Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
I am trying to learn about automated trading, for a few months now I have developed 2-3 strategies a week then thoroughly tested them to see how they would play out and time and again they lose money, i even spend a lot of time trying to layer in more rules about stop losses and time of day. but just no luck, whats the best way to essentially build a functional automated trading strategy. I know its out of the question to try to buy a system from someone but I have just gotten so many vague responses, I mean someone must know, I don't see how automated trading could thrive if everyone involved was losing money on it.
Can you help answer these questions from other members on NexusFi?
in my (not great) experience, strategy building is a long and tiring task, so I would certainly not expect someone to come up with 2-3 good strategies a week..a year, yes!
To be more practical, I would recommend the following books that helped me in setting some good practices in my strategy development. They all refer to automated trading.
1 - Trading Systems: A new approach to system development and portfolio optimisation, by Emilio Tomasini and Urban Jaekle
2 - Quantitative Trading: How to Build Your Own Algorithmic Trading Business, by Ernie Chan
3 - The Evaluation and Optimization of Trading Strategies, by Robert Pardo
Do you have good tools in testing, so that you won't be wasting time with those at least? I am using Multicharts (I understand many traders in this forum are using NinjaTrader), with a data feed from Tradestation.
There is a webinar tonight about this subject at Big Mike's, if you are an Elite member you can watch it. I don't know how useful it will be, but in my experience a trader will pick up small nuggets of wisdom left and right, and slowly build their own knowledge and personality.
Don't hesitate to give me a shout for whatever else you may think I could assist you with - your questions are quite general, feel free to be more specific. In the same spirit though, I could offer this 'typical' advice that you will see in many other places also:
- make sure to use in-sample and out-of-sample data for your tests
- make sure that your strategies have enough data points (it's a whole issue what is enough!)
- try to avoid many rules to avoid overoptimisation
- and ensure that your testing includes realistic slippage and transaction costs.
Finally, expect to create various modest systems, which individually will be unimpressive, but together will make a well-performing and hopefully diversified portfolio that will serve you as per your needs.
The idea of automated trading is a lure. That's what people want so that's what vendors offer. I'm not saying it's impossible but it is most certainly an elusive beast. Look around on futures.io (formerly BMT) and you will see what I mean. No need to look at other forums because the results are the same.
Now that I've said that, I will offer to you that you need to minimize the number of rules for entry and exit. The more rules you have the more likely you will be curve fitting your strategy. Also if using indicators, same thing. Use only 1 or 2 indicators at the most and I mean that literally. Resist the urge to add another indicator/filter to weed out some of the bad trades. You will paint yourself into a corner very quickly. Understand everything there is to know about the bar type you are using especially if it's an exotic bar type like renko, range, kagi, linebreak, etc. Even tick and volume bar types have some gotchas that you need to discover. I wish I could give you more detail but that is a tall order. Main point is if you see a result in back testing that seems too good to be true, then I guarantee that IT IS. Figure out why it's too good to be true. You might want to start a thread on futures.io (formerly BMT) documenting your efforts. That is a way of allowing us to help you along your path to the edge of the cliff so you can at least take in the scenery before you leap.
Regardless whether you ever develop a successful strategy or not, you will learn a tremendous amount about the markets. I make the assumption that you will continue down the path of developing a strategy no matter what anyone says. So, I strongly suggest that you attack this from 2 sides. Continue with your strategy development but spend at least equal time trading without indicators. You will learn a lot about yourself and your chosen market by actually trading it without indicators. Plenty of material on futures.io (formerly BMT) for that as well.
How about a fun challenge to those who believe they can create an automated strategy that is profitable? My intention is to start a discussion where algo traders can show off their strategies and learn something in the process.
This is the first time …
It contains user contributed trading systems that @sam028 runs on his Chicago servers and posts the results each month.
OK, if you'd read my other journal ( here at Big Mike's, you'll know (as of today) that I have attempted 2 TopStepTrader Combines in the past few months. I did not achieve the profit goal in either of them :faint:, but qualified for a rollover …
I feel your frustration. I used to feel the same way. Initially I was trying to build systems to trade very low time frames, like 5min or 1min or some arbitrary Renko/Range/Tick bar size. It's very hard to come up with profitable systems at this level because of commission and slippage, and most of it is noise. In my opinion there are a finite number of positive expectancy trading opportunities in any given day. Be OK with that and move on. I found better results trading higher time frames, like hourly, or daily. You could even code the lowly moving average cross system and make money (over time) with index ETFs, but it may take weeks or months for trades to last. I used to try to design systems that would make me a given daily income, because I wanted machines working for me over and over throughout the day, making money for me every day like a steady job income. I realized the market doesn't owe me a thing and all I could do was to take what the market would give me and be patient in between.
Examine and evaluate your expectations and reasons for doing what you are doing. Then, once you have a grasp on that, look for systems that exhibit the following characteristics.
- trade with the prevailing trend
- buy pullbacks in up-trends, sell up-moves in downtrends
- trade highs/lows of a given period, considering how those levels were reached, determine whether to play a breakout or to fade that level
- look for asymmetric risk/reward ratios in a 50% chance of success setup
- consider seasonality (day of week, time of day, month of the year, etc.) that your system works best in
- focus on ways to filter out losing trades, more than trying to get more winning trades, optimize for limited drawdowns and volatility instead of just absolute profit
And check out the books @moses recommended. This site has some good trading system ideas you can experiment with too.
thanks for the replies thus far, I will reply again once i've gone through them all more thoroughly / watched the videos attached and let you know how it worked for me.
You need to read up on what you've been given before asking new questions, because frankly what you are asking makes no sense. You should draw that conclusion as well once you read up on the research we've suggested.
