Creating new strategies involves:
1. The idea and type of strategy: Break out, Mean reversion, Trend following + a tendency of the market to be exploited
2. Direction: long or short
3. Universe: ETFs, stocks, forex ...
4. Instrum. selection: parameters, indicators, fundamentals
I am thinking of doing a little systematic trading workshop (for free) and create a strategy showing „Real Test“ in action. I would randomly pick 5-10 people. When you are interested please like and retweet! PS: No coding skills required.
I have developed a market filter for my trading and backtested it. I am only allowed to trade when it is green. When it is red I am only allowed to trade on the short side. This puts the odds into my favor.
Today I created one of my best strategies: It is a mean reversion long, uses medium volatility and a variable limit target. The KPIs are incredible: 25,7% CAGR, 8,3% DD and only 7% avg. Exposure + very robust among 100s of variants.
If I could only trade one strategy for the rest of my life, it would be mean reversion short. Why? It works in every market environment and gives top results. There are only a few periods where it has problems e.g. in strong uptrends.
Free sample strategy, all rules disclosed. Read my comments (grey colour). CAGR is 10% but it only uses 12% of the money, so you can combine it with many other strategies. Don't look at CAGR on a single strategy level, look at CAGR/Average exposure, which is 87% in this case!
Because everybody is talking about it, I built a strategy: Buying fresh ATHs. Looks very promising. Will do more robustness testing. Idea to implementation: <20 minutes.
Performance YTD, still struggeling to significantly outperform SP500. I was ahead in the first months, but then had a 16% DD (NDX Momo sucked, Semis weakness), have to reduce vola.
Outperformance continues, made new all-time-highs yesterday and the day before, despite the market's weakness. If we can rebound in one of the next days, I could get a really big move forward. Have to stay humble though.
Evaluating "Real Test" for my systematic trading. It is great! It is super fast. It took me only minutes to get into it, you look at the examples and it is so easy to learn (if you have an understanding of algos and trading). You can customize everything. Some outputs:
A strategy looks smooth on a 20+ years equity curve (MOMO with 24% CAGR), but if you zoom into a year or two, you see the fluctuations much more clearly, but you only feel it when trading it live.
Momentum and TF strategies are often a test for yourself and nothing for the impatient. You stay in a trade for months and then get stopped out with a loss, stocks gapping down, you buy at the top, etc. Galary of horror, momentum strategy (trades end of month only):
I did reach a new ATH yesterday, ending a 40-day DD. Algos made big progress and relative outperformance the last few days. The next DD begins shortly after a new high is reached.
Month end strategy: Buy ten SP500 stocks that are temporary in a low volatile state and in an uptrend, but only when SPY is in an uptrend. Exit when stock or spy entering a downtrend.
If you aren't really keen on programming, stop wasting your time with python. Go for a proven software like RealTest and quality data like Norgate. It took me 9 months in python to get my strategies done. With Realtest? Less than a week, including order management.
Don't buy extended stocks. I did over a million trades in SP500 and closed them after 5 days. It is more favorable to buy when the stock is below 10SMA. The chart shows the distance to SMA10 in %. Each bar has >69,000 trades inside.
We will try to cover the whole process:
✅Idea generation
✅Indicator selection
✅Optimizing for robustness
✅DD reduction
✅KPIs
✅...
Please like, retweet and subscribe.
I am thinking of doing a little systematic trading workshop (for free) and create a strategy showing „Real Test“ in action. I would randomly pick 5-10 people. When you are interested please like and retweet! PS: No coding skills required.
New shorting strategy. I have no real clue why it works, but is is very stable and robust. If it holds through 2020-2023 I will trade it. Have to do further testing.
Some entires for mean reversion:
🔵Extreme RSI readings
🔵Cumulative/Avg. RSI readings
🔵Multiple lower/higher lows/closes in a row
🔵MA stretch
🔵Drop in MA readings
New experiment: I will try walk forward optimization and trading it. Train the algo with Data from 36 month, trade it for 3m, re-train it with the last 36m, etc. It is a very simple strategy, only two parameters. 1st training period was 2000-2002.
Merry Christmas to all of you! 🎄🎁
I hope you are having a wonderful day with your loved ones and enjoying the festive spirit.
