Hello and Welcome to Amxsys!
Amxsys stands for “Advanced Multi Asset System”, because our AI based trading system tracks multiple markets worldwide to select the best looking opportunities for trading. The constant goal is to exit from unclear setups and participate in clear and compelling setups (based on the last 20 years of market analysis, which has been used for creating the AI based trading system).
Collectively, we spend thousands of hours researching and analyzing the global markets, searching for good investing and trading opportunities across the various markets:
- Equity Indices (S&P500, Nasdaq100, Nikkei225, Nifty50, FTSE100, DAX, Hang Seng — and their ETFs like SPY, QQQ, SPXL, TQQQ, XBI, etc )
- Crypto (Bitcoin, Ethereum, Litecoin, Dogecoin, Shib, Cardano, Polygon, etc)
- Commodities (Crude Oil, Gold, Silver, Copper, Wheat, Coffee, Corn, etc)
Amxsys is new system (was under development for last 4 years), and now we are happy to make it live and share with all traders and investors worldwide.
Our AI based Trading System tracks and analyzes key markets worldwide, using our trading systems, to share profitable trading signals with our Members/Subscribers.
Our Trading Systems are the core of our work in financial markets. Without a trading system, we can’t survive in any market even for a week. In fact, there can be no clear starting point without a Trading System.
We believe in technical analysis, and it works best on fundamentally strong assets, like S&P500 and Nasdaq100 indices.
In our 20+ years experience in financial markets, we have seen countless times that the markets have some “reliable setups” and many “unreliable setups”. Waiting for the “reliable setups” needs patience and discipline, and it gets rewarded with profits.
We are constantly improving our Trading Systems so that we can perform better in each key market that we track. Not all experiments give results, but even if one experiment gives good result, then it can be very profitable.
“To invent you have to experiment, and if you know in advance that it’s going to work, it’s not an experiment.” — Jeff Bezos
“Cornerstone” is the name of our Foundation Trading System that has been developed based on our study of last 20-50 years of market data, on various timeframes from hourly to yearly, in each market/index that we cover. This system keeps evolving as it learns from new market data every month, and we make ongoing effort to improve and test the system with both new and old market data.
Following are the key design goals of this system:
-
Give Buy signal when market starts meaningful uptrend.
-
Give Buy signal when a market correction has completed/exhausted, to enable lower level buying.
-
Give Sell signal when market starts meaningful downtrend.
-
Give Sell signal when a market rally has completed/exhausted, to enable near-peak selling, and exit of existing positions.
-
Give near term, medium term, and long term Price Targets in all the markets we track.
-
Give Warning about upcoming volatility.
-
Give Warning about significant divergence in usually correlated assets/markets.
-
Give Alert regarding unusual investor behavior in any of the markets we track.
-
Give Alert regarding any rare price action in any of the markets we track.
-
Give Shutdown signal to close all open trades in case of rare negative event of likely high impact.
The above design goals have been achieved to a good extent, and they were fully tested during the market rally from Oct 2019 to Feb 2020, and during the sharp correction in Feb 2020 and March 2020. Of course, there is always enough scope for improvement, because no Trading System can be perfect. We will keep improving!
Cornerstone foundation system includes 4 Trading Systems, each operating with a different methodology, in different time frames (hourly, daily, weekly, monthly).
- For the two main equity indices (S&P500, Nasdaq100) have all 4 Trading Systems analyzing them in their respective ways – these are the most actively tracked and trained indices in Amxsys, and you can trade ETFs (SPY, SPXL, QQQ, TQQQ), Futures (ES, NQ) and SPY options with the trading signals published by Amxsys.
- For Bitcoin, Hang Seng index of Hong Kong, and FTSE Index of UK market, we are using daily and weekly timeframes (but not hourly).
- We are not yet offering signals on Agri Commodity Futures (like Corn, Wheat, Soybean) but it’s on our product roadmap, and to train our Amxsys AI system for these agri commodities on daily or weekly timeframe.
The output of these Trading Systems is Buy/Sell signals on various time frames.
- Short Term (Alpha) signals are based on Hourly charts, and they can help traders who are trading on timeframes of 1 to 5 days, including Intraday Traders. A vast majority of traders will come in this category, and these signal can be very profitable if the market is moving in a trend.
- Near Term (Delta) signals are based on Daily charts, and they can help traders who are trading on timeframes of 1 to 30 days. Many positional traders will come in this category, and such signals can capture a good part of a market move.
- Medium Term (Gamma) signals are based on Weekly charts, and they can help both traders and investors who are working with timeframes of 1 to 12 months. Many traders and investors will come in this category, and such signals can capture a good part of a market move and a market cycle itself.
- Long Term (Omega) signals are based on Monthly charts, and they can help traders and investors who are working with timeframes of 1 to 3 years. Many investors will come in this category, and such signals can capture a significant part of an entire market cycle, including the major market trends that create big gains.
