Parabolic Trendline in Excel: How to Add and Use It

Charts & pivot tables · Updated November 21, 2026

To add a parabolic trendline in Excel for Microsoft 365 on Windows, select an XY scatter chart’s data series, choose Chart Design > Add Chart Element > Trendline > More Trendline Options, then select Polynomial and set Order to 2. A second-order polynomial is the parabolic curve; Excel can also show its fitted equation on the chart.

What a parabolic trendline represents

A parabolic trendline models a relationship between x and y with a quadratic equation: y = ax² + bx + c. The x² term creates the curve, while the values of a, b and c determine its direction, steepness and vertical position. When a is positive, the curve opens upward; when a is negative, it opens downward. If a is zero, the equation reduces to a straight line.

In Excel, a polynomial trendline can have several orders. Order 1 is linear, Order 2 is parabolic, and higher orders can bend more often. Choosing Order 2 is therefore the direct way to request a parabola. It does not mean Excel has discovered a cause or guaranteed that the relationship will continue; it has calculated a curve that fits the selected data.

A useful distinction: a trendline is an estimated model, not the same thing as connecting the data points. A line chart may connect points in sequence, while a polynomial trendline summarizes their overall pattern. For numerical x-values such as time, distance or temperature, an XY scatter chart generally represents the x-axis more appropriately than a line chart with category labels.

Add a polynomial trendline to a chart

  1. Arrange paired numeric observations in two columns, with x-values in one column and corresponding y-values in the next. Select both columns, including their headings if present.
  2. Choose Insert > Insert Scatter (X, Y) or Bubble Chart, then select a scatter chart style. Confirm that the horizontal axis uses the x-values and the plotted series uses the y-values.
  3. Click a point in the data series to select the series. Choose Chart Design > Add Chart Element > Trendline > More Trendline Options. Alternatively, right-click the series and choose Add Trendline.
  4. In the Format Trendline pane, select Polynomial, set Order to 2, and, if useful, select Display Equation on chart and Display R-squared value on chart.

The equation and R-squared value are optional labels, not separate chart types. To make the curve easier to read, give the chart a title, label both axes with units, and use a clear series name. On Windows, Chart Design appears when the chart is selected. In Excel for Mac, select the series and use the chart’s Chart Design controls or its context menu to add a trendline; the exact ribbon presentation can vary by release. Excel for the web has fewer chart-formatting controls than the desktop app, so use the Windows or Mac desktop version if the polynomial order or equation options are unavailable.

Choose data and chart type carefully

A quadratic fit is meaningful only if the observations support a curved pattern. Make sure the x and y values are numeric, correctly paired, and measured on a consistent basis. Blank cells, text stored as numbers, accidental sorting of only one column, or an incorrect x-axis can produce a chart that looks plausible but represents the wrong data.

Use an XY scatter chart when x-values are actual quantities or unevenly spaced dates. For example, observations at 1, 2, 10 and 20 should be spaced according to their numeric distances. A line chart treats its horizontal positions as categories and can space those observations evenly instead. That distinction can change the apparent shape of a curve.

Before fitting, inspect a scatter plot without a trendline. Ask whether the points show a single broad bend, whether there are outliers, and whether the chosen range includes only observations relevant to the question. A quadratic can be pulled toward an extreme value, and adding or removing points may substantially alter the coefficients.

Read the equation and R-squared value

Suppose Excel displays y = 1.4x² − 2.1x + 5. In this example, the positive coefficient on x² means the fitted parabola opens upward. The equation estimates y for a given x, within the range and context of the data used to fit it. It does not say that every observed point lies on the curve.

R-squared describes how much of the variation in the observed y-values is accounted for by the fitted model in that dataset. A value closer to 1 indicates a closer fit by this measure, but it does not prove that the model is appropriate, that the relationship is causal, or that predictions outside the observed x-range will be accurate. A more flexible model can often fit existing points better while predicting new data worse.

Excel rounds the equation shown on a chart for display. Do not copy rounded coefficients into a calculation that needs precision: small rounding differences can lead to noticeably different results, especially for large x-values or a curve with substantial curvature. Increase the displayed decimal places if needed, or calculate coefficients from the worksheet with greater precision.

Calculate the quadratic equation in worksheet cells

For a reusable calculation, you can fit the coefficients with LINEST. This method uses helper columns so the squared term is explicit. Assume x-values are in A2:A5 and observed y-values are in B2:B5. Put x in C2:C5 and x squared in D2:D5. With x in A2, enter =A2 in C2 and fill down; with x in A2, enter =A2^2 in D2 and fill down. These helper columns should contain the same rows as the y-values.

With observed y-values in B2:B5 and the x and x-squared predictors in C2:D5, enter =LINEST(B2:B5,C2:D5,TRUE,FALSE) in F1. In Microsoft 365, the result spills across F1:H1: the coefficients are returned in reverse predictor-column order, so F1 is a for x², G1 is b for x, and H1 is c, the intercept. LINEST is also available in older desktop Excel versions, though older versions may require selecting three horizontal cells and confirming the formula as an array formula with Ctrl+Shift+Enter. Mac keyboards use Control+Shift+Return for the legacy array-entry method.

