Телеграм канал «mt QUANT - Advanced Options Analytics / Backtesting Platform»

A new update of mtQuant version 1.0.2.118 was rolled out. When you restart the mtQuant, click ‘Yes’ to update the application.

1. Old Html Reports Updated:

a) Many new summary sections and useful graphs were added to the report.

b) In the previous update we had optimized the report and in the case of larger results the days will load as the user scrolls down.

However many users requested that due to this Cntrl+F not working and they had to scroll the complete report down which was taking time. For this purpose we have added a new button “Load Full report” near the existing collapse all days button.

For any suggestions or issue reporting for these functionality, you may DM @mtquantsup.

Update version you can see at the left bottom of Login Screen.

A new update of mtQuant version 1.0.2.117 was rolled out. When you restart the mtQuant, click ‘Yes’ to update the application.

1. Old Reports Updated:

a) Now the old reports would open much faster and be able to open the larger results then the previous versions of reports.

b) Dark theme added in the old html reports.

c) few new useful charts were added.

2. Fix: When you were importing the positional portfolio where “Take ReEntry only on the Leg Execution Day” was untick, then on import it was getting ticked. This was only happening only with positional portfolios which were imported. It's fixed now.

For any suggestions or issue reporting for these functionality, you may DM @mtquantsup.

Update version you can see at the left bottom of Login Screen.

A new update of mtQuant version 1.0.2.115 was rolled out. When you restart the mtQuant, click ‘Yes’ to update the application.


1. MCP Update: MCP Server updated with new functions of Profit Locking and Trailing at the Portfolio Level.

Also added SL Trailing at the Portfolio Level through MCP.

2. In the case of Positional Portfolios, If the end time was specified as 00:00:00, then ReEntry / ReExecute was not working properly. End time in the case of positional portfolio was forced since the last update only. It is fixed now.

For any suggestions or issue reporting for these functionality, you may DM @mtquantsup.

Update version you can see at the left bottom of Login Screen.

If anyone still facing any issues the can follow the above steps. Wrong detection was removed.

!!Attention!!,

Some Users are complaining that mtQuant is not working. It is happening due to the defender and we had already raised an issue to Microsoft. However this may take time.

Till then you can download the latest version from below Url:
https://cdn.jsdelivr.net/gh/mtquant/mt/mtQuant.rar

Extract the File on Desktop:
Users for whom mtQuant is not working can download and extract the file on desktop.
Password: 1234

After extracting the file, cut from the desktop and then paste into C:\Program Files (x86)\mtQuant folder and replace the existing file.

Thanks

A new update of mtQuant version 1.0.2.112 was rolled out. When you restart the mtQuant, click ‘Yes’ to update the application.

CLAUDE / OPENCODE / OTHERS DESKTOP OR CLI BASED LOCAL AI INTEGRATION WITH MTQUANT (BETA)

1. MCP Update: MCP Server updated with new functions for bulk management of Portfolios and Legs.

AI can choose between the available functions at his discretion, however if AI chooses to use these functions for bulk, then Token usage can be reduced.

2. Local AI: There were apprehensions from some users that if they work completely from AI then AI may have their complete strategy.

First of all the usage of MCP Server is completely optional and at user discretion. You can keep using the application in the traditional way as well.

Second, we have added a basic Local AI Functionality by using a small AI model which runs within the computer of the user. No data or prompt leaves the user's computer.

You can click: Options Trading > Local AI Assistant to start the Local AI. For the first time, it will download the necessary files for about 2.5GB on the first click.

Do not use over VPS or virtual, it will kill the complete performance.

However there are multiple aspects to consider :-

a) Minimum Hardware Requirements: AI needs very good hardware requirements, specifically a Good Nvidia Graphics Card. However mtQuant Local AI can work with a decent CPU As well, but speed might be slower in comparison to GPU.

This functionality of Local AI will work on any Windows 10+ computer having minimum 8GB of RAM, however it would be very slow. Preferred minimum RAM is 16GB.

It would perform well if users are having a dedicated 8GB Nvidia Graphics Card. However it can run without Graphics Card.

b) Local AI has limitations in terms of intelligence and reasoning.

c) You cannot expect Local AI to be comparable with the commercial AI. Commercial AI are far superior and have very dedicated hardwares.

d) However you can use Local AI for port creation, modification or starting a backtest etc. Results you can analyse over any commercial AI. In this way your whole portfolio structure was never disclosed to commercial AI.

e) To work properly with Local AI, Treat Local AI as a child and keep the prompts simple and explainable in simple English. Try to provide as much as clear information in the prompt.

f) Perform one task in a Chat, for a new task start a new chat. Suppose you performed some work on a short straddle and then wanted to create an iron condor, then better to create a new chat.

g) If Local AI is unable to understand any Task then try to break that in simple steps.
For a very basic example: Suppose prompt was "Backtest all portfolios starting with ABC from 01 Sep 2026 till date" and AI hallucinated then i can say like,
“list all the portfolios starting with ABC” and then the second prompt could be “Now Backtest all these portfolios from 01 Sep 2026 till date”.

h) If you want to contribute to the betterment of local AI, With every response from Local AI, there are thumbs up / Down options, just click those and send us the logs of Local AI in telegram chat. Rest assured about the logs, we would be interested in the function calls not the parameters so we can adjust accordingly.

i) It is fully local and all information / data will reside within the user's computer.


For any suggestions or issue reporting for these functionality, you may DM @mtquantsup.

Update version you can see at the left bottom of Login Screen.

This is the reference guide for the mtScript. Soon we will add an editor to create / modify mtScript within mtQuant.

However if you’re already using any AI over mtQuant MCP, then MCP will automatically provide the necessary details to the AI and you only need to send the prompt to the AI.

No need to send this document to the AI.