P(A∩B) = P(A)P(B|A) E[X] = Σx·P(x) Var(X) = E[X²] - (E[X])² Z = (X - μ) / σ χ² = Σ(O-E)²/E F = s₁²/s₂² t = (x̄ - μ)/(s/√n) r = Cov(X,Y)/(σₓσᵧ)
RSI = 100 - 100/(1+RS) MACD = EMA₁₂ - EMA₂₆ %K = (C-L₁₄)/(H₁₄-L₁₄)×100 ATR = (1/n)Σ|H-L| CCI = (TP - SMA)/(.015×MD) Williams %R = (H₁₄-C)/(H₁₄-L₁₄) ROC = (P₁ - Pₙ)/Pₙ × 100 OBV = ΣV × sign(ΔP)
ROE = NI / SE ROA = NI / TA D/E = TD / TE P/E = Price / EPS EV/EBITDA = EV / EBITDA FCF = OCF - CapEx ROIC = NOPAT / IC EPS = (NI - PD) / WAS
σₚ² = w'Σw SR = (Rₚ - Rf) / σₚ β = Cov(Rᵢ,Rₘ) / Var(Rₘ) IR = αₚ / σ(εₚ) VaR = μ - zα × σ ES = E[R | R ≤ VaR] MDD = max(DD) Calmar = CAGR / MDD
PV = FV / (1+r)ⁿ FV = PV × (1+r)ⁿ PMT = PV × [r(1+r)ⁿ]/[(1+r)ⁿ-1] NPV = Σ[CFₜ/(1+r)ᵗ] - I₀ IRR: 0 = Σ[CFₜ/(1+IRR)ᵗ] MIRR = ⁿ√(FVₚ/PVₙ) - 1 PI = PV(CI) / PV(CO) DPP = I₀ / CF̄
C = S₀N(d₁) - Ke⁻ʳᵀN(d₂) P = Ke⁻ʳᵀN(-d₂) - S₀N(-d₁) Δ = ∂V/∂S Γ = ∂²V/∂S² θ = ∂V/∂t ν = ∂V/∂σ ρ = ∂V/∂r λ = ∂V/∂q
Y = α + βX + ε R² = 1 - SSR/TSS β̂ = (X'X)⁻¹X'Y DW = Σ(eₜ-eₜ₋₁)²/Σeₜ² AIC = 2k - 2ln(L) BIC = k·ln(n) - 2ln(L) JB = (n/6)[S² + (K-3)²/4] LM = n·R²
ARIMA(p,d,q) φ(L)Δᵈyₜ = θ(L)εₜ ACF(k) = γₖ/γ₀ PACF = φₖₖ GARCH: σₜ² = ω + αε²ₜ₋₁ + βσ²ₜ₋₁ EGARCH: ln(σₜ²) = ω + α|εₜ₋₁| VAR: Yₜ = Φ₁Yₜ₋₁ + εₜ VECM: ΔYₜ = αβ'Yₜ₋₁ + εₜ
MSE = (1/n)Σ(yᵢ - ŷᵢ)² RMSE = √MSE MAE = (1/n)Σ|yᵢ - ŷᵢ| R² = 1 - SS_res/SS_tot Precision = TP/(TP + FP) Recall = TP/(TP + FN) F1 = 2·(P·R)/(P + R) AUC-ROC = ∫TPR dFPR
RWA = Σ(Aᵢ × RWᵢ) Tier 1 = CET1 + AT1 LCR = HQLA / NCOS NSFR = ASF / RSF PD = P(default) LGD = 1 - RR EAD = CCF × UL + DB EC = EAD × PD × LGD
QuantPerks
Market Closed

Portfolio Analyzer

Advanced optimization with Monte Carlo forward-looking projections

Clean Failure Mode • No Look-Ahead Bias • CVaR/EUP Daily | Returns/Vol Annual

Configuration

Machine learning clustering approach - No assumptions about future returns, naturally diversified.

Click to optimize your portfolio using the selected method and parameters.

Portfolio Optimization Ready

Configure parameters and choose from 7 robust optimization methods with optional Monte Carlo forward-looking projections.

Best Practice: Use Portfolio Analyzer data only up to a specific point for optimization, then evaluate the remaining period for out-of-sample validation. This prevents overfitting and provides realistic performance assessment.
Clean Failure Mode: No synthetic fallbacks or data manipulation
Real Data Only: FMP API • No synthetic data
Anti Look-Ahead: All data automatically bounded to prevent future bias