Main Window

The main window is the central workspace in CFTool. It displays your data, fits, and results, and provides quick access to key information about fit quality and performance.

Layout overview

The window is divided into four panels:

  1. Main graph — plots your data, fits, and any additional lines or functions.

  2. Residuals plot — appears below the main graph.

  3. Results panel — to the right of the main graph.

  4. Feedback panel — below the results panel.

Residuals plot

The residuals plot shows how each data point deviates from the fitted curve.

  • If your model is correct, the residuals should resemble random Gaussian noise scattered evenly around zero.

  • If they show a pattern or structure, your model may be incomplete or physically inappropriate.

Use these buttons for control:

  • Hide Residuals – hides the residuals panel to enlarge the main graph area.

  • Show Histogram – opens a separate window showing a histogram of the residuals, which can help visualise their distribution.

Interpreting residuals

Tip: How to read the residuals plot

  • Random scatter around zero → the model is likely suitable.

  • Curvature or waves → the model function may be missing a term (e.g., quadratic or sinusoidal behaviour).

  • Systematic offset → a constant term or background may be missing.

  • Clusters or asymmetry → uncertainties (σ values) may be mis-estimated, or the weighting may need review.

Randomness in the residuals is usually a good sign; structure or bias means the fit can be improved.

Results panel

The Results panel, to the right of the main plot, lists:

  • The best-fit coefficient values.

  • The confidence intervals (errors) for each coefficient.

  • The fit limits used for the analysis.

  • The definite integral of the fit between those limits.

These values update automatically each time a fit is performed.

Fit quality indicators

Below the coefficient table, several numerical indicators help assess the quality of the fit:

  • SSE (Sum of Squares of Errors) — the sum of the squared differences between the data and the fitted model.

    • Smaller values indicate a better numerical match to the data.

  • R² (Coefficient of Determination) — measures how well the model reproduces the variation in the data.

    • Values close to 1.0 indicate an excellent fit; values near 0 indicate a poor one.

Note: High SSE or low R² indicate a poor match, but even a high R² does not confirm that the chosen function is physically valid — it only shows that it reproduces the shape of the data well.

Feedback panel

The text panel beneath the Results section provides messages, warnings, and feedback.
It reports:

  • Problems encountered during fitting.

  • Issues with custom functions (e.g., syntax or undefined parameters).

  • Other messages relevant to particular fitting tools.

If a fit fails or produces unexpected results, check this panel first — it usually indicates what went wrong.

Back to menu