Forum

A zero-shot model (IBM Granite TTM-R2) in MT5 via ONNX: export pitfall, 0.86 ms per forecast, what it can forecast

I wanted to know whether one of the new pre-trained time-series models can run inside MetaTrader 5 itself, without Python, DLLs or WebRequest, so that it also works in the Strategy Tester. It can. These are my notes on IBM's Granite TTM-R2 (Tiny Time Mixers, Apache-2.0), with the numbers I measured

Broker GMT offset in the Strategy Tester: TimeGMT() is server time there, so estimate it from the weekly close (code)

Any EA that works with fixed session times (Tokyo fixing, London open, New York close) needs the broker's GMT offset, the difference between server time and UTC. Live, TimeTradeServer() - TimeGMT() gives it. In the Strategy Tester it does not: the documentation says that during testing TimeGMT() is

Japanese bank holidays and gotobi days in MQL5 without a calendar file (code)

If an EA needs Japanese bank holidays, for example to trade the Tokyo fixing on gotobi days or to stay out of thin Tokyo mornings, there is no need for a calendar file or WebRequest: the holidays follow fixed rules and can be computed. This is the code I use. It covers the equinox days for

New “sk-proj” OpenAI Key Truncated? WebRequest Always Returns “You didn’t provide an API key.”

Hello everyone, I’m attempting to integrate ChatGPT (gpt-3.5-turbo) into MetaTrader 5 using WebRequest() in MQL5. When I call https://api.openai.com/v1/chat/completions , I consistently get a 401 response telling me: "You didn’t provide an API key. You need to provide your API key in an