Potential of Electronic Trading Webinar - Waters Technology
Walters Technology conducted a webinar on electronic trading on . This blogpost summarizes a few of the points mentioned in the webinar.
How to transform your operations to unleash the full potential of electronic trading
Financial firms are finding it essential to shift to electronic trading in generating alpha to stay competitive. As a result, buy-side and sell-side financial firms will need to invest in low latency direct solutions, including consolidated market data feeds & historical data services.
Waters Technology and Refinitiv are delighted to host this webinar, featuring a line-up of technology and data experts who will discuss the latest trends and innovations to support electronic trading and automation of middle and back-office trade processes.
On the agenda:
- Market fragmentation and global pandemic: how is the APAC region overcoming challenges to embrace electronic trading?
- Defining objectives in transforming the manual middle and back-office processes
- How to take advantage of direct data feeds that provide exchange data at a low latency
- The future applications of Al and ML technology to the trading of financial instruments
- Case studies of generating alpha through automation and direct data feeds
Speakers and Moderator
- David Sharratt, Chief Data Officer, Commerzbank AG
- Kerr Hatrick, Executive Director and Quantitative Strategist, Morgan Stanley
- Rob Lane, Head of Real-Time Feeds, Refinitiv, an LSEG business
- Moderator: Wei-Shen Wong, Asia Editor, WatersTechnology
Notes
- Morgan Stanley
- Data and the way one deals with the data is one of the greatest challenges of
electronic trading
- real time price feeds
- order book data
- news feeds
- data that one uses after to validate model and post trade analytics
- Brokerages and buy side firms - Data is critical - Models come later
- Kdb is one of the databases used in Morgan Stanley
- VWAP algo example
- Make sure that the metrics provided by kdb
- Prices are not the place where the problem lies. Many data vendors do a good job of providing price data
- Real time price and Real time volatility stats are stored in kdb
- It is hard to get people who have experience in dealing with data
- Economies of scale do not work well with data. Available of latency options
- Provide some sort of interpretation of what algos are doing
- Different stocks move in different kinds of news
- Sources from RavenPack and MRN
- Algos
- News associated with negative sentiments have a greater effect in some regions
- News associated with positive sentiments have a greater effect in some regions
- Challenges in moving model to production
- News event and secondary news event that follow the main event. What is the way an algo decides that it has already seen a news item
- What does novelty mean for a story ?
- When looking at news and deciphering the sentiment - understanding historical distributions on the sentiment is important
- How do you train novelty score ?
- Large amount of news data per market per stock
- Need other kind of info such as beta of the stock
- Encompass the intuition of the trader
- Increase participation rate as there was there a news event
- Decrease participation rate based on alternative data
- Why get in to ML ?Clients will be able to understand the behavior of the algo
- Getting and using data is all about getting the right people to work and
leverage internal and external datasets
- Bottlenecks for storing and processing data has been removed
- Data and the way one deals with the data is one of the greatest challenges of
electronic trading
- Commerzbank
- Offensive and Defensive data
- Make sure that you have the
- Regulatory reporting - Defensive side - has a lot of information has wide as possible
- Not just about price but enrichment of the data
- Data going in to algos
- Price
- order book data
- news data
- static data on the day
- beta
- mcap
- all attributes that give the algo the kind of stock it trades
- delayed data from crosses
- price data is not all the same and hence need different ways to consolidate and aggregate data
- Regulators are getting more heavily focused
- How do you manage the data flows in an organization ?
- Governance is one of the most critical aspects
- Swim lanes for analyzing data flows in an organization
- Refinitiv
- Changes at the venue level - Continuous push for more data and technology
- 2022 - 400% increase in messages per second since 2017
- Taking the data is easy part
- PCAP data that can be used in the applications
- We don’t get just get the exchange data but add fills and value added attributes
- Give variations of the data
- Cloud
- PCAP
- Tick History
- We need people who understand data, finance, technology, software
- 80% of the data management strategy has not changed since a decade - alarming
- Different challenges
- Consolidated feed - tick history - It is a different challenge
- Inevitable ticks get dropped. Challenge is to make sure that fills are made
- Data in a raw form - PCAP data
- PCAP - taken off the wire
- Interest from the regulatory side
- Give clients various options to consume data
- Costs are rising to collect, normalize and disseminate the data
- Clients opting for what types of distribution - all kinds
- Giving clients to choice options to give easy access to data
- Changes at the venue level - Continuous push for more data and technology
- 75% of the audience in the webinar do not use ML and AI
- Focus on Human Capital - Getting data right is largely a matter of hiring the right people both to archive data and access data are more sophisticated now.