Repository navigation
SMA2DynamicDemo
This is an earlier AlgoQuant usage tutorial record. Its numerical results are irrelevant to the teaching purpose. See the maintained tutorial and the learning directory.
Synthetic-data demonstration. Source · Verified execution · Data-and-methodology · Reproduce-the-results
Quick mode recalibrates monthly using a trailing three-month data window, starting with CNY 3300 cash. Each parameter combination receives a fresh simulator and independent measure accumulators. Full mode covers 2010-01-01 to 2017-03-01 and a 4-by-4 grid of rebalance periods (1 week, 2 weeks, 1 month, 3 months) and data windows (1, 2, 3, 6 months). The published run is quick mode, not the full search.
| Item | Value |
|---|---|
| Instrument label | 000001.SS |
| Requested interval (end excluded) | 2016-01-01 to 2016-04-01 |
| Data | Generated weekday OHLCV bars |
| Mode | Offline, headless; script passes --quick (only dynamic demo changes its workload) |
| NAV observations | 65 |
| Executions | 1 |
| P&L | -1.667327 |
| InformationRatioForTrades | not reported |
| OmegaForTrades | NaN |
The run passed the nonempty/finite NAV check. Execution count counts fills; it is not necessarily a count of completed round-trip trades. NAV observations count the report's valuations.
mvn exec:java -Dexec.mainClass=tech.nmfin.algoquant.demo.model.technical.SMA2DynamicDemo -Dexec.args="--quick"A valid external NM license and Maven authentication are required; see Reproduce-the-results.
| Recalibration time (exchange local) | Selected fast lag | Selected slow lag |
|---|---|---|
| 2016-01-04 15:00 | 7 | 20 |
| 2016-02-04 15:00 | 9 | 17 |
| 2016-03-04 15:00 | 3 | 13 |
Captured total and product P&L: -1.6673265308268128. Maximum drawdown: 14.99955558412421; maximum drawdown fraction: 0.004527190212157116 (approximately 0.452719%). Maximum drawdown duration output is 36.0 for the total row and 36.375 for the product row, with a one-day duration unit configured. These are the exporter values as produced.
Information ratios, Omega and win/loss ratios are NaN in this short run. There was only one execution, so these values must not be presented as meaningful risk-adjusted estimates or replaced with zero.
The CSV exports contain performance measures, executions and trades. Permanent copies are under GitHub Release assets; the original simulation-reports artifact includes logs but is subject to GitHub retention.