> ## 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.

# Benchmarks and intermittent demand

> Benchmark forecasters every model should beat (naive, seasonal naive, random walk with drift, historic and window averages, conformal seasonal pool, constant) and intermittent-demand methods for series with many zero periods (Croston classic, optimized and SBA, ADIDA, IMAPA, TSB).

Toolset `tsforecast_benchmarks`: 2 tools.

| Tool | What it does | Notes |
| - | - | - |
| `tsforecast_baselines` | Forecasts one series (y) or every series of a panel (data) with benchmark methods: naive (last value), seasonal naive, random walk with drift, historic average, window and seasonal window averages, the conformal seasonal pool (seasonal naive with sample-based intervals) and constant, zero or missing forecasts (fallbacks). | |
| `tsforecast_intermittent` | Forecasts series with many zero periods (spare parts, slow-moving stock, rainfall, sporadic trades) with intermittent-demand methods: Croston classic, optimized and Syntetos-Boylan (SBA), ADIDA (temporal aggregation), IMAPA (multiple aggregation) and TSB (demand probability times size, handles obsolescence). | |

Inputs, limits and outputs are described in [Fincept TS Forecast](/guides/fincept-ts-forecast). Full schemas: `fincept_describe_tool`.


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