I have been into Algo Trading since 2014, let me answer this from Indian Stock market perspective. Before getting into the future of Algo Trading, let me explain how Algorithmic Trading in India was few years before, how it is now and how the future of is going to be.
Back then, I was working full time in a MNC and it was difficult to manage my trading and job together, so I was trying to automate a simple breakout strategy.
Whenever Buy/Sell conditions are met, the charts in Amibroker charting tool generates signals like below, and the respective orders has to be placed in the broking account.
But in order to achieve this order execution, I should have
- Real time data feed
- Amibroker charting platform
- A bridge/plug in that can connect charting tool and broker account
- Order Management system that executes the trade.
- A cloud server
I have to subscribe to Real time data feed from authorized vendor , where the price data feed goes to charting tool – Amibroker, and the trading strategy has to be coded in AFL (Amibroker Formula Language) so that you can create Buy/Sell signals. Then I should purchase a Bridge/Plugin that connects my amibroker and trading terminal. So that these buy and sell orders can be directly pushed to my trading account.
In order to avoid any disconnection issue, I need to host all these services in a cloud server, so that none of the service are interrupted and order execution goes smoothly.
The problem I faced with such flow was mainly due to multiple layers of connection,
- There will be days where my real time data feed vendor might face connection issue, so I end up missing trades at times.
- Since Buy/Sell signals are generated from AFL code, at times I try to modify some lines or functions through google help where I end up messing the whole program, as I do not know coding back then.
- The plugin/bridge provider which connects Amibroker and Trading terminal, goes down many times, so the Buy / Sell order generated by my trading strategy would have not been really passed on to my trading terminal, so I end up missing the trades.
- Real time data feed costs – Rs.1500 per month, Amibroker costs around Rs.20,000, the bridge/plugin cost Rs.40000 per year, cloud server costed Rs. 20000 per year. So over a yearly period, I end up paying huge money for infrastructure just to execute my trades.
After getting into fully automated system, it din’t really help me, in fact it required frequent monitoring of my cloud server to ensure everything runs smoothly.
As the technology grows, more and more brokers came into market by providing their own Automated platforms, brokers like Aliceblue started giving out their API platforms for free to retail traders, so that anybody can design their own trading strategy and execute it flawlessly.
So I started focusing on a Order flow system that can eliminate the need of Real time data feed and charting tool Amibroker.
Using brokers API and other tech, we created our Trading bots https://squareoffbots.com/ and deployed on Amazon cloud server based on the specific trading rules, where anybody who trades based on these inbuilt strategy in our platform, can just input their capital, the platform would automatically place the trades. Just with few seconds a day, people can execute trading system flawlessly.
By eliminating the real time data feed, amibroker, plugin, vps, we are able to save considerably lot of money now. So the cost of Fully automated trading has gone down significantly now.
Alternative data sets are information about a particular company that is published by sources outside of the company, which can provide unique and timely insights into investment opportunities. An alternative data set can be compiled from various sources such as financial transactions, sensors, mobile devices, satellites, public records, and the internet
Soon retail traders will start building trading systems based on alternate data sets. As price data is accessible to everyone out there, people would look for other information which can affect the stock price.
In developed markets, many hedge funds/ Investments bank have already started using these alternate data sets years before. Analysts at UBS have used satellite images of Wall Mart parking lots, if the parking lots are crowded, it means more people are walking into Wall mart which means, more sales, more profits and which in turn means, rise in stock price.
So just real time tracking information of satellite images helped these firms making in higher profits.
With the growth in Social networking apps, there are immense amount of data that gets generated every second. When it comes to investing, still traders/investors following the herd mentality dominates, with billions of users in social media, even a single tweet can cause wild move in stock price.
Kylie Jenner made headlines in early 2018 by firing off one single tweet that set off a downward spiral for Snap’s stock ($SNAP). A headline from CNN Money read: “Snapchat stock loses $1.3 billion after Kylie Jenner tweet.”
But these alternate data sets are very huge, a normal retail trader with his laptop at home cant do all the computing tasks and come with an interpretation quickly. Just having the data is not enough, we should have necessary infrastructure to process all these information faster and interpret the data to make investing decision.
Going forward, there would be many startups which would use AI/Machine learning to process all these alternate data much faster and help retail traders with faster access to information at affordable cost.
With regards to Indian markets, many retail traders still rely on news channels to get the latest updates on stock news, and they keep monitoring multiple channels, looking out for hot news which move the stock up or down significantly, so that they can get in and make some quick profits and move out. Other than price data and these news data, many doesn’t focus on other data sets. As per exchange rules, any listed entity should submit their press release information to exchange first and then it only it gets published to other media.
During 2019, there was a news about a stock Amara Raja Batteries, where they informed exchange about the termination of agreement regarding a joint venture. The news first hit the NSE and then after a while went to other TV channels.
But look at the reaction of stock price. There was a immediate reaction in few seconds, stock price started plummeting. So there are certain firms, funds, big investors out there who could process the information much faster and able to make decision much quicker before other retail traders jumps in.
But when do retail traders would come to know about it, may be next day in the newspaper.
If we dig deep, there are certain instances where similar news came out for other companies, like Hero Honda terminated their agreement with Honda, how stock price reaction at that time, studying such historical alternate data sets would provide some inputs to interpret the news faster.
And I believe, the future of Algorithmic trading, specifically in India, will be about solutions offered by startups that are focused into alternate sets that could help retail traders to make investing/trading decisions faster.