Skip to content

Releases

OpenRiskSim 0.2.0

Choose a tag to compare

@fernandevdaza fernandevdaza released this 07 Oct 13:34

🌐 App: https://fernandevdaza.github.io/openrisksim/ · 📖 Docs: https://fernandevdaza.github.io/openrisksim/docs/

⬇️ Descargar / Download

Sistema / System Archivo / File
macOS · Apple Silicon (M1–M5) OpenRiskSim-0.2.0-mac-arm64.dmg
macOS · Intel OpenRiskSim-0.2.0-mac-x64.dmg
Windows · x64 OpenRiskSim-0.2.0-win-x64-setup.exe
Windows · ARM64 OpenRiskSim-0.2.0-win-arm64-setup.exe
Linux · x64 OpenRiskSim-0.2.0-linux-x86_64.AppImage · OpenRiskSim-0.2.0-linux-amd64.deb
Linux · ARM64 OpenRiskSim-0.2.0-linux-arm64.AppImage · OpenRiskSim-0.2.0-linux-arm64.deb

Las apps aún no están firmadas · The apps are not code-signed yet. macOS: clic derecho → Abrir la primera vez (o xattr -cr /Applications/OpenRiskSim.app). Windows: SmartScreen → Más información → Ejecutar de todas formas.

Added

  • Desktop apps for macOS (dmg, Apple Silicon + Intel), Windows (installer, x64 + ARM64) and Linux (AppImage + deb,
    x64 + ARM64), built with Electron; installers are attached to each release and built by CI on every push.
  • Excel-like formula editing: click-to-reference (Point mode) with coloured reference boxes, F4, localized function
    autocomplete with argument tooltips, Insert Function dialog, fill handle and Ctrl+D/R/Enter, copy/cut/paste with
    reference adjustment and paste values, Formulas tab (AutoSum, show formulas, trace precedents/dependents, evaluate
    formula, defined names), explained error cells and syntax checks, expandable formula bar.
  • Tornado / spider: "± % of base value" method (Risk Simulator style) and grouping of assumptions with the same name.
  • Mobile layout: scrollable ribbon, bottom navigation with run button, model drawer, full-screen results, dialogs
    and tools; touch-friendly grid.
  • Documentation: new Formulas guide; tornado methods and why rankings can differ from Risk Simulator.

OpenRiskSim 0.1.0

Choose a tag to compare

@fernandevdaza fernandevdaza released this 06 Oct 01:10

🌐 App: https://fernandevdaza.github.io/openrisksim/ · 📖 Docs: https://fernandevdaza.github.io/openrisksim/docs/

Primera versión pública de OpenRiskSim: simulación de riesgo Monte Carlo para tus hojas de cálculo, en el navegador — una alternativa libre a Risk Simulator / @risk / Crystal Ball.
First public release: Monte Carlo risk simulation for your spreadsheets, in the browser.

First public release.

Added

  • Web app (PWA) that opens and saves .xlsx / .csv with an Excel-compatible formula engine (HyperFormula), including
    Spanish function names and decimal commas. The risk model is stored inside the workbook.
  • Monte Carlo and Latin Hypercube simulation with 33 distributions, truncation, rank correlations (Iman–Conover),
    reproducible seeds, precision control, step mode and a Web Worker runner.
  • Forecast charts with two-tail / left-tail / right-tail certainty, full statistics, percentiles, overlay charts,
    PNG export and report export (.xlsx and printable PDF).
  • Analytical tools: tornado, spider, sensitivity, scenario table, distribution fitting (KS, AD, χ², AIC/BIC),
    bootstrap, hypothesis test, descriptive statistics, overlay chart.
  • Forecasting: moving average, exponential smoothing, Holt, Holt–Winters, ARIMA / auto-ARIMA, trend models,
    multiple and stepwise regression, stochastic processes (GBM, mean reversion, jump diffusion).
  • Optimization: static and stochastic, Nelder–Mead, genetic algorithm, simulated annealing, efficient frontier.
  • Finance: project evaluator (project and equity cash flows, NPV, IRR, MIRR, payback, PI, EAA) with export to a sheet
    with live formulas, NPV/IRR calculator, loans, depreciation, WACC/CAPM, break-even, scenario analysis.
  • Optional acceleration (Simulation → Settings → Acceleration): multi-core CPU, formulas compiled to exact JavaScript,
    and a WebGPU compute-shader backend (Metal / Direct3D 12 / Vulkan). Accelerated runs are validated against the
    spreadsheet engine and fall back automatically; GPU trials that fail in f32 are re-evaluated exactly on the CPU.
  • Configurable confidence levels: per-forecast certainty band and confidence of the mean; up to three
    prediction-interval levels in time series, confidence level for regression coefficients, fan-chart levels for
    stochastic processes.
  • Five example models, English/Spanish interface, light/dark themes, autosave.
  • Documentation site (Spanish and English).

Example workbooks are attached below (also in examples/).