C

Adding C Longs to A+B

Portfolio ContributionAs ConfiguredBacktested

How did the tested long-only C configuration affect an equal-capital A+B portfolio?

Annual €300 reports are stitched separately for each engine, using live-condition files where available (2021–2025), otherwise demo (2014–2020). Live-condition means simulated execution, not observed live trading.

A+B starts with 50% in each engine; A+B+C longs starts with one-third each. Allocations reset at the start of the selected window, then drift with returns. Each engine’s annual return is compounded into its next annual run. There is no risk scaling or portfolio rebalancing: differences in the engines’ configured risk remain in the comparison.

The shared annual €300 coverage ends in 2025. C’s 2026 report is excluded because comparable A/B annual €300 reports are not available. A/B’s live-balance-rebased FPT reports are separate from this study.

C is planned to launch longs only in January 2027. The decision below documents the evidence, its limits and the pre-launch conditions; portfolio capital allocation and Book risk weighting remain separate.

Research decision · reviewed 2026-11-05

Proceed toward a monitored C-long launch with frozen v1 parameters

Target: January 2027, longs only. The published evidence supports controlled forward deployment, subject to final configuration, sizing and execution checks. The decision concerns C as an independent engine; its Book risk weight and portfolio capital allocation remain separate decisions.

Why proceed

  • C2–C4 are profitable out of sample: PF 1.56 / 1.33 / 1.64. C1 is breakeven at PF 1.00.
  • Adding C lowered maximum closing-balance drawdown in 12/12 individual-year comparisons. Full-history weekday correlation with A+B is 0.15.
  • In 2025, A+B returned 61.69% versus 61.86% with C; maximum closing-balance drawdown fell from 22.47% to 18.40%.

Limits of that conclusion

  • Return improved in 4/12 individual years. Without rebalancing, C’s influence can shrink as A/B compound faster. Equal capital does not mean equal risk.
  • The seeded 10,000-path bootstrap puts 95th-percentile maximum balance drawdown at 33.74%; 12.03% of paths exceed 30%. These simulations use the €1,000 reference configuration, not a direct forecast for a €300 account.
  • C3/C4 OOS periods overlap; the annual frozen-configuration reruns are descriptive backtests. There is no dedicated frozen-v1 C-long live track record yet. LAB and FTMO configurations are separate.
Conditions to complete before launch
  • Record the code and parameter fingerprint of the tested long-only v1 configuration. Freeze that artifact before deployment; later strategy changes require a new research version.
  • Document the intended capital, risk limits and actual minimum-lot exposure with the target broker’s costs. The €300 annual reports and €1,000 stress reference do not establish identical realized risk.
  • Verify order direction, stop protection, monitoring, execution logging and protective shutdown procedures. Confirm the January launch date after this review; the site’s launch plan is not evidence that these checks have passed.
  • Preserve frozen strategy logic during forward tracking. Document execution failures and protective actions; review deviations against the recorded references rather than re-optimizing to recent outcomes.

This assessment uses the published historical tests and 2014–2025 portfolio record, independent of the chart window selected below. Bootstrap paths independently resample historical returns without preserving regime clustering; they do not establish a future drawdown limit.

Review WFA and Monte Carlo evidence →

Past performance, including backtested and simulated results, does not guarantee future returns. Trading leveraged instruments involves risk of loss.

A+B versus A+B+C longs

Combined ROI previewReturn history. Interactive chart controls become available after the page loads.
SRR/DDCalmarDDDD AveDD MaxDTDT AveDT MaxROI
A+B · 50/502.352228.4010.224.98%5.16%24.86%36d22.3d182d+55397.94%
A+B+C longs · ⅓ each2.361497.959.164.93%4.76%24.77%36d21.9d182d+37104.33%
A+B · 50/50A+B+C longs · ⅓ each

2021-01-01 → 2025-12-31. Curve statistics use closing balances, including drawdown depth and duration. All returns are backtested.

Does C help when A+B loses?

C / A correlation

0.17

C / B correlation

0.13

C / A+B correlation

0.16

A+B losing days improved by C

570 / 641

Mean A+B return on its losing days

-0.88%

Mean A+B+C return on those days

-0.83%

Pearson correlations of weekday closing-balance returns; missing trade-event dates carry the latest balance. “Improved” means A+B+C lost less or gained on a day when A+B lost. This does not require C itself to be profitable that day.

Individual engines · same window

Combined ROI previewReturn history. Interactive chart controls become available after the page loads.
SRR/DDCalmarDDDD AveDD MaxDTDT AveDT MaxROI
A2.071076.868.118.80%6.24%25.67%30d28.1d237d+27642.99%
B2.253396.7711.615.90%5.61%24.48%40d21.6d182d+83152.90%
C longs1.7730.542.600.00%3.90%16.93%0d40.7d303d+517.09%
ABC longs