10-Year Demo

MEΩ Backtest Simulator

A reproducible simulator measuring portfolio performance in the MEIc numeraire?returns are currency-neutral and inflation-aware. Tune trading costs and rebalancing cadence; outputs follow the explicit equations below.

Demo returns are synthetic; no external data is fetched. Snapshot date is fixed at 2025-10-08.

Simulation Parameters

Upper bound on projected fee + market impact per trade, plus the trading clock for policy decisions.

Last run: 02:59

Cost Cap (bp)60 bp

Upper bound on projected fee + market impact per trade. If the bound is breached, the trade is skipped (cost gate).

Rebalance Period7d

Trading clock for policy decisions (weekly in this demo). Adjust to explore cadence sensitivity.

Current Settings

Period:2014-11 to 2024-11

Policy:Weekly re-weighting

Risk gate:Median-σ (see §3.2)

Cost cap:60 bp

Performance (MEΩ-native)

Live from controls

GINI™ (annualized)157 bp

Total Return (MEΩ)10.8%

Sharpe (weekly → annualized)3

Max Drawdown-0.8%

Trade Stats

Total rebalances521

Skipped (Cost Gate)23

Avg cost per trade18.7 bp

Total cost impact-93 bp

Cumulative Returns

MEΩ baseline

USD (with its own drift)

Rounded to two decimals to avoid long binary fractions (e.g., -9.99% instead of -9.999999999999998%).

Methodology (precise but readable)

Demo uses synthetic but structured signals so graphs respond to the controls while preserving realistic cost/vol behavior. In offline mode (MEO_OFFLINE=1) the weighting snapshot is pinned to 2025-10-08 from the bundled CSV/JSON; no live feeds are queried here.

r_{i,\ME\Omega}(t) = \Delta \ln\!\left(\dfrac{P^{\USD}_i(t)}{P^{\ME\Omega}_{\USD}(t)}\right)

Copy: r_i_MEΩ = Δ ln(P_i_USD / P_USD^{MEΩ})

r_{p,\ME\Omega}(t_k) = \sum_i u_i(t_{k-1})\, r_{i,\ME\Omega}(t_k)

Copy: r_p_MEΩ(t_k) = Σ_i u_i(t_{k-1}) r_i_MEΩ(t_k)

\mathrm{d}w_j = w_j\!\left[r^{(\lambda)}_j - \sum_k w_k\, r^{(\lambda)}_k \right]\mathrm{d}t,\quad r^{(\lambda)}_j=\mathrm{clip}\!\big(r_j,-5\sigma_j,+5\sigma_j\big)

Copy: dw_j = w_j [ r_j(λ) - Σ_k w_k r_k(λ) ] dt

c_{bp}(t_k) = f + 10000 \cdot \gamma \sqrt{\dfrac{|\Delta q(t_k)|}{\mathrm{ADV}_{10}(t_k)}},\qquad c_{bp}(t_k) > \mathrm{Cap}_{bp} \Rightarrow \text{skip trade}

Copy: c_bp = f + 10000·γ·sqrt(|Δq|/ADV10); skip if c_bp > Cap_bp

Invariants to assert after each run:

\big|\sum_j w_j(t) - 1\big| < 10^{-9}

Copy: |Σ w_j - 1| < 1e-9

\dfrac{\left|P_{\USD}^{\ME\Omega}(t) - \kappa \sum_j MC^{\USD}_j(t)\right|}{P_{\USD}^{\ME\Omega}(t)} < 10^{-6}

Copy: |P_USD^{MEΩ} - κ Σ MC_j^{USD}| / P_USD^{MEΩ} < 1e-6

Mathematical Appendix (for quants)

P^{\ME\Omega}_{\USD}(t) = \kappa \sum_{j \in \mathcal{C}} MC^{\USD}_j(t),\quad \kappa = 10^{-6}

Copy: P_USD^{MEΩ} = κ Σ_j MC_j^{USD}

w_j(t) = \dfrac{MC^{\USD}_j(t)}{\sum_{k} MC^{\USD}_k(t)},\qquad \sum_j w_j(t)=1

Copy: w_j = MC_j^{USD} / Σ_k MC_k^{USD}

Risk & cost model options

Vol: realised σ (window W) or GARCH(1,1) on $r_{i,\ME\Omega}$ .
Tails: EVT (POT–GPD) for VaR/CVaR in MEΩ units.
Cost: commission $f$ (bp) + impact $\,\,10000\cdot\gamma\sqrt{|\Delta q|/\mathrm{ADV}_{10}}$ ; gate when $f + \mathrm{impact}_{bp} > \mathrm{Cap}_{bp}$ .

Disclosure

Data proxy: stylized signals inspired by FRED, LBMA, CoinGecko behaviors; math and cost/risk machinery are real.

No advice: educational use only. Export the trade log and report to reproduce aggregates within rounding tolerance.