1. The design comes first | 2. Feeding paired losses | 3. Reading the result | 4. What the package refuses | 5. Inconclusive is an outcome | 6. Boundaries

The data | A first attempt: classical I-MR | The right tool: regression chart with phase detection | Why log-log? | Methodological caveats | A note on the package's history | References

A simple example: linear drift | Sigma from residuals | Phase detection | A worked example with many phases and visible violations | The model menu | A growth-curve example | Interpreting violations | References

1. The two articles | 2. The adapted chart in package terms | 3. Why the chart has to be adapted (SBPO Fig. 2) | 4. Brazil (SBPO Fig. 3, RBE Fig. 6) | Retrospective reconstruction with the article's phases | Automatic detection | 5. Recife (SBPO Fig. 4) | 6. Pernambuco and São Paulo (RBE Figs. 7 and 8) | 7. Phase II: the "Monitoramento" segment | 8. Differences from the originals | References

1. What "simulação" means in the articles | 2. The replay, done explicitly | The loop | Recife and Brazil | How close is the replay to the article? | The chart the replay produces | 3. A Monte Carlo evaluation of the strategy | Design | A fast replay, identical to the package loop | Results | Reading the table against the theory | 4. Practical guidance | References

1. The design comes first | 2. Feeding paired losses | 3. Reading the result | 4. What the package refuses | 5. Inconclusive is an outcome | 6. Boundaries

p chart with variable n | c chart and Poisson honesty | np chart for constant n | u chart for variable exposure | References

A textbook example | Interaction with shewhart_regression(model = "auto") | When not to transform | References

Five minutes, three charts | 1. An I-MR chart on individual measurements | 2. An Xbar-R chart on a subgrouped process | 3. A c chart on defect counts | Working with the result | Phase I and Phase II | A roadmap for the rest of the documentation | References

A worked example | Phase I: calibration | Phase II: monitoring | Why the trim step matters | When this matters most | References

What the panels show | A diagnosis-driven fix: Box-Cox | When non-normality is a feature, not a bug | References

When to use which | I-MR | Robust alternatives | Xbar-R | Xbar-S | Reading the results | References

When the Shewhart chart is too slow | EWMA — Exponentially Weighted Moving Average | Choosing lambda | CUSUM — Cumulative Sum | Choosing k and h | Phase II monitoring | Which one should I use? | References

Closed-form benchmarks | Adding rules sharpens detection | A quantitative comparison: WE-7 vs Nelson 2 | Beyond normal residuals | References

When univariate charts are not enough | What the chart computes | Diagnosing an alarm with contributions | Phase II workflow | When not to reach for a Hotelling chart | References