Should You Modify Algo Trading Signals?
A while back, I received an e-mail from a relatively new trader. He was using an algorithmic trading system (also know as an "algo") he purchased, based on its alleged stellar performance. "I've been using this trading system for a few weeks, and sometimes I just KNOW the algo trading signals it gives will be wrong. What do you think if I'd take the signals, and combine them with work of my own, to get modified signals?"
Here is what I told him...
Obviously, since you bought the trading system, you can do whatever you like with the signals, since the final decision to trade or not rests with you. The signals are only a recommendation - you control what happens in your trading account. Don't underestimate the magnitude of this fact.
I can tell you that in trading strategies I develop, I use the signals AS IS, with zero modifications. I trade most of my algo trading systems fully automated, where it is almost impossible for me to interfere. I'm not smart enough to know when the automated trading signals will be good or bad, so I've taken myself out of the equation. It is funny, but the days I think "why did the system go short/long TODAY?" usually turn out the best - the trading system is typically right, and my "hunch" is usually wrong.
But that's me. I make sure I have the confidence in the system BEFORE I start trading it. It sounds like you need to develop that confidence, too. Here is how to get it:
Here is what I told him...
Obviously, since you bought the trading system, you can do whatever you like with the signals, since the final decision to trade or not rests with you. The signals are only a recommendation - you control what happens in your trading account. Don't underestimate the magnitude of this fact.
I can tell you that in trading strategies I develop, I use the signals AS IS, with zero modifications. I trade most of my algo trading systems fully automated, where it is almost impossible for me to interfere. I'm not smart enough to know when the automated trading signals will be good or bad, so I've taken myself out of the equation. It is funny, but the days I think "why did the system go short/long TODAY?" usually turn out the best - the trading system is typically right, and my "hunch" is usually wrong.
But that's me. I make sure I have the confidence in the system BEFORE I start trading it. It sounds like you need to develop that confidence, too. Here is how to get it:
Modifying Algorithmic Trading System Signal Tips
1. If you are developing your own algo trading systems, follow a development process that you are confident in. If you are buying a system or signals, only buy trading systems or follow signals from trading system developers you trust. Ask a lot of questions before you buy. Establish a general comfort level.
2. Evaluate the historical performance in detail. For example, can you trade that type of algorithmic trading system, whether it scalps, swing trades or position trades? If the system adds to losers, will you be able to always do the same? Do the historical results look too good to be true? That's a tipoff that the trading system might have been curvefitted or over optimized.
3. See if you can track the performance in real time for a month or so, before you buy. If you develop your own algo trading system, this should not be an issue (and I recommend 6 months or more of real time watching in this case). It is amazing how many algo trading signals that look good today will look terrible after another month's worth of real trades. Treat this as a "cooling off" period to prevent buyer's remorse.
4. If you decide to buy, start small. Give the system a chance, but with limited capital. You'll quickly find out if the system is something you can handle, both financially and psychologically.
5. Once you are ready to go "live" with full size trading, and you've satisfied all the conditions above, NEVER override a signal, unless you have the ability to fully backtest and evaluate it yourself. Modifying trading signals on a whim or a feeling is almost like flipping a coin - it rarely gives you any kind of extra "edge." And make sure you are using a reliable trading platform, like Tradestation.
Here is another way to think about modifying trading system signals: you can buy a new car and modify it to get better performance. It might work much better for a short while, but then maybe the engine will seize (and maybe it won't). You just don't know. It is the same thing with modifying or overriding signals from an algorithmic trading system - you might make improvements in the short term that really hurt the long term performance. You just don't know.
Because of the possible unintended and unknown consequences, I recommend, for systems you believe in, that you follow signals as is. The key is to believe in the system before you develop or buy it.
About The Author: Kevin Davey is an award winning private futures, forex and commodities trader. He has been trading for over 25 years.Three consecutive years, Kevin achieved over 100% annual returns in a real time, real money, year long trading contest, finishing in first or second place each of those years.
Kevin is the author of the highly acclaimed trading book "Building Algorithmic Trading Systems: A Trader's Journey From Data Mining to Monte Carlo Simulation to Live Trading" (Wiley 2014). Kevin provides a wealth of trading information at his website: http://www.kjtradingsystems.com
Copyright, Kevin Davey and KJ Trading Systems. All Rights Reserved. Reprint of above article is permitted, as long as the About The Author information is included.
Kevin is the author of the highly acclaimed trading book "Building Algorithmic Trading Systems: A Trader's Journey From Data Mining to Monte Carlo Simulation to Live Trading" (Wiley 2014). Kevin provides a wealth of trading information at his website: http://www.kjtradingsystems.com
Copyright, Kevin Davey and KJ Trading Systems. All Rights Reserved. Reprint of above article is permitted, as long as the About The Author information is included.