Strategies

Rules

1. Use point-in-time DJIA membership for every historical signal date.
2. Rank eligible DJIA constituents with the Market Move Forecast 3-month Momentum model.
3. Momentum score = equal-weight average of three consecutive 1-calendar-month cross-sectional percentile scores.
4. Hold a maximum of two stocks.
5. Initial positions target global Momentum ranks #1 and #2.
6. Review the portfolio weekly.
7. Keep an existing holding while it remains inside the global Momentum Top 10.
8. If a holding leaves the global Top 10, sell it at the next US market open.
9. Replacement candidate = the highest-ranked Momentum stock not already held.
10. Allow new purchases only when the DJIA close is at or above its 200-session moving average.
11. When the DJIA is below the 200-session moving average, do not force-sell valid Top-10 holdings; only block new purchases.
12. Monitor the two invested slot values at each weekly signal close. The money-management trigger is reached when the larger slot is at least 1.80 times the smaller slot.
13. Do not sell or trim an otherwise valid Top-10 holding solely because the two slots have become imbalanced; the money-management rule acts only when a normal Momentum exit creates an execution event.
14. If the smaller slot exits while the 1.80:1 imbalance is active, keep its proceeds in cash and wait for the larger slot to exit before rebuilding both slots 50/50, subject to the market filter.
15. If the larger slot exits first while the 1.80:1 imbalance is active and buying is allowed, use available proceeds to move the surviving smaller holding toward 50% of portfolio value and allocate the other 50% to the replacement slot. Any top-up is a new buy order and pays the normal commission.
16. No stop-loss or trailing stop is used in the locked reference; risk exits are driven by the Momentum Top-10 rule and the market filter controls new exposure.

Execution & activity

Backtest methodology

• No look-ahead: rankings, slot imbalance and the market filter use information available on or before the signal date.
• Orders are executed at the next US market open.
• Point-in-time DJIA membership is used to avoid current-constituent survivorship bias.
• Historical EUR/USD conversion is used.
• Fractional shares are allowed.
• Cash earns no interest.
• Transaction cost assumption: EUR 8.50 per executed buy or sell order, including money-management top-ups when they occur.
• DIA is used only as a data proxy for the Dow Jones benchmark in this historical comparison.
• The public backtest window is fixed at exactly nine years: 28 August 2017 through 28 August 2026.

Reference status

• MA200 was retained as the locked market-regime filter after neighboring-length robustness tests.
• A 1.80:1 slot-imbalance threshold was retained after comparison with 1.25, 1.33, 1.50, 1.67 and 2.00 thresholds on 5-year and approximately 10-year windows.
• Active trailing and hard-stop variants were not retained because protection rules that actually triggered did not improve the historical risk/return profile in the tested samples.

Historical EUR/USD conversion

Rules

1. Use point-in-time CAC 40 membership for every historical signal date, with explicit pragmatic exclusions for unavailable legacy price series.
2. Rank eligible CAC 40 constituents with the Market Move Forecast 3-month Momentum model.
3. Momentum score = equal-weight average of three consecutive 1-calendar-month cross-sectional percentile scores.
4. Hold a maximum of two stocks.
5. Initial positions target global Momentum ranks #1 and #2.
6. Review the portfolio weekly.
7. Keep an existing holding while it remains inside the global Momentum Top 10.
8. If a holding leaves the global Top 10, sell it at the next Euronext Paris market open.
9. Replacement candidate = the highest-ranked Momentum stock not already held.
10. Allow new purchases only when the CAC 40 close is at or above its 200-session moving average.
11. When the CAC 40 is below the 200-session moving average, do not force-sell valid Top-10 holdings; only block new purchases.
12. Monitor the two invested slot values at each weekly signal close. The money-management trigger is reached when the larger slot is at least 1.80 times the smaller slot.
13. Do not sell or trim an otherwise valid Top-10 holding solely because the two slots have become imbalanced; the money-management rule acts only when a normal Momentum exit creates an execution event.
14. If the smaller slot exits while the 1.80:1 imbalance is active, keep its proceeds in cash and wait for the larger slot to exit before rebuilding both slots 50/50, subject to the market filter.
15. If the larger slot exits first while the 1.80:1 imbalance is active and buying is allowed, use available proceeds to move the surviving smaller holding toward 50% of portfolio value and allocate the other 50% to the replacement slot. Any top-up is a new buy order and pays the normal commission.
16. No stop-loss or trailing stop is used in the locked reference; risk exits are driven by the Momentum Top-10 rule and the market filter controls new exposure.

