This strategy trades based on the trading signals generated by the Faytterro Estimator. The Faytterro Estimator is an indicator that judges trends by calculating the convergence and divergence rate of prices. This strategy combines the trading signals of the Faytterro Estimator and some additional conditions to generate long and short signals of different sizes at ideal points.
## Strategy Logic
The core of this strategy is the Faytterro Estimator. Its calculation is: first calculate the convergence and divergence rate (CR) of prices, then construct a quadratic function, which can reflect the shape of the CR curve by setting different coefficients. By observing the inflection points of the quadratic curve, it judges the change of price trends.
Specifically, the strategy first calculates the CR of prices. Then it constructs an array dizi of length 2*len, and fills it with quadratic function values sequentially. The coefficients of the quadratic function reflect the value of CR. After that, by observing the two values at index len+1+5 and len+1+4, it determines whether the quadratic function has an inflection point. If there is an inflection point, it generates buy or sell signals.
On this basis, the strategy also sets some additional conditions, such as setting the minimum distance between price breakthroughs to avoid frequent trading, generating signals of different sizes, etc. These conditions are used to filter out some undesirable trading points.
## Advantage Analysis
This strategy has the following advantages:
1. Using the Faytterro Estimator to judge trends, which is sensitive to price fluctuations and can capture trend changes early.
2. Constructing a quadratic function to reflect the shape of the CR curve and finding inflection points directly and effectively.
3. Generating signals of different sizes allows pyramid-trading at ideal points, increasing profit potential.
4. Increasing the minimum distance setting effectively filters signals and avoids ineffective frequent trading.
5. Many adjustable parameters can be optimized for different trading products, improving adaptability.
6. The strategy logic is clear and easy to understand, and the code is highly readable, making it easy to learn from.
## Risk Analysis
There are also some risks to note for this strategy:
1. Faytterro Estimator has the risk of curve fitting, and may underperform in some trading products.
2. Judging solely based on inflection points of the quadratic curve may be too crude, leading to misjudgements.
3. Frequent pyramid trading increases the cost of commissions.
4. A large number of adjustable parameters increases the difficulty of optimization.
5. It cannot effectively deal with misjudgements in period of price oscillation.
6. Lack of stop loss mechanism may lead to greater losses.
The corresponding solutions are:
1. Optimize parameters for different products to improve robustness.
2. Add other indicators for filtration to avoid misjudgements relying solely on inflection points.
3. Set proper stop loss to control single loss.
4. Use big data methods to auto-optimize parameters.
5. Add oscillation identification to avoid trading in oscillating periods.
6. Set reasonable stop loss logic.
## Optimization Directions
The optimization directions include:
1. Add stop loss logic to control single loss, such as trailing stop loss or time stop loss.
2. Add other indicators to avoid misjudgements relying solely on Faytterro Estimator. For example, combining with MACD, KDJ etc.
3. Add confirmation mechanisms to avoid being stopped out by short-term pullbacks. Consider re-entry confirmation.
4. Optimize adjustable parameters for different products using genetic algorithms, Bayesian optimization etc.
5. Identify oscillating markets using ATR, DMI etc and avoid trading during oscillation.
6. Optimize pyramid logic to prevent chasing trends. For example, dynamically adjust pyramiding positions based on trend strength.
7. Test parameters on different timeframes to find the optimal timeframe.
## Conclusion
This strategy makes decisions based on the trading signals of Faytterro Estimator, and adds logic judgments and different sized entry signals on top of it to form a trend-following strategy with pyramid characteristics. The strategy is intuitive and easy to understand, with strong trend catching capabilities. But it also has problems like indicator misjudgements, no stop loss, difficulty in parameter optimization. Future optimizations include adding filtration mechanisms, stop loss logic, parameter optimization etc to improve robustness and adaptability. Overall, this strategy provides a way of using indicators to judge trend changes that is worth learning from.
buy = array.get(dizi,len+1+5)>array.get(dizi,len+1+4) and array.get(dizi,len+1+5)<cr[len]
sell = array.get(dizi,len+1+5)<array.get(dizi,len+1+4) and array.get(dizi,len+1+5)>cr[len]
bb=buy? hlc3 : na
ss=sell? hlc3 : na
sbuy= buy and close<(close[ta.barssince(buy or sell)])[1]-len4 and close<ta.highest(fixnan(ss),len2)-len3*3
ssell= sell and close>(close[ta.barssince(buy or sell)])[1]+len4 and close>ta.lowest(fixnan(bb),len2)+len3*3
buy:= buy and close<(close[ta.barssince(buy or sell)])[1]-len4 and close<ta.highest(fixnan(ss),len2)-len3 //and close>ta.highest(fixnan(ss),len2)-len3*3
sell:= sell and close>(close[ta.barssince(buy or sell)])[1]+len4 and close>ta.lowest(fixnan(bb),len2)+len3 //and close<ta.lowest(fixnan(bb),len2)+len3*3
