I recently developed an RPA app and am wondering if anyone would like to team up to build a trading robot. I'm more of a long term investor rather than a trader so I'm not quite sure what to tell the bot to do. The upside to RPA is that it sees the screen and uses the mouse/keyboard just like you do so there's no need for API access. The downside is that it needs to visually see what to trade so it will not be as fast as a system that trades non-visually. Let me know what you think.
Hey A, you and millions of others. Folks develop these apps, algos, robots & whatever, yet they don't know how to trade/invest in the first place. Arnie said : ,,You can have the best ship in the world, but if you don't know how to opperate it, where are you sailling, - you will simply drift around" Not that many, if any, for free, will join to a such project. One has to ,,Bring more on the table" than - i have this robot. Or do it on your own. And if you don't want to do it by yourself, then probably, it was started for the wrong reasons (?)
Applying RPA with trading models is like applying the steam engine to the automotive industry. Before most technologies fell into their respective categories and developed buzz words, the financial industry already tried them. Why create a layer of software to automate a trading strategy when there are platforms that automate trading natively? However, here is where you may apply RPA: Legging a spread. Spreaders run about $400 a month. Build it, license it and make some money.
Good points, there are definitely downsides to RPA but there are also downsides to APIs. The level of expertise needed is a huge barrier and limits progess in some ways. Perhaps instead of using less efficient RPA within the trading layer maybe use it in a less critical data gathering layer.
I hear what you're saying. But I think the lone wolf attitude of needing to know and do everything yourself can hold traders back. It certainly takes more than one person to sail the best ship in the world. RPA is geared to be affordable and easy to use for non-programmers. At the very least, one should use it to prevent getting carpal tunnel from running thousands of backtests.
I could be interested in it. Could you provide more info about the type of algorithm and the broker you are using? You can write here to get more people interested or in pvt.
Hi Nakesha - it could be used with any broker's platform because RPA operates a PC just as any real person would. Their only clue might be that the user is clicking sooo fast...i accidentally triggered Twitter's bot detection algo this way. The algo is up to you...I don't even need to know it. Tbh I'm much better at long term investing but would really enjoy seeing someone use RPA to create an awesome trading bot.
but, ...this looks like work. CQG Spreader A Competitive Edge https://www.cqg.com/products/cqg-spreader This sophisticated order management and ultra-low-latency trade execution interface routes orders to the CQG Spreader Core, our co-located exchange servers. The CQG Spreader Core uses proprietary algorithms to manage your orders, allowing for modification and execution of the synthetic spread leg orders to occur in less than one millisecond.
I do support this idea. I support any idea where individual traders help each other because the entire system is rigged against us. Brokers, Exchanges, Software companies all POOL resources. But traders are all atomized. Thus, its just a matter of time before the entire cash balance of the trader ends up in the hands of the Brokers, Exchanges, Software companies. Traders need to POOL resources. We need open source software for everything. The more we voluntarily share with each other the less we are forced to "share" with Brokers, Exchanges, Software companies. We need a platform to enable this sort of collaboration.
I'm successfully using ArcadeTrader.com to develop and backtest strategies. Collective2 has more users, but it's not FREE and doesn't support Javascript. I love Arcade Trader, but the "share" feature is not available yet (I wrote multiple times to the team, but it's not clear when it will be available). With that feature you could easily share a strategy with other author and co-work on it. Otherwise a github repo with just scripts in it, but how can we backtest and have the same results?