Why stockpredictions.net

Owing to the fact that I was trading with emotions, I often lose money by making decisions based on my feelings at that moment. Too tight stop losses or thinking that it will soon get better cost me a lot of money. Of course, there is a piece of advice on the internet, but you have to look for it. It's not coming right into your mailbox, or at least not free. 

Trailing Stops

For better trading, I started using trailing stops. A trailing stop is a stop price set at a defined percentage below the current market price of the position. As the market price rises, the stop price also rises. If the market price falls, the stop price does not change. But then the question arises, what is the distance to keep? 2 x ATR or 1.5, and is this also true in an uptrend? And of course, is this the best value for every stock different? As a programmer, I created a web application which calculates the best values for each stock by running all the possible trailing stop distances (%) values against the historical rates of the stock. And the best one is used in a control run, that's a timeframe later than the learning timeframe. The prediction rate in the control run is used as the quality standard of the trained model.

Uptrend and downtrend

For better predictions, we need to know if the stock is in an uptrend or downtrend. We have to adjust our trailing stops by the current trend to finetune the stop losses and set the sell orders. We also need to know where to set our buy orders of course. Using the same training model as above, we calculate some Technical Analysis indicators and let the weight of each indicator be calculated bij machine learning. So each stock has different weights of the moving average or dmi by determining the bullish or bearish trend. In this scenario, false signals of TA indicators are neutralized by the other indicators, and bad indicators get less weight in the overall advice. The main difference is that we use the trailing stop strategy, so we always wait until the price is doing what we think it must do. Sometimes we are a little bit later with buying, but the risk is lower. 


Now, we are bringing the calculated trends and trailing stop strategies to practice. How much risk are we willing to take, and when do we adjust stop-loss orders or make it to a sell order? This is why I created the bots. Of course, I have a connection to a real broker, but it's safer to run bots against a virtual broker. You can get updates about the buy and sell action of the robot by making an account and selecting your favorite bot.