London, August 16, 2023
Today, after about 4 years and 1,240 users, I’m announcing that Arcade Trader will stop receiving new subscriptions for its Algo Trading cloud platform. In about a month, the service will be suspended and any pending credits from the remaining subscriptions will be automatically refunded.
The users that are running Arcade Trader in standalone mode now have the option to request the entire codebase to maintain and evolve the product if they choose to. If you’re one of them, please drop an email to support — at — arcadetrader.com.
Why are we shutting down Arcade Trader?
Unfortunately, our SaaS model didn’t become profitable. This is the end of a long history with multiple pivots.
Initially, our goal was to provide a marketplace of strategies, where algo-traders and end users can meet to buy and sell strategies. That was our mission: to provide easy access to edge-fund-level strategies with proven and documented track records. With the Marketplace, anyone could invest and trade by using an algorithm that runs on their own trading account, completely autonomous.
Even if technically we were ready to launch the Strategy Marketplace, our major challenge was purely legal, especially because our users are based all around the world and each country has different rules.
About 2 years ago, we permanently halted the release of the Marketplace. Instead of shutting down the service, we decided to go all-in with a Freemium SaaS Subscriptions model only for the algo traders.
With Arcade Trader, you could backtest your strategy against the last 10 years of US Stock Market data. Backtesting at the tick level is extremely expensive in terms of CPU and IO resources. For this reason, we created an elastic cluster where the user could buy resources by purchasing “Arcade Tokens.”
We doubled down with new features, like the support of Natural Language to write algorithms in plain English. This was a pretty cool feature you can’t find in any other products.
Also, we created the new Portfolio Management where you could copy an ETF composition right into your trading account, keeping your share allocation balanced without paying any fees for ETF trading.
Then we decided to open up the Professional Trader’s most requested feature for everyone: the ability to base a strategy on the real-time stream of ticks (trades and quotes) received directly from the exchanges. We’re talking about thousands of events per second with <5ms of latency between the event and the order.
It seems that not many algo traders out there were interested in a cloud solution. At least not enough to make the service profitable.
After all, the “Algo Trader” is an exceptionally rare human species: half developer and half financial advisor.
Here at Arcade Data, we love to play with data. After all, Arcade Trader was the perfect challenge for our ArcadeDB DBMS to crunch real-time data from the US Stock market, collecting more than 10TB of data in about 4 years.
Thanks to Arcade Trader, our team was able to implement new features in ArcadeDB and improve the overall performance of the engine with Graphs & Time Series structures.
Our next challenge will be focused 100% on AI. We just released the Vector Model for ArcadeDB. In the next weeks, we will announce a new revolutionary project that will be entirely based on ArcadeDB, Vectors, Embeddings and LLM.
Thanks to all of our 1,240 users for using Arcade Trader!
Founder & CEO Arcade Data