(To the tune of "Do You Want to Build a Snowman" from "Frozen")
Do you wanna build an algo?
Come on, let's buy and sell
I'll never lose anymore
This program's for sure
Winning trades until the bell
We used to call our brokers
Now they sit alone
And I won't tell them why!
Do you wanna build an algo?
It doesn't have to be an algo.
Managing emotions is absolutely the toughest aspect of trading. Even the best traders can let their emotions get the better of them on a bad day. I do believe that's the biggest appeal behind algorithms. I'm not talking about high frequency traders here, which clearly rely on computers to make trades, sometimes hundreds of them, in the blink of an eye. I'm talking a more simple algorithm -- buy if "this" appears and sell if "that" appears.
While I'm still a big proponent of trading, whether it be technical or fundamental or a hybrid of both, I have also spent the last several months developing some automated systems. I thought sharing my experience might be worth a read for anyone interested in pursuing the same idea. It's a big lengthy, so we'll divide it up over the Christmas week here. My suffering and recounting this challenge will be my present to you.
If you've ever been on vacation and taken a shuttle from the airport to your hotel, then what I'm about to tell you may be no surprise ¿ well, as long as you were on a shuttle with Vegas virgins, first timers to Sin City. My last trip out, I was the last person off the shuttle. I love taking the shuttle from the airport. It's like a party. Everyone is ready for a wild time and big winners. Spirits are high, folks are friendly and laughter fills the air. The eyes of the first timers beam with optimism.
And then ... well, there is the ride from the hotels back to the airport. This, this is not the same. The ride back, especially early in the day, is a stark contrast to the first ride. While the ride from the airport to the hotel is full of optimism, the ride from the hotel to the airport is a living art of realism. Heads are hung much lower for most. Eyes are bloodshot for some, with murmurs of folks just wanting to get back home. At least these folks still have their airline tickets. The easy money, for many, wasn't so easy.
I say this because it is the same for trying to build your own algorithm. Thirty minutes and a thesis aren't going to deliver what you want. The market isn't a pizza place. While one or two folks may find a gem quickly, the result and setup is likely to take tens or even hundreds of hours, just to find something you are even willing to try with cold, hard cash. Come on, seven!
So where does one start? At the beginning, of course. The starting line is the easiest to find. Simply start with an indicator or measurement, either technical or fundamental, which catches your interest or one you've heard referenced many times before. I'm a technical guy, so for me concepts like RSI (Relative Strength Index), Moving Averages (specifically crossovers), Stochastics and Parabolic Stop and Reverse along with some other indicators are where I would start.
Once you pick your poison, then it is time to study history. For me, that meant looking at lots and lots of charts for a single security, often an ETF. I was working on an intraday trading system, so I scanned hundreds of daily charts. The initial view is not just about pattern recognition, but also about commonality and studying the biggest moves. Isolate the biggest moves first.
One school of thought is to simply test your thesis to see if it works. To me that is too random, so I study those big moves. Are there any? Are there a lot? Do they happen both in moves higher and lower, or are there just a few for decoration? I really don't want to try and catch a lot of very small moves. Commissions and slippage could kill you on a lot of small move trades.
If I have a good sampling of big moves isolated, I bring in a half-dozen indicators. While I have my preferred indicators, I want to keep an open mind here and am willing to use anything as long as I understand what it is. If I don't, then I learn. If I can't get a grasp on what it tells me, then I toss it aside.
I want to see if there is one, or preferably two, indicators that have the same pattern around the big moves as the trigger. Simply put, I want a confirming mechanism to make sure I catch the big moves with hopes the mechanism will help me avoid chop, or some bad entries.
Later this week, I'll delve into the back-testing process, which includes fine-tuning entries and establishing sell points. It's a pretty lengthy process, so it's best to take it one step at a time.
At the time of publication, Timothy Collins had no positions in any of the securities mentioned.