Model
Rows a sample needs to see a rare defect
Sampling is cheapest exactly where it is weakest: the expensive defects are the rare ones.
- Prevalence
- the share of rows with the defect
- Miss probability
- chance of seeing it zero times (confidence 1 − β)
How it behaves
One input on the horizontal axis, the others held fixed.
95% confidence99% confidence
Data table
| Prevalence p | series | Rows to sample |
|---|---|---|
| 0 | 95% confidence | 30M |
| 0 | 95% confidence | 25.2M |
| 0 | 95% confidence | 21.2M |
| 0 | 95% confidence | 17.8M |
| 0 | 95% confidence | 15M |
| 0 | 95% confidence | 12.6M |
| 0 | 95% confidence | 10.6M |
| 0 | 95% confidence | 8.94M |
| 0 | 95% confidence | 7.52M |
| 0 | 95% confidence | 6.33M |
| 0 | 95% confidence | 5.33M |
| 0 | 95% confidence | 4.48M |
| 0 | 95% confidence | 3.77M |
| 0 | 95% confidence | 3.17M |
| 0 | 95% confidence | 2.67M |
| 0 | 95% confidence | 2.25M |
| 0 | 95% confidence | 1.89M |
| 0 | 95% confidence | 1.59M |
| 0 | 95% confidence | 1.34M |
| 0 | 95% confidence | 1.13M |
| 0 | 95% confidence | 947K |
| 0 | 95% confidence | 797K |
| 0 | 95% confidence | 671K |
| 0 | 95% confidence | 564K |
| 0 | 95% confidence | 475K |
| 0 | 95% confidence | 399K |
| 0 | 95% confidence | 336K |
| 0 | 95% confidence | 283K |
| 0 | 95% confidence | 238K |
| 0 | 95% confidence | 200K |
| 0 | 95% confidence | 168K |
| 0 | 95% confidence | 142K |
| 0 | 95% confidence | 119K |
| 0 | 95% confidence | 100K |
| 0 | 95% confidence | 84.4K |
| 0 | 95% confidence | 71K |
| 0 | 95% confidence | 59.8K |
| 0 | 95% confidence | 50.3K |
| 0 | 95% confidence | 42.3K |
| 0 | 95% confidence | 35.6K |
| 0 | 95% confidence | 30K |
| 0 | 95% confidence | 25.2K |
| 0 | 95% confidence | 21.2K |
| 0 | 95% confidence | 17.8K |
| 0 | 95% confidence | 15K |
| 0 | 95% confidence | 12.6K |
| 0 | 95% confidence | 10.6K |
| 0 | 95% confidence | 8.94K |
| 0 | 95% confidence | 7.52K |
| 0 | 95% confidence | 6.33K |
| 0.001 | 95% confidence | 5.33K |
| 0.001 | 95% confidence | 4.48K |
| 0.001 | 95% confidence | 3.77K |
| 0.001 | 95% confidence | 3.17K |
| 0.001 | 95% confidence | 2.67K |
| 0.001 | 95% confidence | 2.25K |
| 0.002 | 95% confidence | 1.89K |
| 0.002 | 95% confidence | 1.59K |
| 0.002 | 95% confidence | 1.34K |
| 0.003 | 95% confidence | 1.13K |
| 0.003 | 95% confidence | 947 |
| 0.004 | 95% confidence | 797 |
| 0.004 | 95% confidence | 671 |
| 0.005 | 95% confidence | 564 |
| 0.006 | 95% confidence | 475 |
| 0.008 | 95% confidence | 399 |
| 0.009 | 95% confidence | 336 |
| 0.011 | 95% confidence | 283 |
| 0.013 | 95% confidence | 238 |
| 0.015 | 95% confidence | 200 |
| 0.018 | 95% confidence | 168 |
| 0.021 | 95% confidence | 142 |
| 0.025 | 95% confidence | 119 |
| 0.03 | 95% confidence | 100 |
| 0.036 | 95% confidence | 84.4 |
| 0.042 | 95% confidence | 71 |
| 0.05 | 95% confidence | 59.8 |
| 0.06 | 95% confidence | 50.3 |
| 0.071 | 95% confidence | 42.3 |
| 0.084 | 95% confidence | 35.6 |
| 0.1 | 95% confidence | 30 |
