---
title: "The models view"
description: "The Review step's fifth view: what the second model said, shown as Hybrid Analysis on the hybrid algorithms and GF Calibration on the solver."
---

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

# The models view

import { GfSolver } from "@/components/react/PlannerScreens";
import { trimixDeco } from "@/demo/plans";

Three algorithms run two decompression models rather than one, and the Review
step then grows a fifth view holding what the second model said: **Hybrid
Analysis** under **Hybrid OC** and **DiveLogic Frontier**, **GF Calibration**
under **Thalmann GF Calibration**. Under **Bühlmann ZHL-16C**, **Thalmann E-L**
and **VPM-B** there is no second model and no view.

## Hybrid Analysis

The chart is the dive with two ceilings over it, one from Bühlmann and one from
Thalmann, and the gap between them filled: their disagreement, minute by minute.
Where they part company the deeper of the two governs your stop. Four figures
sit under it:

- **Max disagreement**, in metres: the widest the two ceilings ever got.
- **Mean disagreement**, in metres, across the dive.
- **Bühlmann controlled** and **Thalmann controlled**, as percentages: which
  model set the ceiling, and for how much of the ascent.

Below them, a **Decompression stops** table gives each stop its depth, duration
and gas, both models' ceilings, and a **Controlling** badge naming which model
produced it. A schedule one model controls throughout is one you could have got
from that model alone. **DiveLogic Frontier** also shows the family of schedules
it swept, with the one it executed marked.

> **Caution**
>
> If the hybrid solver could not run, a banner says so and replaces the stop
> table. The plan you are looking at is then a plain Bühlmann schedule, and the
> disagreement figures describe nothing you dived.

## GF Calibration

**Thalmann GF Calibration** does not take a gradient-factor pair, it produces
one. The view opens with the answer:

- **GF Low** and **GF High**: the pair the planner fitted, as percentages.
- **TTS Delta**, in minutes: how much longer or shorter the fitted Bühlmann
  schedule is than the Thalmann schedule it was fitted to.
- **Max divergence**, in metres, between the two ceilings.

Below them the chart draws both ceilings and the depth profile, as Hybrid
Analysis does, over a row naming the Thalmann parameter set used as the
reference. A **Decompression stops** table puts the two schedules side by side.

**GF Low** and **GF High** fields sit at the top of the card. Left alone it
reads *Auto-solved*; type into either and it switches to *Manual comparison
active*, comparing your pair against the Thalmann reference instead. **Reset to
auto-solve** returns it.

A fitted pair is a translation between two vocabularies, not an equivalence: it
holds for one dive on one parameter set, and changes with the depth, the bottom
time or the gases.

## On a rebreather

This is the one Review view the bailout rail does not appear on, and it is
unavailable under the hybrid and calibration algorithms with bailout
enumeration, because those algorithms produce no scenarios.

## Next

[Warnings](/planning/warnings), and [The GF solver](/theory/gf-solver) for what
a fit does and does not tell you.

Source: https://docs.divelogic.ai/planning/review-models/index.mdx