After several months of development, we're excited to release StrongholdTrainer 1.0! 🎉
Here's a rundown of some of the main features in this first full release:
- Stronghold Runs: The central feature of the mod. Timed stronghold runs (both IGT and RT) and functionality to review these runs.
- Timing: Runs are timed from leaving the starter staircase to entering the portal room. Features both a live timer and an RT time only visible on run completion.
- Statistics: Utilizing a machine learning model developed by Geosquare, Matthew Bolan, and XeroOl, the mod will automatically analyze your run and display chess website-style statistics to help you understand if and where you made mistakes in stronghold runs. You can also compare yourself to Feinberg ('s average room times)!
- Custom Inventory: Allows you to automatically reset and load a custom inventory (save your current inventory using
/inventory save) every time you go to a new stronghold. - Player Trace: You can see your path through the stronghold.
- Models and Hints: Another central feature is the ability to load stronghold navigation machine learning models, as well as hints that are powered by those models. More details will be documented in the repository.
- ML: Currently, there are several pre-loaded models including Geo's original model, and more advanced models that support backtracking developed by Flourish and neoprene1337.
- Hints: Hints are outlines visible around stronghold doors, as well as labels for rooms. These doors display the output of whatever the loaded model is, as well as the perfect choice and the model choice of exit. Only some of the models support backtracking.
Feel free to report issues on the issue tracker, but make sure that the issue hasn't already been reported.