priori knowledgeFundamental EmpiricalGut feeling 10. Visualization DatavolumeMathSpeed panopticon.com 11. RTS Index and S&P Index, 2010-10-11RTSISPX16:40 16:5017:00 17:1017:20 17:3017:40 12. Data Hypothesis Model Testing 13. ModelAlpha Risks Transaction costs 14. EDGE ? 15. Model: mathPrototype 16. Our experience: RDomainLibrariesOpen and free SlowNo realtime Open and free 17. Data Hypothesis Model Testing 18. TestingDataPrototype Results Historical data R / Python/ Alpha ModelingJava Risksmarket impact Cluster / cloud Transactionand order flow GPU costs Realtime 19. I. Algorithmic tradingII. Develop the modelIII. Launch the strategyIV. Analyze the resultsV. Trends 20. 21. Realtime dataSpeed Depth CoverageLow-latency L1AmericasUltra low-latency L2 EuropeSub millisecondRawAsia 22. Strategy LanguageInfrastructure ControlHigh-level (C++, Java,Client Manual C#, etc)DSL (Slang, etc) Server AutomaticVisual (diagrams) CloudGUI Strategy sandboxDataOrdersNYSEMFGRobot 1Robot 2Robot 3Robot 4 LSEJP 23. Arbitrage exampleLondon Server (Telehouse)Arbitrage strategyGAZPRUOn new tick:LIMIT (LSE)(MICEX)ogzd_rub = convert(ogzd, usd_rub)spread = normalize(ogzd_rub/gazpru) Filled (size)changedSpread() OGZD (LSE) On change spread: if (spread > threshold)MARKET (MICEX) place_limit(OGZD, price, size)USD/RUBFilled (price)(FOREX)On limit fill: If (limit_is_filled) place_market(GAZPRU, size)Parameters: threshold 24. VWAP example Moscow Server (MacomNet)VWAP strategySBER bid/askOn new tick:(MICEX)vwap = recalculateVwap(trades)execute_vol =recalculate(average_volume, volume) MARKET (MICEX)SBER volume executeOrder(execute_vol)(MICEX)Filled (size) on market fill:SBER trades(MICEX)our_vwap = update(price, size) vwap_delta = our_vwap - vwapParameters: average_volume 25. I. Algorithmic tradingII. Develop the modelIII. Launch the strategyIV. Analyze the resultsV. Trends 26. Gather results dataMarket snapshotOrdersDataLatencyStrategy parameters 27. Export the results data Excel R, MatlabExportVisualization Model 28. Compare with the model 29. Optimize the parameters Model Results Testing Trading 30. I. Algorithmic tradingII. Develop the modelIII. Launch the strategyIV. Analyze the resultsV. Trends 31. Adoption100 FORTSCME GLOBEX Vol, % Msgs, %90 E-mini S&P 500 Futures 51.66 69.9390 EuroFX Futures 69.32 83.4180 Eurodollar Futures 51.29 64.4670 60Crude Oil Futures35.34 71.2460Algorithmic Trading and Market Dynamics July 15, 20105040 Foreign Exchange Buy-side, %Sell-side, %30 Order Routing25 9220 Time-slice 25 1510 Liquidity42 46 0 Alpha92 39Vol, %Msg, % FX Hedging 25 39Estimated by FORTS 09.2011 Streambase 2011 Special Report on FX 32. Dodd-Frank Swap Execution Facility.SEC 15c3-5 Eliminate naked access to exchange.MiFID II Crossing networks, derivatives, HFT. 33. Algotrader: a new breed MathematicsTechnology Finance 34. StrategiesMulti-assetULL DMAHFT tradingBuy-side or sell-FX, Eqty, Debt,5s / km side? DerivsMarket making Europe, USA, Asia< 100s / algoLiquidity search Feeds andand aggregationexecution106 msg / sec Fast and reliableStat arbdata 35. TechnologiesSoftwareOverclocking FPGAMulti-coreGPGPUCloud x32x200x30000 36. Our experience: cloudPowerDiversity Cost controlEngineering Compliance Latency 37. Service! 38. Anatoliy Knyazev ak@exante.com.mtVladimir Maslyakov vm@exante.com.mt