- For example, we have been asking all investors in our network to invest in the Nasdaq ETF (QQQ) since year 2010, when it was $50 per share, and now it hit $400 per share in 2 Such large moves can been seen with long term analysis. We are in the “Tech Century” and the Nasdaq ETF (QQQ) has much higher targets, and all investors can continue to accumulate it ongoing basis, and even more so during 5-10% corrections.
The Trading Systems use a combination of linear and non-linear analysis.
And they are currently being used in real-money accounts of various sizes.
Because we focus on limited number of markets/indices, we can study them in detail, and try to do better analysis with each passing quarter, and enable our Trading Systems to give better trading signals to our customers. Please see the Dashboard page to learn more.
Using Neural Networks for creating AI based Trading Systems – part 1
Neural networks are being used for creating trading systems. Neural networks are powerful machine learning models that can be trained to recognize complex patterns in data. In the case of trading systems, neural networks can be trained to recognize patterns in historical market data and use those patterns to make predictions about future market movements.
Here are some ways neural networks are being used for creating trading systems:
- Prediction of asset prices: Neural networks can be trained to predict future asset prices based on historical price data and other relevant features such as economic indicators, news articles, and social media sentiment.
- Portfolio optimization: Neural networks can be used to optimize portfolios by selecting the best mix of assets based on historical performance and risk factors.
- Trading signal generation: Neural networks can be used to generate trading signals based on patterns in historical market data. For example, a neural network might be trained to recognize a particular pattern in stock price movements that has historically been associated with a high probability of a price increase in the near future.
- Risk management: Neural networks can be used to manage risk by predicting the likelihood of extreme market events, such as crashes or spikes, and adjusting portfolio allocations accordingly.
Overall, neural networks offer a powerful tool for creating trading systems that can adapt to changing market conditions and make better-informed trading decisions. However, it’s important to note that building effective trading systems using neural networks requires extensive data preparation, model tuning, and rigorous testing to ensure the reliability and accuracy of the system.
Using Neural Networks for creating AI based Trading Systems – part 2
Following are the steps for creating a trading system using neural networks that gives buy/sell signals based on a combination of technical indicators:
- Define the problem: The first step is to define the problem you want to solve. In this case, you want to create a trading system that generates buy/sell signals based on a combination of technical indicators.
- Choose the technical indicators: Select the suitable technical indicators you want to use in your trading system. Some of the most commonly used indicators include Simple Moving Average (SMA), Exponential Moving Average (SMA, )Relative Strength Index (RSI), Commodity Channel Index (CCI), MACD, and Bollinger Bands.
- Collect and preprocess data: Collect historical data for the securities you want to trade and preprocess the data to prepare it for use in your neural network model. This may involve cleaning the data, normalizing the data, and splitting the data into training and testing sets.
- Train the neural network model: Use the preprocessed data to train the neural network model to recognize patterns in the data and generate buy/sell signals based on the technical indicators you have selected.
- Test and validate the Model: Test and validate the neural network model to ensure that it is generating accurate buy/sell signals. Use a separate validation set to test the model’s performance and adjust the model as needed.
- Implement the Trading system: Once you are satisfied with the performance of the neural network model, implement the trading system by using the model to generate buy/sell signals in real-time.
- Monitor and Refine the Trading system: Monitor the performance of the trading system over time and refine the system as needed. This may involve adjusting the technical indicators used in the model, optimizing the model parameters, or adding new features to the system.
Overall, creating a trading system using neural networks requires extensive data preparation, model training, and testing to ensure the reliability and accuracy of the system. It is important to have a solid understanding of the underlying technical indicators and market conditions, as well as the ability to continually monitor and refine the system to adapt to changing market conditions.Without ongoing/regular training, the neural network can become obsolete. Therefore, constant learning is needed by the system for best results.
How do we get better results from our AI system using Neural Networks?
There are several ways to improve the predictive capability of a neural network:
- Increase the amount of training data: The more data that the neural network has to learn from, the better its predictive capability will be. By increasing the amount of training data, you can help the neural network identify more patterns in the data and improve its accuracy.
- Improve the quality of the training data: Ensuring the quality of the training data is also important for improving the predictive capability of a neural network. This can involve removing noisy or irrelevant data points, correcting errors, and filling in missing values.
- Optimize the neural network architecture: The architecture of the neural network, including the number of layers and nodes, can have a significant impact on its predictive capability. By experimenting with different architectures and parameters, you can fine-tune the neural network to achieve optimal performance.
- Regularly update the neural network: Markets and trading conditions are constantly changing, so it is important to update the neural network regularly to ensure that it remains effective. This can involve retraining the model on new data or adjusting its parameters to better reflect current market conditions.
- Combine multiple models: Combining multiple neural network models, each trained on different subsets of data or with different architectures, can help improve the predictive capability of the overall system. This can be achieved through techniques such as ensemble learning or stacking.
- Regularly evaluate the performance of the neural network: Evaluating the performance of the neural network regularly is crucial to identify areas for improvement and ensure that the system remains effective over time. This can involve measuring metrics such as accuracy, precision, and recall, and using techniques such as cross-validation to test the model’s performance on new data.