For a checkable example, let A2:A5 contain 0, 1, 2 and 3, and B2:B5 contain 2, 4, 8 and 14. These values follow y = x² + x + 2 exactly. The helper columns therefore contain x values 0, 1, 2, 3 and squared values 0, 1, 4, 9. LINEST returns 1, 1 and 2 across F1:H1, respectively. The fitted value at x = 3 is 1×9 + 1×3 + 2 = 14, matching the example observation.

With coefficients a, b and c in F1:H1 and the x-value to estimate in A2, enter =$F$1*A2^2+$G$1*A2+$H$1 in the prediction cell. The formula assumes F1 holds a, G1 holds b, H1 holds c, and A2 holds the x-value. For the example where A2 is 0, it returns 2. Fill it down beside a list of x-values to create fitted estimates, then add those estimates as another series if you want a curve based on worksheet values.

Understand what the curve can and cannot tell you

A parabolic fit is often useful when a measured response rises and then falls, falls and then rises, or changes at an accelerating rate over a limited interval. Examples might include a controlled experiment measuring response at several settings, or a dataset in which a quantity has a visible turning point. The curve’s vertex is the turning point of the mathematical model, but whether that point is useful in practice depends on the measurements and the situation being studied.

Be cautious about interpreting the vertex when it lies outside the observed x-range. That location is an extrapolation, not a feature supported directly by nearby data. Also consider whether the response should plausibly be quadratic: many real processes have thresholds, plateaus, cycles or other shapes that a parabola does not describe well. A visually smooth curve is not evidence by itself that the underlying process follows a quadratic law.

For analysis, compare the curve with the plotted observations and, where possible, assess predictions against data not used to fit it. If the objective is forecasting, a fit that closely follows historical points is not automatically the best forecast. Report the data range and the equation’s units, since coefficients have units determined by the x- and y-measurements.

Compare a quadratic fit with other trendlines

Excel offers linear, exponential, logarithmic, power, polynomial and moving-average trendlines. Each represents a different kind of pattern. A linear model describes a constant change in y for each unit of x. An exponential model describes a proportional rate of change and requires appropriate positive data. A moving average smooths variation rather than fitting a quadratic equation.

A polynomial of order 2 is appropriate when a single broad bend is supported by the data. Higher polynomial orders can create multiple bends and may follow noise or outliers, particularly when there are few observations. A lower R-squared for a quadratic than for a higher-order polynomial does not by itself justify the more complex curve. Prefer a model that fits the subject matter and performs reasonably on relevant data, rather than choosing solely by the chart’s fit statistic.

Fix common trendline problems

  • Polynomial is missing or the order control is unavailable: Check that a data series is selected and that the chart is a supported chart type. A standard XY scatter chart is a reliable starting point.
  • The curve looks straight: The fitted x² coefficient may be close to zero, or the axis scale may make curvature hard to see. Do not force a more complex order just to make a visible bend.
  • The curve looks wrong: Verify the x/y columns, the selected series, and the plotted range. Check for outliers, text values, or an unintended category axis.
  • The equation disagrees with worksheet calculations: Check the displayed rounding, coefficient order, and whether the worksheet model uses the same observations. For LINEST, predictors in C:D return coefficients in reverse order: x² first, then x, followed by the intercept.
  • The fit changes sharply when one point is removed: The dataset may be small or that point may have strong influence. Investigate the observation rather than deleting it solely to improve the curve.

Frequently Asked Questions

How do I make a parabolic trendline in Excel?
Select an XY scatter chart’s data series, choose Chart Design > Add Chart Element > Trendline > More Trendline Options, select Polynomial, and set Order to 2.

Is a second-order polynomial the same as a parabola?
Yes. A polynomial with a nonzero x² term and no higher powers has the quadratic form y = ax² + bx + c, whose graph is a parabola.

Can I use a parabolic trendline to predict future values?
You can calculate estimates from its equation, but predictions beyond the observed x-range are extrapolations and may be unreliable. Check whether a quadratic model makes sense and validate predictions with relevant data.

Why does my trendline equation look different from my calculation?
The chart label may round coefficients, or the chart and worksheet may use different data ranges. Use unrounded worksheet coefficients and confirm the x-values, y-values and coefficient order.

What does R-squared mean on a polynomial trendline?
It summarizes how closely the fitted curve accounts for variation in the selected y-values. It is not proof of causation, a guarantee of future accuracy, or a substitute for checking the model and data.

Can Excel calculate the quadratic coefficients without a chart?
Yes. Use LINEST with columns for x and x² as predictors. In Microsoft 365, its coefficients can spill into adjacent cells; for older versions, the array-entry procedure may differ.

Related

Try it: the Percentage Formula Generator: percent of, percent change, increase and decrease, with the formula.

More in Charts & pivot tables

Sources: Function behaviour and menu paths: Microsoft Support documentation for Excel. Method & sources

Spotted an error in this guide? Let us know.