Published 16 October 2026
Reference report / Edition 01
A complete, frozen rendering of the synthetic reference fixture. The values are retained with their input identity and method version.
Synthetic monthly observations, January 2023–August 2024. No live account, model or monitoring service is represented. Historical-looking dates are fixture keys. The declared delivery timestamp is invented metadata.
within example bounds
FIXTURE RESULT| Record | Retained value |
|---|
| Engine version | illustrative-monitor/1.0.0 |
| Model identity | SYNTHETIC-MONTHLY-RESEARCH / example-1.0 |
| As-of timestamp | 2024-09-02T12:00:00Z |
| Last-updated timestamp | 2024-09-01T12:00:00Z |
| Periods | 20 explicitly listed monthly dates |
| Review conclusion | Within the selected example thresholds; no operational approval |
Data quality before interpretation.
| Metric | Result | Interpretation |
|---|
| Observed / expected | 20 / 20 | Counts reconcile; identity checks are also required |
| Duplicate dates | 0 | Repeated date keys |
| Missing expected dates | 0 | Declared schedule only; no inferred exchange calendar |
| Null feature share | 0.000% | Observed rows are the denominator |
| Declared delivery age | 24.000000 hours | As-of minus supplied last-updated timestamp |
Frozen distribution comparison.
The reference fixture deliberately keeps the same feature distribution. Fixed edges are −1, 0 and 1; values exactly on an edge enter the bin to its right. Additive smoothing is 0.5 per bin, normalized separately within each sample. The mean-shift denominator is the baseline sample standard deviation.
| Bin bounds | Baseline count | Current count | Smoothed baseline share | Smoothed current share |
|---|
| −∞ to -1 | 4 | 4 | 0.204545 | 0.204545 |
| -1 to 0 | 6 | 6 | 0.295455 | 0.295455 |
| 0 to 1 | 6 | 6 | 0.295455 | 0.295455 |
| 1 to +∞ | 4 | 4 | 0.204545 | 0.204545 |
| Diagnostic | Value |
|---|
| PSI | 0.000000 |
| Standardized mean shift | 0.000000 |
Aligned outcome measurements.
All twenty model and benchmark returns are complete and aligned on the declared periods. Returns compound from wealth 1; running drawdown includes that initial peak. Tracking error uses the sample standard deviation of arithmetic active returns and an explicit 12-period annualization assumption.
| Metric | Result |
|---|
| Model cumulative return | 11.921% |
| Benchmark cumulative return | 9.766% |
| Difference of cumulative returns | 2.155% |
| Mean arithmetic active return / period | 0.100% |
| Annualized tracking error / 12 periods | 0.940% |
| Model maximum drawdown | -1.200% |
| Benchmark maximum drawdown | -0.900% |
What this report cannot establish.
These metrics cannot prove causal attribution, strategy validity, future performance, operational independence or regulatory compliance. Example thresholds are chosen for teaching. A complete period window does not prove the source observations are economically correct. Any real report needs authorized source evidence, ownership, review procedures and durable cloud execution.
Read the metric definitions ↗ · Read the aligned-window guide ↗