Yes, it is.
You can create a similar model, with the same specifications, the same data, the same technical indicators, then train and backtest your model with Scikit-learn, to finally check the prediction similarities between your Python-coded model and your model on Swiss33. If the results are similar enough according to your threshold, you can be sure that the bot will only follow the indications of the model.
Swiss33 provides only the trading platform and is currently not offering support in machine learning model building.
Also, we recommand to practice with paper mode before jumping to live trading.
No. However, if you are worried about your account by sharing your API keys, be sure to disable all money transfer features and trade in paper mode.
Yes, these gaps are normally only visible for stock trading and can be due to several things:
Swiss33 supports 138 technical indicators. They appear at tick t as an indicator at time t-1 (i.e., for example, on 2022-01-01T10:20:00 the RSI displayed is that of 2022-01-01T10:19:00 if the selected granularity is "1 minute"). The default parameters are those of the TA-LIB library. To know these default parameters, you can refer to the documentation of this library: https://mrjbq7.github.io/ta-lib/doc_index.html. The possibility to modify the parameters of each indicator is currently under development. For more information on how each technical indicator is computed: tadoc.org.
For this you can create exactly the same model in Python with Scikit-Learn and on Swiss33, with the same technical indicators. You should therefore see similarities in the results. If you are still not convinced, here is the piece of code that allows us to get your preferences in terms of technical indicators:
Can't find your answers?Contact us