L5 Portfolio
Multi-Symbol Portfolio Backtest
Score-weighted blend of the top radar-eligible strategy/symbol pairs. Legs share a calendar via date intersection and are combined in $1-unit returns space, then scaled to the initial capital.
score-weightedbuy-and-hold within legsbenchmark SPYAll legs use real Yahoo daily OHLCV data.
Methodology / Assumptions
- Weights are proportional to radar score, with equal-weight fallback if scores are not usable.
- Effective bets and correlation are computed from daily return series on the shared calendar.
- Benchmark comparison uses SPY or the first available substitute on the same dates.
- Portfolio results are historical research diagnostics and do not imply future P&L.
+14.3%
+4.6%
-67.2%
2.10
2.1%
-1.5%
0.15
2.8
Portfolio vs SPY
Loading chart
Daily return correlation
| Leg | AAPL | WMT | CAT | CAT |
|---|---|---|---|---|
| AAPL | 1.00 | 0.09 | -0.01 | 0.07 |
| WMT | 0.09 | 1.00 | 0.10 | 0.07 |
| CAT | -0.01 | 0.10 | 1.00 | 0.59 |
| CAT | 0.07 | 0.07 | 0.59 | 1.00 |
Leg contributions
| Symbol | Strategy | Weight | Annualized | Sharpe | Max drawdown | Contribution |
|---|---|---|---|---|---|---|
| AAPL | Defensive Trend Pullback | +26.7% | +3.9% | 1.65 | -2.1% | +3.2% |
| WMT | Low-Volatility Rotation | +25.4% | +4.4% | 1.32 | -3.5% | +3.4% |
| CAT | EMA Continuation Signal | +24.4% | +5.8% | 1.16 | -5.7% | +4.5% |
| CAT | Quality Momentum Breakout | +23.5% | +4.3% | 1.42 | -2.0% | +3.2% |
Methodology
- • Eligible legs: radar candidates and continue-observing strategies with Sharpe > 0.5, capped at four.
- • Legs the concentration gate flagged as near-duplicates (> 80% return correlation with a higher-ranked leg) are excluded — the portfolio reuses the same diversification decision as the paper-trade queue.
- • Effective bets = N / (1 + (N−1)·ρ̄), the same N_eff measure shown on the radar panel and home overview.
- • Weights are proportional to the composite radar score; falls back to equal-weight if all scores are non-positive.
- • Calendar = intersection of every leg's equity-curve dates so the blend uses only days when all legs traded.
- • Each leg's equity is normalized to a $1 base; the portfolio re-aggregates daily — no intra-day rebalancing assumption.
- • Benchmark is SPY (or the first available substitute) sampled on the same calendar.
- • Correlation cells use Pearson on daily returns; the off-diagonal average drives the diversification badge.