Data and methodology

Data and methodology

Synthetic inputs

DemoSupport generates one OHLCV bar per weekday at the instrument exchange's closing time. Exchange holidays are not modeled. Requested intervals include their start and exclude their end. The calendar-date formula is identical for every stock:

d = calendar days since 2009-01-01
close = 100 + 0.02*d + 8*sin(d/11) + 3*cos(d/29)
high = close + 1.5 + 0.4*sin(d/7)
low = close - 1.5 + 0.4*cos(d/13)
open = close - 0.2
volume = 1,000,000 + 1,000*(d mod 200)
adjusted close = close

Using calendar dates means overlapping calibration windows get identical prices for the same date. The generated trend and oscillations exercise strategy logic; they do not model real assets, realistic cross-asset correlations, market microstructure, or corporate actions.

SMA examples consume order books adapted from adjusted closes. The ADL and CMF examples also consume full OHLCV events for their volume indicators. --live selects cached Yahoo downloads instead, and is not used in this report.

What was verified

All six executable demos compiled against the public obfuscated combined JAR and ran in separate JVMs without desktop charts or a market-data service. Each returned a nonempty NAV series containing only finite values. Each process had a five-minute timeout. Seven Java source files include the shared helper; there are six simulations.

Interpretation limits

P&L is a monetary amount rather than a return percentage. Trade-level information ratios and Omega are recorded with the library's measure names. Infinity, NaN, and the negative CMF Omega are preserved as emitted, rather than silently normalized or interpreted as conventional portfolio statistics. Read the implementation/API definition before comparing such measures with other systems.

Executions are fills, not completed trades. The dynamic quick run's one execution is insufficient to infer a distribution of completed-trade returns. These smoke runs do not verify realistic costs, slippage, shorting constraints, index tradability, execution timing or absence of look-ahead bias in a real-market research setup. Synthetic results are API examples, not investment backtests.

Home