| 0 | 99% confidence | 46.1M |
| 0 | 99% confidence | 38.7M |
| 0 | 99% confidence | 32.6M |
| 0 | 99% confidence | 27.4M |
| 0 | 99% confidence | 23.1M |
| 0 | 99% confidence | 19.4M |
| 0 | 99% confidence | 16.3M |
| 0 | 99% confidence | 13.7M |
| 0 | 99% confidence | 11.6M |
| 0 | 99% confidence | 9.73M |
| 0 | 99% confidence | 8.19M |
| 0 | 99% confidence | 6.89M |
| 0 | 99% confidence | 5.8M |
| 0 | 99% confidence | 4.88M |
| 0 | 99% confidence | 4.1M |
| 0 | 99% confidence | 3.45M |
| 0 | 99% confidence | 2.91M |
| 0 | 99% confidence | 2.44M |
| 0 | 99% confidence | 2.06M |
| 0 | 99% confidence | 1.73M |
| 0 | 99% confidence | 1.46M |
| 0 | 99% confidence | 1.23M |
| 0 | 99% confidence | 1.03M |
| 0 | 99% confidence | 867K |
| 0 | 99% confidence | 730K |
| 0 | 99% confidence | 614K |
| 0 | 99% confidence | 517K |
| 0 | 99% confidence | 435K |
| 0 | 99% confidence | 366K |
| 0 | 99% confidence | 308K |
| 0 | 99% confidence | 259K |
| 0 | 99% confidence | 218K |
| 0 | 99% confidence | 183K |
| 0 | 99% confidence | 154K |
| 0 | 99% confidence | 130K |
| 0 | 99% confidence | 109K |
| 0 | 99% confidence | 91.9K |
| 0 | 99% confidence | 77.3K |
| 0 | 99% confidence | 65K |
| 0 | 99% confidence | 54.7K |
| 0 | 99% confidence | 46.1K |
| 0 | 99% confidence | 38.7K |
| 0 | 99% confidence | 32.6K |
| 0 | 99% confidence | 27.4K |
| 0 | 99% confidence | 23.1K |
| 0 | 99% confidence | 19.4K |
| 0 | 99% confidence | 16.3K |
| 0 | 99% confidence | 13.7K |
| 0 | 99% confidence | 11.6K |
| 0 | 99% confidence | 9.73K |
| 0.001 | 99% confidence | 8.19K |
| 0.001 | 99% confidence | 6.89K |
| 0.001 | 99% confidence | 5.8K |
| 0.001 | 99% confidence | 4.88K |
| 0.001 | 99% confidence | 4.1K |
| 0.001 | 99% confidence | 3.45K |
| 0.002 | 99% confidence | 2.91K |
| 0.002 | 99% confidence | 2.44K |
| 0.002 | 99% confidence | 2.06K |
| 0.003 | 99% confidence | 1.73K |
| 0.003 | 99% confidence | 1.46K |
| 0.004 | 99% confidence | 1.23K |
| 0.004 | 99% confidence | 1.03K |
| 0.005 | 99% confidence | 867 |
| 0.006 | 99% confidence | 730 |
| 0.008 | 99% confidence | 614 |
| 0.009 | 99% confidence | 517 |
| 0.011 | 99% confidence | 435 |
| 0.013 | 99% confidence | 366 |
| 0.015 | 99% confidence | 308 |
| 0.018 | 99% confidence | 259 |
| 0.021 | 99% confidence | 218 |
| 0.025 | 99% confidence | 183 |
| 0.03 | 99% confidence | 154 |
| 0.036 | 99% confidence | 130 |
| 0.042 | 99% confidence | 109 |
| 0.05 | 99% confidence | 91.9 |
| 0.06 | 99% confidence | 77.3 |
| 0.071 | 99% confidence | 65 |
| 0.084 | 99% confidence | 54.7 |
| 0.1 | 99% confidence | 46.1 |
Worked examples, recomputed
Every number the mechanisms quote for this relationship is recomputed from it, so none of them disagrees with it.
Where it appears
Each mechanism states it in its own names. The general symbol each name stands for is shown with it.
- Data Quality & Validation
How do you check that data is right, when "right" is a claim about the world and not about the bytes?