AdamMT5

Adam: Neural Network-Based Trading Robot

Adam is a trading robot developed based on my custom neural network, containing everything necessary for efficient operation. The latest update to version 1.21 is, hopefully, the final one for this application. I've made a significant effort to optimize the entire system as much as technically possible. Adam now has a slightly faster optimization process, but keep in mind that this is a highly complex system, so optimization on a large time frame can still be resource-intensive. Ultimately, it depends on your hardware. If you're using a regular personal computer, the optimization process will be demanding, but the same task becomes much easier with a dedicated workstation.

As I mentioned earlier, there’s no need to go too far back in history. You can train the neural network through the optimization process, but not for more than a year and a half from the current date. For the starting date, we can set June of last year, the current day as the end date, and for the forward/custom period, we can choose February or March of this year. While some might think this is a small dataset, believe me, for neural network-based robots, this is a large dataset. I've just provided a maximum range here—you can opt for a smaller time frame during the forward optimization process, covering just a few months back if needed.

Application Overview

Position Size is determined based on the input “Fixed Lot Size,” which denotes the fixed lot size. I don’t think this needs further explanation. Exit Rules such as Take Profit, Stop Loss, Trailing Stop, and Step are calculated based on the ATR indicator. The values are expressed in coefficients, as two independent ATR indicators are used as multipliers within the system. If you're wondering where Break Even is, don't worry—it's included within the system. By adjusting the Trailing Stop and Step options, you are, setting both Trailing Stop and Break Even simultaneously. The robot also provides standard functionality for placing Buy and Sell orders. Through the drop-down menu “Choose Trade Mode,” you can choose to execute either Buy orders only, Sell orders, or both, which is the default. 

The next option is “Order Type”. By clicking on the drop-down menu, you can select which types of orders the robot will work with. You can choose between Market orders, Limit orders, or Stop orders. Next is “Order Expiration in Bars”. This option applies only if you’ve selected pending orders in “Order Type” (Stop or Limit). It determines when pending orders will expire if they are not activated. Expiration is measured in bars. For example, if you’re using an H1 time frame and set the expiration to 1, this means that the order will expire after 60 minutes if not activated. In the M15 time frame, it expires after 15 minutes. I hope this is clear. Another option is “Max Trades per Day”. When this is set to 0, it means this input is not considered, in other words, it's inactive. This setting allows you to define the maximum number of trades the robot will open within a single day. If you set it to 1, the robot will open only one trade today.

Magic Number: This is an arbitrary number that can be any integer. If you’re using the robot on only one chart and with no other robots, you can leave the default value, like “5003”. However, if you’re using the robot on multiple charts, time frames, or in combination with other robots on different currency pairs, this number must be unique. The magic number is essential as it serves as a unique identifier for the robot's operation. Comment: This input allows you to write any note you think will be useful later. It’s an optional field.Exit on Friday determines the time (hour and minute) when all open trades should be closed on Friday. For example, if you set 20:00, the robot will close all positions at 8 PM on Friday, regardless of whether they are in profit or loss.

Neural Network Inputs

The Adam Neural Network includes nine input parameters. Let’s go through them one by one.

• Epoch: This defines the number of epochs required to train the neural network during robot initialization. For instance, if you set this to 1000, the robot will complete the process in 1000 epochs or 11000 ticks.
• Session: This is a multiplier for the number of epochs. If you set it to 2, the robot will train the neural network for 2000 epochs (2 * 1000).
• Threshold: This sets the weight for the neural network. Use appropriate weights, as the system compares the total sum of all significant parameters to this threshold. If the sum is less than the threshold, the system won’t operate (e.g., it won’t open market orders).
• Global Factor: This is a negative value with a specific purpose within the system. The recommended range is between -1.0 and -10, with -2.5 being a standard and widely used value.

• Platform type: MT5
• Account type: Any
• Currency Pairs trading 
•  Check section  >> What's New <<
• If you buy now, you reserve the right to own new upgraded versions of the robot, which will appear at specific time intervals continuously.