Where I was at with HFT in 2011.. now there are Vampire Systems that feed on HFT.
The development of the ARM (Arbitrage Risk Module) project started in Nov of 2010. It’s a propriety component built on top of the XXX existing architecture. The following software components had to be developed to allow automated trading of ARM strategy:
1.Direct Liquidity Integration
Currently the Aggregation Module is connected to 3 destinations for Forex liquidity. These include HotspotFX, PFGForex (Institutional) and Interactive Broker. The ARM projects success is dependent on direct access to as many liquidity hubs and direct bank feeds as possible. We plan on adding 12 more liquidity solutions to the Aggregation system, this is an ongoing tasks that requires development and maintained.
2.Liquidity Aggregation Module
All the data from multiple destinations are aggregated and sorted in realtime without adding any latency to the streams. The aggregated streams are then passed on for analysis.
3.Trading Model
The trading analysis modules seeks out opportunities based on various models that have been developed. These models will determine the risk, velocity and available liquidity in order to generate signals.
4.Smart Order Routing
Once a trading opportunity is detected, order are routed to different destinations based on preset of rules which seek best execution using a sophisticated set of algorithms.
5.Risk module
The risk module tracks and monitors margins, fills, latency, and the overall “health” of each liquidity provider. Automated triggers will stop trading automatically based on several risk criteria that have been identified. Additional risk measure will be added as the project develops further and more realtime trading data is available.
6.Order Management System
The order management system is visual monitoring tool that allows full control over every aspect of the trading process including manual overrides, alerts, connectivity monitors, interaction with orders and more.
Risks involved:
Essentially the biggest risk to the project is related to latency:
·
Latency in communication systems
Co-location and deployment of fast servers will reduce latency to its absolute minimum.
·
Latency in software technology
ARM Module latency has been reduce to its lowest possible limits (2.4 Millisecond turnaround)
·
Speed of order execution by each destination
This will be the biggest ongoing learning curve and risk. The team has to continuously adapt to changes to liquidity providers execution profile and aggressively add new liquidity destinations to the aggregation system.
·
Issues arising from any technological breakdowns
Contingency measure will be implement to factor for each of the known risk scenarios.
Backtesting results:
The software components have been rigorously tested under stress conditions, and the models have been validated mathematically.
Liquidity streams of data from various destinations and their respective signatures, changes drastically depending on the sources of liquidity and the nature of the relationship with each bank (liquidity providers). This change in data profile and the nature of the trading model essentially invalidates any backtesting results. Therefore the only way to test the validity of this model is to forward test the strategies.
Simulated forward testing results:
Simulated trading tested have been conducted to tested out the technology components of this project. Please watch the following movie:
Notes:
Change the resolution in Youtube to
HD 1080p for the best quality viewing. (The button is towards the left side of the video player near the bottom.)
This is private link for private viewing only. Please do not distribute.
This is a low quality video and many frames are missing, therefore the data flow appears slower then normal.
Notice:
·The aggregated time and sales for 2 destinations, HOTSPOTFX and PFG.
·The symbol being tracked is EURUSD.
·The P&L is being tracked.
·Notice the trades are being executed in realtime.
