Opinion: 1. The algo's running within broker infra in co-location is entirely different from algo's run by retailers through broker API from their PC.
The algo's running in co-location have latency advantages over others and also pose direct risk to market due to any erroneous code, so those algo's must be audited and regulated for fair access, but not for those run by retailers from the PC.
2. The orders placed through API is just another way of routing orders other than broker's Desktop, Web and Mobile applications.
3. Since the orders are still going through RMS and OMS of the broker, any kind of erroneous orders will be filtered at OMS level, and only those validated orders reach the exchange.
4. The very purpose of an API is automation; API's can have throttle limit but not regulations.
5. Broker's API provides a cost effective and simple solutions for retailers who can't have access to co-location servers.
6. Trying to regulate API is regressive in nature. Regulating because of some outliers will bring chaos to the existing ecosystem.
Suggestion:
1. Instead of regulating the usage of API by retailers, SEBI should draft a framework to ensure that the broker's OMS is capable of handling any kind of abnormal traffic.
2. Separate servers (i.e. isolated from OMS) may mandatorily be commissioned by broker's for routing order requests by API, thus preventing any direct risk to OMS. (like using cloud flare for preventing DDOS)
3. SEBI may enforce browser based login for all brokers API
4. SEBI may bring uniform throttle limit across all brokers for APIs (like orders per min/day)
We can't avoid completely but we can contain it to some extent. Your fill price depends on your order priority and the size of your precedent orders. (2/n)
There is no one method that fits to all systems and users. You have to adopt different methods that suit your trading system. Ex: The method used to exit option buying may not work for exiting option selling, the method to enter pull back may not work in breakout entry. (3/n)