Choose your reading route
eduKateSG · Why Mathematics? · Circular Block Bootstrap: Wraparound Blocks and Edge Balance
Start with the question, audit the worked example, or jump to the assumptions and practice. Every chapter stays open for uninterrupted reading.
Understand the question
Define the quantity and follow the mechanism.Route 2 · Chapters 6–11
Reproduce the example
Audit the numbers one step at a time.Route 3 · Chapters 12–18
Check assumptions
See where design and calibration matter.Route 4 · Chapters 19–24
Explore applications
Connect the method to a concrete investigation.Route 5 · Chapters 25–35
Study and verify
Practise, inspect failure cases and read further.
Full chapter index · Practice and checking · Mathematics Learning Hub
Why is mathematics important when fixed blocks preserve time dependence but observations near the start and end should have equal opportunities to appear? The circular block bootstrap treats a finite series as a ring. Blocks can begin anywhere and wrap from the final observation to the first, balancing edge inclusion while preserving fixed-length local order.
Did You Know? In a circular series of N observations with block length l, every observation belongs to exactly l possible starting blocks. That symmetry removes the edge under-sampling of ordinary moving blocks.
Full chapter index
Chapters 1–5 · Understand the question
Chapters 6–11 · Reproduce the example
Chapters 12–18 · Check assumptions
Chapters 19–24 · Explore applications
Chapters 25–35 · Study and verify
A non-wrapping moving block of length l can start only at 1 through N−l+1. Observations near the middle appear in more possible blocks than observations at the ends. The circular construction permits all N starting positions.
The arch documentation defines the circular block bootstrap as using equal-length blocks with end-to-start wraparound. Read the class documentation and time-series overview.
For start s and within-block offset j, the source index is ((s+j−2) mod N)+1 in one-based notation. Modulo arithmetic turns the finite line of indices into a ring without changing internal order.
| Start | Block |
|---|---|
| 4 | 4, 6, 2 |
| 5 | 6, 2, 3 |
For series [2,3,5,4,6] with N=5 and l=3, the block starting at index 4 is [4,6,2], and the block starting at index 5 is [6,2,3]. Both cross the boundary using wraparound.
Draw starts 4 and 2. Concatenate [4,6,2] and [3,5,4] to get [4,6,2,3,5,4], then truncate to N=5: [4,6,2,3,5]. Its mean is 20/5=4.0.
| Replicate | Mean |
|---|---|
| 4, 6, 2, 3, 5 | 4.0 |
Each observation is contained in l of the N circular blocks. This is a combinatorial balance property. It does not imply that all dependence patterns are faithfully represented or that the artificial boundary is scientifically natural.
Every selected block has length l, except the final concatenated portion may be truncated. Fixed length makes the dependence horizon easy to explain but produces block boundaries at regular intervals in each replicate.
The observed last and first values may not have been temporal neighbours. Approximate stationarity makes their joint behaviour representative in distribution, but a strong trend or irreversible process makes wraparound implausible.
Small l creates many joins and weak dependence preservation. Large l creates fewer independently selected pieces. Dependence estimates, block-length selection rules and sensitivity analysis can guide the choice.
If the series covers complete repeating cycles, circularity can sometimes align naturally. However, wrapping an incomplete seasonal record may join incompatible phases. Model seasonality or align blocks with the actual cycle.
When resampling several variables observed at the same times, use the same index blocks for every column. Resampling columns separately destroys contemporaneous relationships.
Repeat block selection B times and calculate the target statistic. The empirical spread approximates sampling uncertainty under the block model. Store block starts so any surprising replicate can be reconstructed.
A larger B stabilises the estimated bootstrap distribution and p-values. It cannot add independent time information or repair a block length that fails to cover the dependence horizon.
Daily pollutant readings may have local weather dependence. Circular blocks can balance index inclusion, but a seasonal trend and exceptional events must be handled rather than hidden inside wraparound.
Queue lengths sampled every minute can persist over short runs. Blocks preserve congestion episodes, while the wrap assumption is more credible within a stable operating regime than across a system redesign.
Both methods wrap. Circular blocks use fixed l; stationary blocks have geometric lengths with expected size. The latter avoids deterministic boundaries, while the former offers simple, auditable block construction.
Moving blocks use only non-wrapping starts and can under-sample edges. Circular blocks restore symmetry at the price of an artificial end-to-start transition. Report that trade-off explicitly.
Show interval endpoints or standard errors across several l values, and compare moving, circular and stationary variants when edge effects are material. Do not select only the result that supports a preferred claim.
State N, l, circular indexing, target statistic, replicate count, seed and interval method.
Circular blocks do not fix nonstationarity, structural breaks, very short series, long memory, irregular spacing, measurement bias or causal design.
Students can use a clock-face diagram for indices and practise modulo arithmetic while seeing exactly which local transitions a block preserves.