Thank you for your thoughts and kindness throughout this year!
Working on a Weinstein trailing stop, I like a lot: when a stock moves to (1) above the previous swing high (2) you can move your stop from (3) to (4) the current swing low.
In our groundbreaking study with
@BearBullTraders
, we unveiled the remarkable success of the 5-min ORB strategy on QQQ and TQQQ, showcasing consistent profits from 2016 to February 2023. 📈
Post-publication, this strategy has continued to yield robust results, collecting more
Thrilled to hit 3,000 followers 🎉🚀 Thank you all for joining me on this journey through the world of systematic trading. Let's keep exploring, learning, and growing together!
Don't buy the beginners course for systematic trading launched recently. It is not worth the money, even if it looks cheap. There is a real test course for free offering much more insights and content:
Testing a Bollinger Band strategy for $SPY, getting almost 5x returns vs. B&H. Trading SPY is a good start for systematic trading, as risk is limited and liquidity is very high. Can trade it with Alpaca and Tradestation at no costs.
Most People can't diet, eat less, or drink less. They can't exercise regularly, so why should they be able to execute a strategy consistently? It's just not human nature.
Still developing an all weather portfolio, making good progress. I had a market filter and could remove it without a much bigger DD. The version without a filter looks pretty decent now! Max. DD% is increased by 8%. Which version would you choose?
I switched to systematic trading because I am really bad in doing things consistently and following soft/weak rules, I am good at analytics, data, optimization and coding. Know your strengths and weaknesses.
To create a robust trading strategy, follow these steps:
1. Define clear objectives: Establish your goals, such as risk-adjusted returns, drawdowns, or trade frequency. This will guide your strategy development process.
Ridiculous fact: Best time in Europe to buy US ETFs is not during US opening hours, it is around noon. Interestingly 14.00-16.00 is the most expensive time to buy (in terms of spreads). Xetra is providing nice stats:
Heureka! Couldn't sleep well this night and had an idea what I wanted to test this morning. This is a monthly momentum strategy on NDX stocks. Look at the DD and smooth EQC! They key for my strategies is about combining elements I read about or look at different ways to use it.
Had my first look-ahead-bias moment. The curve already looked too good to be true. When I wanted to translate it into python, I recognized, I had looked at the next bar. 🤷♂️
I would like to join forces. 💪 I have a strategy that has clearly an edge, but I would like to improve it. It is a break out strategy. You can see the top10 worst and best performance of 300 test runs below. How would you proceed? Will select 2 or 3 people with good answers.
He took the current(!) S&P 500 constiuents and ran a backtest against it, a typical beginners fault. I am not getting the exact same figures bc of some ranking difference, I guess, but you see very similar figures.
@InvestmentTalkk
You have to delete this, this is fooling people.
Here's a seemingly stupid investment strategy.
Buy the top 10 S&P 500 stocks from the past year, equally weighted, and replace them annually.
Plot twist: It actually slaps.
From 2013 to today, it would have generated a 37% CAGR relative to the S&P 500's 12% CAGR.
In my new job, I have to hand out reports (financial figures) to departement heads. I automated it using VBA and chatgpt and saved about 4-5% of my monthly working time. 🥳
Most ETF rotational (eg all-weather-strategies) aren't robust bc rotational dates are either favourable or not. Covid crash was favourable for end-of-month rebalancing (chart 1). If you introduce a different rebalancing week (every 4 weeks check), you get worse results (chart 2).
In my former job I had to close down a company in Germany. Couldn’t finalize it because there were tax issues, went to court. This is more than 10years ago and the case isn’t settled (will have to pay interest for that time). 10 years! Germany is a becoming a failed state.
I am not good at doing the same things over and over again, I get bored and deviate from rules. That is my personality. I am really good at analytics, algos, data crunching, etc. That was the reason why I completely switched to systematic trading and left discrectionary trading.
Momentum trading has one big advantage, it does not care about pattern, break outs, etc. Just buying the strongest stocks is a big advantage in a rally like this when stocks are already extended and going higher.
Thread. I will show you, why it is so hard to trade during bear markets and not really rewarding: Imagine being a swing trader running a 2.5:1 reward/risk system with a 40% batting average. You are making good profits. 1/3