Chart & Technical · COST
Strategy Cup Tier for COST
AI-generated plain-English read of strategy cup tier for COST. Updated daily.
Strong tier — borderline validation
lowStrong val-window result (an unknown margin) with borderline statistical support. Solid edge in the numbers, but verification is incomplete — follow with discipline, not conviction.
What will probably happen
- The edge is above average but borderline — treat entries as tentative until the next val window confirms.
- Walk-forward stability is high — the edge held up across multiple rolling out-of-sample windows, reducing the chance of parameter overfitting.
- Fewer than 10 trades — a single trade's outcome can dominate the result. Don't trust aggregate statistics from this sample.
What to do
- If you HOLD this stock and follow this strategy
- Strong tier but incomplete validation — hold with reduced position size. Set a drawdown trigger to re-evaluate if performance decays.
- If you're EVALUATING whether to use this strategy
- Borderline validation — paper-trade for 3-6 months before allocating. If the edge persists on the next val window, add capital.
- If you trade OPTIONS on this strategy's signals
- Don't increase options size on this strategy's signals — the edge may not persist out-of-sample. Treat signals as weak directional hints only.
Weak credibility profile — fewer than 10 trades — sample is too small.
Technical details
Technical analysis
- Cup tier
- cup_3 (strong)
- Robustness
- Stable
- Tier grade
- Strong
- Trade count
- 2
- MC significant
- No
- Sample quality
- Thin Sample
- Validation window
- —
- Statistical support
- Borderline
- Deflated Sharpe Ratio (DSR)
- 1.000
- Delta return vs buy-and-hold
- —
- Walk-forward stability score
- 1.09
Key levels
| Level | Value |
|---|---|
| dsr | 1 |
| cup tier | 3 |
| num trades | 2 |
| robustness score | 1.086 |
Caveats
- Cup tiers reflect val-window performance. Live forward returns may differ from backtest — even validated strategies have drawdown risk.
- Fewer than 10 trades — single-trade outcomes can dominate the result. Aggregate statistics (Sharpe, delta_return) are unreliable at this sample size.