What is an OC Curve?

An Operating Characteristic (OC) Curve is a statistical visualization tool used in quality control and acceptance sampling. It illustrates the probability of accepting a batch or lot of products based on the percentage of defects within that batch. By mapping the fraction defective on the x-axis against the probability of acceptance (Pa) on the y-axis, manufacturers can determine the effectiveness of their sampling plans.

How to Interpret the OC Curve

The curve typically starts at a probability of 1.0 (100% acceptance) when there are zero defects and slopes downward as the percentage of defects increases. A steeper curve indicates a more discriminating sampling plan—meaning the plan is better at distinguishing between high-quality and low-quality lots. Two critical points often identified on this curve are the Acceptable Quality Level (AQL) and the Lot Tolerance Percent Defective (LTPD).

Producer's vs. Consumer's Risk

The OC Curve helps manage two types of statistical risks. The Producer's Risk (α) is the probability of rejecting a good lot (type I error), while the Consumer's Risk (β) is the probability of accepting a bad lot (type II error). By adjusting the sample size (n) and the acceptance number (c), engineers can fine-tune the sampling plan to strike a balance between inspection costs and risk mitigation.

How to Use This Calculator

To use this tool, simply enter your Sample Size (n)—the number of items inspected from the lot—and your Acceptance Number (c)—the maximum number of defective items allowed to still accept the entire lot. Click "Generate Curve" to visualize how the probability of acceptance changes across various defect rates (from 0% to 20%). This tool utilizes the Binomial Distribution formula, which is the industry standard for most acceptance sampling scenarios where the lot size is significantly larger than the sample size.

Frequently Asked Questions

Q: Does batch size (N) matter?
A: When the batch size is very large compared to the sample size (typically N > 10n), the Binomial distribution is highly accurate. For smaller batch sizes, the Hypergeometric distribution is technically more precise, though the difference is often negligible in industrial practice.

Q: How does increasing 'n' affect the curve?
A: Increasing the sample size (n) makes the curve steeper, reducing both producer and consumer risk but increasing inspection time and costs.