Success Rate Curve Calculator

Results

Calculated values, comparisons, targets, and curve diagnostics appear here.

Enter data or load the sample, then calculate the curve.

Interactive success-rate chart

Interpretation

Model ranking

Rank	Series	Average	Final	Target

Detailed calculations

Series	X	Successes	Attempts	Rate	Failure	Cumulative	Rolling	Smoothed	Change	CI	Target

Data input

Use counts, percentages, or binary outcomes across one or more series.

Input mode

X-axis meaning

Missing-value policy

Series filter

Enter comma-separated series names to include.

X value	Successes or outcome	Total attempts	Success rate (%)	Series name

Binary outcomes accept 1, 0, yes, no, success, or failure.

Paste comma, tab, semicolon, or whitespace-separated data

CSV file

Recognized headers include x, successes, attempts, rate, group, model, and series.

Curve and calculation options

Displayed curve

Rolling window

Smoothing method

EMA alpha or smoothing strength

Confidence interval

Confidence level

Target success rate (%)

Baseline reference (%)

Random-chance reference (%)

Selected threshold or milestone

Minimum X filter

Maximum X filter

Sort points by X value

Merge duplicate X values

Normalize numeric X values to 0–1

Clamp rates between 0% and 100%

Show original curve with smoothing

Show confidence bands

Show point markers

Chart appearance

Chart title

X-axis label

Y-axis label

Line style

Line pattern

Line width

Marker size

Y-axis display

Legend position

Y-axis minimum

Y-axis maximum

Show gridlines

Use logarithmic X-axis

Series colors

Colors update when series names change or data is calculated.

Formula used

Success Rate = (Successful Attempts ÷ Total Attempts) × 100

A success rate describes the share of attempts meeting your rule. Counts provide the strongest basis for weighted summaries. Larger samples usually produce narrower confidence intervals.

Cumulative rates combine all observations through each selected point. Rolling rates use only the latest chosen window. Smoothed rates reduce noise while preserving the broad pattern.

How to use the calculator

Select counts, percentages, or binary outcomes.
Enter X values and assign meaningful series names.
Choose rolling, smoothing, confidence, and target settings.
Customize labels, references, filters, and chart appearance.
Calculate, inspect diagnostics, then export needed reports.

Use counts whenever successful and total attempts are available. Rate-only data can still create useful descriptive curves. Binary observations are aggregated into rates automatically.

Worked example

Suppose an agent succeeds 34 times during 40 attempts. Its success rate equals 34 divided by 40. Multiplying by 100 gives an 85 percent rate.

Episode	Successes	Attempts	Success rate	Series
1	22	40	55.00%	Baseline
4	30	40	75.00%	Baseline
8	38	40	95.00%	Optimized

Frequently asked questions

What does a success-rate curve show?

It shows how often a defined success occurs across ordered observations.

Should I use counts or percentages?

Use counts when possible because weighted summaries remain more accurate.

What is a cumulative success rate?

It combines every success and attempt through the current point.

What does the rolling window control?

It controls how many recent observations form each rolling estimate.

Which confidence interval should I choose?

Wilson intervals are dependable for many binomial success-rate datasets.

Can I compare several models?

Yes. Give each model a different series name before calculating.

How are duplicate X values handled?

They can be merged using summed counts or averaged percentages.

What does first target crossing mean?

It is the earliest point reaching or exceeding your target.

Can the chart use thresholds?

Yes. Select threshold as the X-axis meaning and enter values.

Why might a curve look unstable?

Small samples, changing conditions, or rare outcomes can increase variability.