top of page
Capstone Trading

Perfect or Fake Trading System with Artificial Intelligence?


In this video, we discuss how to develop an incredible looking trading system with a straight-line equity curve that won't work in live trading. Some of these strategies fall under the disguise of AI. This is not an Artificial Intelligence based trading system.


Renko charts create amazing visualizations of the trend and could be useful in trading. In order for this to be useful in trading, it is critical to create an accurate backtest. Renko and Kase charts can be accurately backtested using our latest Money Management Algorithm update that will send signals from Renko or Kase charts to a time series chart.



In this video, we show how a "perfect" trading system can be developed using a Custom Renko chart with a Brick size of 5 and an offset of 1 with a simple moving average crossover that generates a straight-line equity curve with almost no drawdown on the E-mini Nasdaq futures.


It's always exciting to find these types of strategies at first but it isn't reality. If you are new to algorithmic trading, it is important to understand the accuracy of your backtest.


Two important points on developing trading systems in this video.

1.) It is important to accurately backtest a trading system. If a strategy starts to degrade or underperforms, the signals should still match the live trades with slippage and commission. Occasional data errors will cause trade differences. With custom charts such as Renko charts, depending on how it is coded, every trade changes once the chart is refreshed. Live automation in SIM mode shows a big loss instead of a big gain.


2.) It is important to develop realistic expectations for a trading system. Technology and trading platforms allow us to find straight line equity curve trading systems. We can deceive ourselves to think that we are being reasonable to believe that while the backtest is not completely realistic, that it should be partially realistic. It is possible that the backtest is completely false.


The most accurate way to backtest a trading system is on time series charts down to one-minute intervals.


Additionally, no strategy is perfect, and strategies degrade over time. Real trading system have choppy equity curves and drawdowns.


In conclusion, it is important to make sure the backtest will match future trading signals with slippage and commission. We don't want to see trade signals change when the chart is refreshed. With time series this happens sometimes if there is the occasional data error. With Renko charts, it happens on every trade if it isn't coded correctly.


Renko charts can still be useful when sending signals to a time series chart to accurately backtest. We demonstrate this in the video, showing more realistic results that are in line with the Live automation in SIM mode.

22 views0 comments

Comments


bottom of page