What is Broker Speedtest?
Broker Speedtest is a new aid for algo traders to assess and compare broker API responsiveness. It’s been designed and engineered at AlgoTest to offer transparent data on the API performance of different brokers.
This tool is relevant foralgo traders, traders who plan to move into algo trading, and even traders trading manually using third party apps built on top of broker APIs (check out AlgoTest’s Strategy Builder). We hope that with this tool we can help traders understand the nuances involved in firing trades from AlgoTest to different brokers’ APIs.
Why Did We Make ‘Broker Speedtest’?
At AlgoTest we believe in transparency when it comes to our customers.
So Broker Speedtest has been made with that very intention. But there’s more to it as well.
There are a lot of misconceptions about broker API speeds out there. Rajat, one of the founders and CTO of AlgoTest initially tweeted a snapshot of different broker API speeds.
🚀Part 1🚀
— AlgoTest (Automate TradingView + Free Backtesting) (@AlgoTest_in) April 20, 2024
90% of orders placed on 16th April 2024(Banknifty & Finnifty weekly expiry) took less than 'n' milliseconds to get an acknowledgement.
Only APIs through which more than 1000 orders were placed are taken in consideration.
But placing an order is not the only thing… https://t.co/2gYt5p1IDC pic.twitter.com/bmbMcCXDyc
It stirred the pot and caused some controversy and so in order to address that, Raghav, our other co-founder, sent a longer tweet with a lot more context. He also used the same time period but a different dataset to rejig the broker rankings we had come up with in the first tweet.
It seems that this tweet caused a bit of a stir last weekend.
— Raghav (@rogue_hft) April 27, 2024
Some interpretations were off base, so here’s a clarification:
Sample size problem
The sample size was way too small to draw any meaningful conclusions:
1. Only one day’s data – The timing and relative positioning… https://t.co/dP8T5hwKYa pic.twitter.com/xuNkK7k7QO
But here’s a couple of misconceptions that we really wish to address :
- Speed is paramount when it comes to API trading for retail traders
- The fastest broker is the best broker
Let’s understand how orders from AlgoTest work
In order to address these misconceptions let’s first understand how AlgoTest fires orders to brokers and what we measure and present.
Step 1
AlgoTest servers fire an order to the broker’s servers based on the strategy parameters.
The Broker sends us an acknowledgment of the order.
Step 2
The Broker’s RMS approves the order and only then is it forwarded to the exchange.
Step 3
The Exchange acknowledges that order, generates an Order ID and then provides that Order ID to the broker.
We only measure Step 1 of the process.
We only have data for when we fire an order(s) and when we receive the acknowledgment from the broker.
We cannot track the other steps because those are performed inside the broker’s internal infrastructure. We encourage the brokers to release metrics that’d help “Api wale retail traders” to make informative decisions.
But how do we interpret this data?
Here’s an example of Zerodha latencies for last week. These latencies are represented as a box plot to give a more accurate picture of how communication over the network should be measured.
Here are Zerodha and Finvasia latencies on the same chart to compare their responsiveness for placing new orders (Place Order requests).
Now it might seem that Finvasia is faster than Zerodha overall. However, let’s examine the 15 mins just after market open and 15 mins just before market close – the peak traffic hours where the highest volume of trading happens
Next, let’s zoom in and see data for a specific time – specifically this particular period.
We find here that on average, Zerodha might be slower in acknowledgment but Finvasia’s max time is higher than Zerodha’s. Overall, variance in Finvasia’s response time is higher during this time. BUT, again, this data doesn’t reflect what happens to your order after the broker has acknowledged your request, which is much more important!
How much time did the RMS check take? When does the broker forward it to the exchange? When does it actually reach the exchange’s firewall?
We have no idea here.
As Chintan, our CFO, would say – “Seedhi baat ye hai ki isse mere PnL pe koi farak nahi padta.”
Does that make Zerodha better than Finvasia?
Not necessarily. We would need to study a lot more data in order to come to that conclusion but also this is not a decision that can be made without factoring in other factors like broker service, customer care, etc.
Holistic Evaluation:
While speed is important, it is not the only factor to consider when choosing a broker. Quality of service and reliability are crucial aspects that should not be overlooked. A broker’s ability to provide consistent and reliable service is paramount for successful algo-trading.
So if speed doesn’t matter, what affects your strategy?
The main effect driving your PnL will affect your strategy. Loosely speaking, if implied vol is higher than realised volatility over the lifetime of the option, your short option strategy will make money in the long run. In this case, IV is a much slower moving parameter, not usually affected by a few seconds here and there. On the other hand, if you’re trying to capture intraday momentum, a few seconds could make all the difference. Bottom line, your entry and exit times don’t matter as much for option selling as they do for a momentum strategy.
In fact, even for a short term momentum strategy which is more price sensitive, this latency won’t matter if the strategy is setup the right way.
Which is why we have features like the SLL order which allow for trades to sit at the exchange waiting for the right entry/exit conditions to trigger.
Instead of focusing on speed, focus on understanding how best to set up your strategies to take advantage of the factors that actually drive your PnL in the market.
How you can use our Broker Speedtest
We aim to provide a number of data points related to API performance. Here’s how you can make the most out of our offerings going forward:
* Stay tuned for further updates:
Response Times: Broker Speedtest gives you detailed information about broker response times for Place Order, Modify Order, and Cancel Order.
Comparative Analysis: The tool offers you the ability to do a comparative analysis between the brokers of your choice. This allows you to see how each broker stacks up against the others in terms of API performance.
Daily Updates: The data is updated every 15 minutes during the market hours itself. So you can track your broker’s API speeds and performance and get an idea in almost real time if your broker’s infrastructure has blown up.
To know more about this headover to openbroker.in.
Should You Switch to a Broker Based on Speed?
No, but Consider Trying New Brokers
While speed is an important factor, it should not be the sole reason for switching brokers. If you are curious and want to try a new broker, go ahead and do it. However, make sure to evaluate the broker on various other parameters as well.
Focus on Service Quality and Reliability
When choosing a broker, focus on the quality of service and reliability. A good broker should provide a seamless trading experience with minimal issues. Ensure that the broker can meet your trading needs comprehensively, beyond just offering fast API speeds.
We’re operating in the land of milliseconds here. The exchange might be executing 10s of 1000s of trades every second for every option contract. Retail traders always operate at a latency. By the time they get the information to make a trade, the market has moved already.
The HFTs which are constantly after your money operate in nanoseconds. They provide liquidity so they need to operate in the nanosecond landscape. You as a retail trader are taking liquidity, and a few milliseconds or even seconds here and there should not affect your PnL in the long term.
While it is important that you get in the queue as fast as possible, you should first make sure that you are actually getting in the queue and that too reliably. Only and only then should you worry about whether your position in the queue is 10 or 50.
Broker Speedtest is engineered for Algo-Traders
Broker Speedtest by AlgoTest is a valuable tool for any algo trader looking to make informed decisions about broker selection. By providing transparent and detailed data on broker API performance, we hope to give you all the information you need to choose the right broker for your trading strategy. Remember, while speed is important, focusing on the overall quality and reliability of the broker will lead to better trading outcomes.
We’re trying to help you in making an informative decision. Stay tuned for more updates!
