Disorderly Market Testing - software and testing as a service.
Generate New Revenue Streams, Secure Competitive Advantage, meet New Testing Regulations
MiFID II and MiFIR Level 1 were published in the Official Journal of the European Union on 12 June 2014. ESMA ran Level 2 consultation processes in both 2014 and 2015 with the final draft of new Regulatory Technical Standards published on 28 September 2015. These new regulations have added substantially to the current Level 3 regulations laid down by the April 2012 ESMA Guidelines on systems and controls in an automated trading environment for trading platforms, investment firms and competent authorities. This has changed the landscape irrevocably for testing of execution systems and trading algorithms placing new requirements on Trading Venues, their Members and Market Participants.
Trading venues have the responsibility to offer isolated pre-live testing facilities and to require their members to certify that the algorithms they use to trade have passed disorderly market testing and stress testing for each venue in addition to enhanced conformance tests before admitting them to live trading. Investment firms (both buy and sell-side) have the duty to stresstest their trading algorithms for propensity to create or contribute to disorderly trading Investment firms can choose to use the trading venues test systems, develop their own disorderly trading test systems, or use a tbird party service. This offers trading venues a new income stream and additional competitive advantage to early movers by helping members and participants meet their new testing responsibilities.
AlgoGuard Arena by TraderServe offers the ideal solution connecting transparently to a trading venue's own test matching engine and generating dynamic responsive orderbooks statistically representing each instrument and isolated from the live market. Members interact seamlessly by sending orders to and receiving updates from the Trading Venue test system. Test schedules with antagonist algorithms provide the necessary disorderly market and stress testing capabilities with pass/fail drill down reporting and logging.
Both Trading & Execution algos can misbehave. When they do so they can seriously affect the proper functioning of an electronic market – a matter of increasing concern to financial regulators. No matter how carefully an investment firm has tested an algo, there is always a threat that it will go off the rails once committed to live trading, particularly if it meets extreme market conditions not experienced previously in simulated trading. How it will behave in the company of other algos is also a key concern given the effect this could have on the market.
AlgoGuard is an independent algo test service based on TraderServe’s unique orderbook emulation, antagonist algorithm and stress test technologies. The aim is to give an investment firm confidence that its algos will behave stably in live trading and that it can satisfy the regulators that it has gone to sufficient lengths to ensure that the algos will not contribute to disorderly trading conditions and striven to meet the stringent new algo testing requirements by including independent testing yet without having to disclose their algo IP.
AlgoGuard allows your algo’s orders to interact in the same order book with other antagonist algorithms and with a realistic emulation of the market but completely isolated from live trading. It examines how stable the algo’s behaviour is in normal and extreme conditions, and can also explore the algo’s disposition to contribute to the sort of emergent market disorder that can be caused by the interaction of diverse algorithms in a live order book. No amount of testing with market replays can address these concerns, but AlgoGuard can offer independent protection to help prevent going live with a vulnerable algorithm. Investment firms receive an orderbook feed from the AlgoGuard server and send their orders and cancellations exactly as in live trading. AlgoGuard tests in both normal and stressed market conditions and with and without different classes of antagonist algorithms. Its reports then provide pass/fail testing with drill-down analytics as well as independent, audit-trailed evidence of algorithm stability against documented test schedules.
AlgoGuard accepts limit and market orders and allows matching within its order book on a FIFO or Pro-Rata regime. Where published, other order types and matching rules can be added on a bespoke basis.
Next Generation Algorithmic Stability Testing Trading without disclosing your Algo IP
- Seamless & Transparent: Client's access to AlgoGuard can be customised to be similar to the intended trading venues.
- Representative & Isolated: Statistically samples historical orderbook data and maintains a representative market in the test matching engine whilst being completely isolated from any risk of live trading.
- Responsive & Dynamic: The emulated market closely resembles the real market in all important respects including spreads, volatility, volume, fill to cancel ratio. In consequence it responds realistically and dynamically to the algorithms under test.
- Private: Clients' algorithms never leave their own servers, only orders are sent to AlgoGuard in response to data and fills sent to the client's algorithms. These orders combine in a dedicated test matching engine only with AlgoGuard algorithms and are isolated from other clients' orders.
- Stress Tests: Customisable test schedules under the control of the client can stress test the market in a variety of ways so that the potentially disruptive behaviour of connected client algorithms and execution systems can be assessed.
- Compliance & Logging: Helps clients to meet the increasingly stringent testing requirements for algorithms and execution systems being imposed in many jurisdictions. Stores test results for as long as required by regulation.
- Pass/Fail Reporting & Traceability: Dashboard provides control and reporting to client's staff. Provides Pass/fail reporting for member algorithm's provocation of market disorder with drill-down analytics.
- Multi-market: For stability testing of multi-market client algorithms there is the option of including customisable test matching engines to emulate multiple trading venues published rules.