> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fincept.in/llms.txt
> Use this file to discover all available pages before exploring further.

# Forecast comparison

> Compare forecasting models on their losses: superior predictive ability (SPA) and reality check p-values and the StepM set of models that beat a benchmark, and the model confidence set of the best models, all with block-bootstrap inference.

Toolset `vol_comparison`: 2 tools.

| Tool | What it does | Notes |
| - | - | - |
| `vol_spa` | Tests whether any of several forecasting models beats a benchmark on their losses, with block-bootstrap inference that accounts for data snooping: Hansen's SPA (lower, consistent and upper p-values, critical values, the models significantly better), White's reality check, or Romano-Wolf StepM (the set of superior models at the given size). | |
| `vol_mcs` | The model confidence set of Hansen, Lunde and Nason: the models whose losses are not significantly worse than the best, at confidence 1 - size, found by sequential elimination with block-bootstrap p-values. | |

Inputs, limits and outputs are described in [Fincept Volatility](/guides/fincept-volatility). Full schemas: `fincept_describe_tool`.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.