Execution & activity

Backtest methodology

• No look-ahead: rankings, slot imbalance and the market filter use information available on or before the signal date.
• Orders are executed at the next Euronext Paris market open.
• Point-in-time CAC 40 membership is used to avoid current-constituent survivorship bias; identities without reliable legacy OHLC remain in membership but are excluded from ranking until valid data exist.
• All strategy and benchmark prices are EUR-denominated, so no FX conversion is applied.
• Fractional shares are allowed.
• Cash earns no interest.
• Transaction cost assumption: EUR 8.50 per executed buy or sell order, including money-management top-ups when they occur.
• Amundi CAC 40 UCITS ETF Dist (CAC.PA) is used only as a tradable data proxy for the CAC 40 benchmark in this historical comparison.
• The public backtest window is fixed at exactly nine years: 28 August 2017 through 28 August 2026.

Reference status

• MA200 is reused unchanged from the locked DJIA reference; no CAC-specific optimization is performed.
• The 1.80:1 slot-imbalance threshold is reused unchanged from the locked DJIA reference; no CAC-specific optimization is performed.
• No stop is used, matching the locked DJIA reference; this CAC publication is a direct out-of-sample transfer of the same rules rather than a CAC-specific optimization.
• This CAC40 publication reuses MA200, Top10 retention, MM 1.80:1, two slots and no-stop rules unchanged from the DJIA reference; no CAC-specific parameter search was performed.

EUR-denominated / no FX conversion

Rules

1. Use point-in-time EURO STOXX 50 membership for every historical signal date.
2. Rank eligible constituents with the Market Move Forecast 3-month Momentum model.
3. Momentum score = equal-weight average of three consecutive 1-calendar-month cross-sectional percentile scores.
4. Hold a maximum of three stocks.
5. Initial positions target global Momentum ranks #1, #2 and #3.
6. Review the portfolio weekly.
7. Keep an existing holding while it remains inside the global Momentum Top 27.
8. If a holding leaves the global Top 27 or becomes unranked, sell it at the next available local-market open.
9. Replacement candidate = the highest-ranked eligible Momentum stock not already held.
10. Allow new purchases only when the EURO STOXX 50 price-index close is at or above its 180-session moving average.
11. When the EURO STOXX 50 is below MA180, do not force-sell valid Top-27 holdings; only block new purchases.
12. When multiple slots are empty and buying is allowed, divide available cash equally among the empty slots.
13. No slot money-management overlay, rebalance trigger or top-up rule is used.
14. No stop-loss or trailing stop is used; exits are driven by the Momentum Top-27 rule and the market filter controls only new exposure.
15. For the fixed historical research window, all remaining holdings are liquidated at the final 28 August 2026 close and normal sell commissions are charged.

Execution & activity

Backtest methodology

• No look-ahead: rankings and the MA180 market filter use information available on or before the signal date.
• Orders are executed at the next available local-market adjusted open after the signal.
• Point-in-time pragmatic EURO STOXX 50 membership is used to reduce survivorship bias.
• Historical membership is retained even when a legacy price series cannot be recovered reliably; such securities remain non-scored until reliable data are available.
• Validated same-security continuities may be used, but successor-security price splicing is not allowed.
• Total-return adjusted opens are used for execution and adjusted closes for portfolio valuation.
• Historical research uses fractional shares; the LIVE model launched in September 2026 uses whole shares and is reported separately.
• All strategy and benchmark instruments are EUR-denominated in this historical comparison; no historical EUR/USD conversion is used.
• Cash earns no interest.
• Transaction cost assumption: EUR 8.50 per executed buy or sell order.
• EXW1.DE is used only as the EURO STOXX 50 benchmark proxy in this historical comparison.
• The public research window is fixed at 28 August 2017 through 28 August 2026.

Reference status

• Top27 emerged as the strongest stable retention region in the later robustness phases rather than from a single best-CAGR configuration.
• MA179, MA180, MA181 and MA182 were economically identical across every Phase-6 CORE cost-by-start stress; MA180 was selected as the central round reference.
• The no-money-management version was preferred because it removed a parameter while retaining a 100% full-period benchmark-beat rate in the Phase-5 execution-delay and stale-signal tests.
• All 17 baseline independent start dates in Phase 6 beat the benchmark.
• Across the Phase-6 CORE cost-by-start matrix, the benchmark-beat rate was 99.4%.
• The sole unique CORE miss was the deliberately harsh EUR 50 per order plus 20 bps slippage-per-side scenario starting 28 May 2021; the strategy still delivered 11.12% CAGR versus 12.23% for the benchmark, an excess of -1.11 points.
• Phase-5 full-period anti-timing benchmark-beat rate for the simple no-MM configuration: 100%.
• The configuration is locked as historical reference V1; parameter changes require a new research version rather than silently rewriting this record.

EUR-denominated / no FX conversion