False Positive Rate: Formula and Calculation

Calculate the false positive rate by dividing the number of false positives by the total number of actual negatives, then multiplying by 100 for a percentage. Enter the false-positive count and true-negative count to find the rate, along with the total number of actual negatives for context.

Formula

The false positive rate (FPR) is the share of actual negative cases that a test or classifier incorrectly labels positive:

FPR = FP / (FP + TN) × 100%

• FP is the number of false positives: negative cases incorrectly called positive.
• TN is the number of true negatives: negative cases correctly called negative.
• FP + TN is the total number of actual negatives.

Enter counts, not percentages, in the calculator. Both counts must refer to the same set of cases and the same decision threshold. The denominator must be greater than zero; if there are no actual negatives, the rate is undefined.

Worked Example

Using the calculator’s default values, there are 10 false positives and 90 true negatives:

1. Total actual negatives = FP + TN = 10 + 90 = 100.
2. False positive rate = FP / (FP + TN) × 100%.
3. FPR = 10 / 100 × 100% = 10.00%.

The result means that, among the 100 cases that were actually negative, 10 were incorrectly classified as positive. It does not mean that 10% of all positive results are wrong.

What the Rate Measures

A false positive occurs when a system signals a condition or category that is not actually present. The term is used in fields such as diagnostic testing, spam filtering, quality inspection, and machine learning. The precise meaning of “positive” depends on the task: it is the outcome the system is trying to detect, not necessarily a favorable result.

The rate is also known as the false alarm rate or, in statistical hypothesis testing, the Type I error rate under the relevant assumptions. It measures errors among actual negatives. That makes it different from a count of mistakes, which can rise or fall simply because the dataset gets larger or smaller.

For example, a filter reviews 1,000 messages: 200 are actually spam and 800 are not. If it incorrectly marks 40 of the non-spam messages as spam, then there are 40 false positives and 760 true negatives. The FPR is 40 / (40 + 760) × 100% = 5%. The 200 actual spam messages do not belong in this calculation because they are positive cases.

How to Read It Alongside Other Metrics

The FPR is one part of a confusion matrix, which sorts outcomes by actual class and predicted class. Its counterpart among actual positives is the false negative rate (FNR): the proportion of actual positives incorrectly labeled negative. Sensitivity, also called recall or the true positive rate, is 1 − FNR when the rates are written as proportions.

Do not confuse FPR with the false discovery rate (FDR), sometimes described as the proportion of positive predictions that are wrong:

• FPR = FP / (FP + TN): false positives divided by all actual negatives.
• FDR = FP / (FP + TP): false positives divided by all predicted positives.
• Precision = TP / (TP + FP): true positives divided by all predicted positives.

These metrics answer different questions. FPR asks how often actual negatives trigger a positive result. FDR asks how often positive alerts turn out to be false. A system can have a low FPR but still have a high FDR if the positive condition is rare, because even a small share of false alarms may outnumber the true detections.

Thresholds, Data, and Common Errors

Many classifiers produce a score and label a case positive when it passes a chosen threshold. Lowering that threshold often catches more true positives but can also increase false positives; raising it often reduces false alarms while missing more positives. Comparing FPRs is meaningful only when the threshold and evaluation conditions are stated.

• Use the right denominator. Divide by all actual negatives, FP + TN, not by all cases or all predicted positives.
• Keep classes straight. A false positive is an actual negative predicted positive; a true negative is an actual negative predicted negative.
• Use one consistent dataset. Counts should come from the same evaluation sample and classification rule.
• Do not mistake a rounded rate for an exact one. Keep counts for the calculation and round only the displayed result.
• Check for zero negatives. With FP + TN = 0, no FPR can be calculated.

For a useful estimate, the test data should represent the conditions where the system will be used. A rate observed on a small or unrepresentative sample may not carry over to another population. When reporting results, include the number of actual negatives and, where relevant, the threshold; a percentage alone hides how much data supports it.

Frequently Asked Questions

What is the formula for false positive rate?
FPR = FP / (FP + TN). Multiply by 100 to express it as a percentage. FP is false positives and TN is true negatives.

Is false positive rate the same as false discovery rate?
No. FPR divides false positives by all actual negatives. FDR divides false positives by all positive predictions, FP + TP.

Can the false positive rate be greater than 100%?
No, not when the counts are valid and FP is part of the actual-negative total FP + TN. The rate ranges from 0% to 100%, inclusive.

What does a 5% false positive rate mean?
It means that 5 out of every 100 actual negative cases were labeled positive in the measured sample or setting. It does not say that 5% of positive alerts are false.

What if there are no true negatives?
If there are false positives but no true negatives, all observed actual negatives were misclassified, so the FPR is 100%. If there are no actual negatives at all (FP + TN = 0), the rate is undefined.

Related

• 6 Sided Dice Probability Calculator for Sums
• 5 Number Summary Outlier Calculator: Formula and Examples
• 5 Number Summary Calculator and Boxplot Guide

in fp | False positives | count | 10 | 0
in tn | True negatives | count | 90 | 0
out rate | False positive rate | % | fp / (fp + tn) * 100 | 2
out neg | Actual negatives | count | fp + tn | 0