Engine C - Initial Candidate

ZenWave C 1

First candidate readout for Engine C. The overlay benchmarks its return path against ZenWave A and ZenWave B, then the sections below test portfolio impact, stress behavior, and independence before any deployment decision.

Full-backtest relative overlay (C vs A/B).

SeriesWin % / YearMedian YearWorst YearWorst 12M% Below High
ZenWave C 192%(12/13)+33.06%-0.75%-8.97%89.1%
Win RateProfit FactorSharpeReturn / DDCalmarExp/TradeDD AveDD MaxDT AveDT MaxTrades
ZenWave A17.0%1.982.18134578.093.720.36%7.55%37.37%21.8d498d3,330
ZenWave B16.3%1.862.0670827.433.050.35%7.77%42.34%20.2d289d3,245
ZenWave C 139.0%1.522.23338.161.780.24%4.35%23.84%20.0d392d1,956
DD Duration FilterShowing 28 DD episodes of exactly 1d (1/98)

Book Simulation

Trade-level compounded portfolio simulation. The current A+B book (each at 0.25%) is shown alongside a hypothetical A+B+C book using the same 0.25% per-engine sizing.

BookWin % / YearMedian YearWorst YearWorst 12M% Below High
Book (A+B @0.25%)100%(13/13)+44.43%+1.09%-3.63%90.2%
Book + ZenWave C 1 (A+B+C @0.25%)100%(13/13)+54.61%+7.29%-3.63%89.5%
WRPFSRR/DDCalmarExp/TradeDD AveDD MaxDT AveDT MaxN
Book (A+B @0.25%)16.4%1.691.253586.863.620.10%3.23%19.15%17.1d289d6,863
Book + ZenWave C 1 (A+B+C @0.25%)21.4%1.641.249463.913.870.09%3.12%21.96%14.5d289d8,819

Monte Carlo Stress Distribution

Synthetic forward paths generated from the candidate’s historical trade distribution.

Equity Path Simulation

10,000 paths · 1,224 trades resampled
Full BT Median 25th–75th 5th–95th

Median Max Drawdown

24.8%

95th pctl: 37.1%

Original Max DD

26.8%

Backtest (single path)

Median Final Multiple

×29.2

5th–95th: ×8.69 – ×99.8

Original Final Multiple

×29.1

Backtest (single path)

Drawdown Probability

>20%

83.8%

>30%

21.2%

>40%

2.4%

>50%

0.1%

1,224 trades · 10,000 paths · bootstrap with replacement · seeded PRNG for reproducibility

Drawdown Comparison

Each series rebased against its own running peak. Shows how the candidate’s drawdown profile lines up with ZenWave A and ZenWave B over the same period.

ZenWave A DDZenWave B DDZenWave C 1 DD

Independence vs ZenWave A

Correlation, sensitivity, and stress-regime co-movement of the candidate relative to ZenWave A.

Daily return relationship with ZenWave A is near zero, indicating independence.

Correlation

+0.168

Pearson correlation of overlapping daily returns with ZenWave A

Sensitivity to ZenWave A is near zero, indicating very limited directional dependence.

Beta

+0.073

Sensitivity of returns to ZenWave A returns

Stress-regime relationship with ZenWave A remains limited, indicating independence during ZenWave A drawdowns.

Drawdown Correlation

+0.156

Correlation during periods when ZenWave A is below its rolling max

Tail-event relationship with ZenWave A turns positive, indicating some sensitivity on the worst ZenWave A days.

Worst 10% ZenWave A Days

+0.056

Correlation on the most negative decile of ZenWave A daily returns

Independence vs ZenWave B

Correlation, sensitivity, and stress-regime co-movement of the candidate relative to ZenWave B.

Daily return relationship with ZenWave B is near zero, indicating independence.

Correlation

+0.164

Pearson correlation of overlapping daily returns with ZenWave B

Sensitivity to ZenWave B is near zero, indicating very limited directional dependence.

Beta

+0.070

Sensitivity of returns to ZenWave B returns

Stress-regime relationship with ZenWave B remains limited, indicating independence during ZenWave B drawdowns.

Drawdown Correlation

+0.133

Correlation during periods when ZenWave B is below its rolling max

Tail-event relationship with ZenWave B turns positive, indicating some sensitivity on the worst ZenWave B days.

Worst 10% ZenWave B Days

+0.014

Correlation on the most negative decile of ZenWave B daily returns

Research Profiles

Exploration of multiple optimization profiles derived from the same underlying signal (Engine A), evaluating how different objectives shape return, drawdown, and consistency for future deployment.

Current work: deep optimization of the Consistency Profile. Other profiles will follow.

Engine A V2 Profile Exploration vs Live Baseline (Frozen Reference)

All profiles shown here are derived from the same core signal (Engine A). Entries and signal generation remain unchanged. Each profile represents a different optimization objective applied to trade management - including return maximization, drawdown control, and consistency improvement.

Initial results shown here are based on early-stage parameter exploration. Current work is focused on further developing the Consistency Profile, with similar depth of optimization planned for the other profiles.

ProfileWin % / YearMedian YearWorst YearWorst 12M% Below High
A V1 – Live Baseline (Frozen Reference)Live Reference92%(11/12)+51.43%-8.00%-19.52%91.5%
A V2 – Consistency ProfileWin Stability FocusIN ACTIVE RESEARCH100%(12/12)+77.97%+5.10%-19.42%91.9%
A V2 – Return ProfileGrowth Focus100%(12/12)+57.86%+4.78%-23.25%91.2%
A V2 – Risk ProfileDrawdown Control100%(12/12)+73.36%+13.03%-11.66%90.9%

Profiles currently reflect early optimization passes (e.g. limited parameter sets). Each profile will undergo further refinement before any deployment decisions are made.

Win RateProfit FactorSharpeReturn / DDCalmarExp/TradeDD AveDD MaxDT AveDT MaxTrades
A V1 – Live Baseline (Frozen Reference)17.0%1.982.23124767.593.900.37%7.43%37.37%18.5d498d3,330
A V2 – Consistency Profile19.2%2.152.1995356.134.630.34%6.84%29.29%17.8d430d3,379
A V2 – Return Profile17.4%2.082.30161430.093.990.38%6.86%37.98%17.0d370d3,325
A V2 – Risk Profile18.3%2.132.3483664.974.790.35%6.54%27.49%16.5d339d3,256
DD Duration FilterShowing 38 DD episodes of exactly 1d (1/96)

Profiles are optimized under different objective functions. Results reflect how trade management affects outcome distribution, not changes to the underlying signal.

All profiles are part of ongoing research. Current focus reflects development sequence, not final preference.