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

# Mental model

> A short guide to OIPD's four main objects and the fit-then-query workflow.

OIPD turns option prices into two kinds of objects: volatility objects and probability objects.

Volatility comes first. Probability is derived from fitted volatility.

<Tip>
  Only need probabilities? You can skip the volatility objects and use `from_chain()` directly: `ProbCurve.from_chain(chain, market)` for one expiry, or `ProbSurface.from_chain(chain, market)` for many expiries. See [Probability layer](/concepts/probability-layer).
</Tip>

## Core objects

| Use case                       | One expiry  | Many expiries |
| ------------------------------ | ----------- | ------------- |
| **Implied volatility**         | `VolCurve`  | `VolSurface`  |
| **Market-implied probability** | `ProbCurve` | `ProbSurface` |

An **expiry** is the date when an option contract ends.

A **curve** is one expiry. A **surface** is many expiries linked through time.

## Workflow

Most work follows three steps:

1. Create an object, such as `VolCurve()` or `VolSurface()`.
2. Fit it to an options chain and `MarketInputs`.
3. Query it for implied vol, prices, Greeks, or probabilities.

```python theme={null}
from oipd import VolCurve

vol = VolCurve().fit(chain, market)
prob = vol.implied_distribution()

print(prob.prob_below(100))
```

This is similar to fitting an econometric model: choose the model, estimate it, then ask it questions.

## Object choice

Use `VolCurve` when you have one expiry and want implied volatilities, option prices, Greeks, or a fitted smile plot.

Use `VolSurface` when you have several expiries and want implied volatility across strikes and time.

Use `ProbCurve` when you want probabilities for one expiry, such as `P(S < 100)` or the median future price.

Use `ProbSurface` when you want those probability questions across several future dates.

## Warnings

Fitted objects keep structured warning details in `.warning_diagnostics`.

Use them when a result looks odd. They record issues such as stale quotes, skipped expiries, or repairs applied while building a probability distribution.