List starts 1 through N, compute each block with modulo indexing, draw starts with replacement, concatenate and truncate. Verify that every original index appears in exactly l source blocks.
Use a documented circular bootstrap, pass aligned arrays together and set a seed. Unit-test the N=5, l=3 wrap blocks and compare output with a hand-built modulo function.
Circular blocks are used in finance, environmental science, signal analysis, operations and reliability. Careers also require recognising when the end-to-start join is scientifically defensible.
Simulate a stationary AR process and a trending process. Compare moving and circular block intervals. Explain why equal edge inclusion helps the first but does not cure the second.
Edge balance is not proof of validity. Wraparound does not mean the real process is periodic, fixed l is not automatically optimal, and more replicates do not replace more time observations.
The moving block guide explains non-wrapping blocks. The stationary bootstrap guide explains random lengths.
Students should draw all circular blocks for a five-point series. Parents can ask whether the artificial wrap makes sense and whether results survive another l.
- Build starts 4 and 5.
- Recreate the replicate mean.
- Prove equal inclusion count.
- Compare circular and moving blocks.
- Diagnose a trending series.
### Does every block have length l? Source blocks do.
### Why use modulo arithmetic? To wrap indices.
### Does circular mean periodic? Not necessarily.
### Are edges balanced? Yes, combinatorially.
Report the ordered data unit, resampling unit, tuning parameter, statistic, number of replicates, random seed, interval or test rule and sensitivity checks. Separate Monte Carlo variation from uncertainty in the original study.
There are N circular starts. For any fixed observation, exactly l consecutive starts place it within a length-l block. Thus its inclusion count is l, regardless of whether it lies near the displayed boundary.
The final selected block may be only partly used after concatenation reaches N. Across many replicates, symmetry of random starts still balances positions, but replicate-level counts vary.
Confirm time order, spacing, stationarity, l rationale, wrap plausibility, multivariate alignment, seed and sensitivity. Report seasonal or break structures that make the boundary artificial.
Continue through the Mathematics Learning Hub. Reproduce one resample by hand before trusting automated repetitions.
Extended practice laboratory
Practice case 1: Order the series
Setup: Start from the article’s circular block bootstrap worked mechanism for case 1. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 2: Choose l
Setup: Start from the article’s circular block bootstrap worked mechanism for case 2. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 3: List N starts
Setup: Start from the article’s circular block bootstrap worked mechanism for case 3. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 4: Apply modulo index
Setup: Start from the article’s circular block bootstrap worked mechanism for case 4. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 5: Build start-one block
Setup: Start from the article’s circular block bootstrap worked mechanism for case 5. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 6: Build end block
Setup: Start from the article’s circular block bootstrap worked mechanism for case 6. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 7: Verify equal inclusion
Setup: Start from the article’s circular block bootstrap worked mechanism for case 7. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 8: Draw first start
Setup: Start from the article’s circular block bootstrap worked mechanism for case 8. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 9: Draw second start
Setup: Start from the article’s circular block bootstrap worked mechanism for case 9. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 10: Concatenate blocks
Setup: Start from the article’s circular block bootstrap worked mechanism for case 10. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 11: Truncate to N
Setup: Start from the article’s circular block bootstrap worked mechanism for case 11. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 12: Calculate statistic
Setup: Start from the article’s circular block bootstrap worked mechanism for case 12. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 13: Repeat B times
Setup: Start from the article’s circular block bootstrap worked mechanism for case 13. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 14: Estimate standard error
Setup: Start from the article’s circular block bootstrap worked mechanism for case 14. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 15: Change l
Setup: Start from the article’s circular block bootstrap worked mechanism for case 15. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 16: Compare moving blocks
Setup: Start from the article’s circular block bootstrap worked mechanism for case 16. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 17: Compare stationary blocks
Setup: Start from the article’s circular block bootstrap worked mechanism for case 17. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 18: Add trend
Setup: Start from the article’s circular block bootstrap worked mechanism for case 18. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 19: Add seasonality
Setup: Start from the article’s circular block bootstrap worked mechanism for case 19. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 20: Align variables
Setup: Start from the article’s circular block bootstrap worked mechanism for case 20. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 21: Set seed
Setup: Start from the article’s circular block bootstrap worked mechanism for case 21. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 22: Reject periodic claim
Setup: Start from the article’s circular block bootstrap worked mechanism for case 22. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 23: Report sensitivity
Setup: Start from the article’s circular block bootstrap worked mechanism for case 23. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 24: Name limits
Setup: Start from the article’s circular block bootstrap worked mechanism for case 24. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 25: Write conclusion
Setup: Start from the article’s circular block bootstrap worked mechanism for case 25. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling unit, verify the probability or resampling identity and state one validation check before interpretation. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.