How the Kiploks Robustness Score Works

1. Formula (multiplicative)

Score = (Validation^0.4 × Risk^0.3 × Stability^0.2 × Execution^0.1) × DQG × 100. Each module is a 0–100 value; DQG ([Data Quality Guard]) is a gate (0 or 1). The exponents reflect weights: Validation 40%, Risk 30%, Stability 20%, Execution 10%. DQG does not have a weight share but multiplies the product: if DQG fails, overall = 0.

2. Blocking thresholds

If any module is below its blocking threshold, the overall score is forced to 0 and the block shows "(FAIL)" and "Blocked by …":

• [Data Quality Guard] (DQG): threshold 0. If test period < 365 days or DQG fails, overall = 0.
• [Validation (Walk-Forward & OOS)]: threshold 0. WFE, consistency, failed windows; if validation fails, overall = 0.
• Risk Profile: threshold 0. Profit factor, kurtosis, recovery factor; if risk module fails, overall = 0.
• [Parameter Stability]: threshold 0. Fragile parameters, edge t-stat; if stability fails, overall = 0.
• Execution Realism: threshold 10. Slippage tolerance at 10 bps. If execution score < 10, the edge does not survive 10 bps slippage and overall = 0.

3. Breakdown (contributing factors)

The block shows each module with its weight, a 0–100 value, and a short verdict (e.g. "→ Adequate data period", "→ BLOCKED"). Rows that are blocking are highlighted (red border). [Data Quality Guard] has no weight percentage but acts as a gate; the other four have 40%, 30%, 20%, 10%.

4. Methodology note

Weights (40 / 30 / 20 / 10) are optimized via cross-validation on 100+ public strategies to maximize the correlation between the Kiploks Score and future out-of-sample performance. Validation (40%) is the primary lead indicator of strategy survival.

5. Deployment rule

Score < 60 = deploy forbidden. A score of 60 or above does not automatically mean "deploy"; it means the strategy has passed the robustness bar. The [final verdict] (ROBUST /CAUTION / DO NOT DEPLOY) and the [Strategy Action Plan] use this score together with other checks.

6. Diagnostics when blocked

When the score is blocked, the block may show a Diagnosis box:

• Execution diagnostic: If Execution is the blocking module and its value is below 10, the text explains that the edge does not survive 10 bps slippage and suggests reviewing costs, turnover, or edge.
• Structure diagnostic: If blocked by Validation/Risk/Stability but [Data Quality] and[Parameter Stability] are high, the text explains that the problem is lack of statistical edge, not data or overfitting.