Your Trading Questions Answered #2
Welcome to The Cypher List Q&A column, where I take questions from readers and answer the best ones from the perspective of an active automated trader.
Anyone can submit a question via twitter or email, but special priority is given to our premium subscribers, so join them now to take advantage and ask away.
What do you see as the dominant programming languages over the next 10-20 years?
I suspect JavaScript will continue to hold ground as the dominant language of the web. Frameworks like React and Node are securing it’s position as a popular ecosystem for both front and back end development.
Long time readers know I love Python and I only see it gaining in prominence over the next few decades. Not only is it cementing it’s spot as the best language for new developers to learn, but people are also realizing that it is more than capable of doing heavy back end work both on the web and in the financial sector.
There are other languages I think are interesting and will very likely be trending upwards over the next few years but given you asked for the dominant ones I’ll stick to those two.
How do you evaluate your "work-in-progress" algos before giving them real money?
Forward testing is king. I talk a lot about how to effectively conduct backtests for your algos in progress, but there are always going to be potential pitfalls when it comes to retroactive testing on historical data.
My process is that I use backtesting primarily to form the core of my strategies then switch to forward testing to verify and refine them over time. There are two main factors to consider when it comes to sample size: time under test and trade count. You want a minimum of a month for the former and a minimum of 20 trades for the latter.
The kind of thing I’m looking out for in my forward tests is significant deviation from backtest results. I don’t expect them to match up 1-1 but if a strategy is killing it on historical datasets then completely flops on fresh data, I know something is likely wrong with my backtesting methodology and the strategy needs work.
I’m a professional developer but never traded, do you think I should work on an automated system? Or manually trade to learn first, then automate parts as I go?
If you’re in this position, definitely start out by learning basic market mechanics before doing anything else. You shouldn’t even be thinking about bots right now. The first thing you should do is pick a sector you are interested in (crypto, precious metals, pharmaceutical stocks, etc).
After you’ve picked a sector, make a few manual trades in those markets with a small amount. Your aim is not to make lots of money from these, it’s to learn the mechanics of trading. How the broker feels, what the order flow is like, placing stop losses. All these things are fundamental.
After a while of doing this, you’ll start to notice some basic patterns. Build some basic strategies around what you see. If you know that you want to have bots trading for you eventually, perfect. Automate gradually. Piece by piece. There’s no rush to do it all at once.
Feel free to DM me for advice along the way.
Do you think retail traders can compete in HFT (High Frequency Trading)?
No.
Site: cypherscope.com
Twitter: @CYPH3RSCOPE
Masterclass: cypherscope.com/buy-masterclass