Skip to content

ArchLabWorks/QUANTXT

Repository files navigation

QUANTXT

Screenshot

Screenshot: QUANTXT intro screen running on an IBM 5160 (8087 FPU), displayed on an IBM 5153 Color Display.

Sovereign Risk Modeling Engine
C89 Standard· Watcom Open C V1.9 · Real-mode DOS · Intel 8087 FPU


Releases

Latest Full Release can be found here


Built to answer a real question: given current US sovereign conditions, what does the data say the stress probability is? The answer runs on a 1981 coprocessor. Calibrated against 60 scenarios — MSE 0.0038, RMSE 0.062 on a [0,1] scale.


QUANTXT — TL;DR

QUANTXT is a fully self‑contained macro‑risk modeling engine built in pure ANSI C89 for real DOS hardware.
It loads economic scenarios, runs a deterministic 13‑factor stress model, applies nonlinear system transforms, and renders results in a clean XT‑style dashboard UI.

  • Two‑pass stress engine Explain QUANTXT stress modules with damping, smoothing, and regime classification
  • Deterministic outputs — no randomness, no external dependencies
  • Runs on DOS, DOSBox, and retro toolchains (build instructions)
  • BIOS‑text dashboard with centered layout and color‑coded risk bars
  • Lightweight, modular architecture (state.h) as the shared root, no cycles
  • Educational reference implementation showing how modern risk concepts fit into 1980s constraints

QUANTXT is designed to be transparent, inspectable, and easy to reason about — a modern macro‑risk engine expressed through the simplicity of DOS‑era software engineering.


The Full Explanation.

QUANTXT is a deterministic, nonlinear, multi-factor macro-financial simulation engine implemented in C89 and executed on IBM XT-class hardware. It computes a sovereign risk index by aggregating 13 economic, political, and market-based factor modules through a weighted nonlinear pipeline with reflexive feedback.

It is currently an alpha demo. It is not a tutorial project. The goal is to build a deterministic analytical instrument — built to model US sovereign risk dynamics using real FRED macroeconomic data and market inputs, cross-referenced against contemporary macro frameworks including Paul Tudor Jones's bond market stress thesis and Garrett Jones's structural fiscal narrative. It is an attempt to build deterministic tooling for evaluating macroeconomic narratives under constrained and reproducible execution.

The engine originated as a YAML schema, was prototyped in Python, and was then ported into C89/8087 with successive tightening of constraints and calibration at each stage. Rather than relaxing assumptions during implementation, each layer forced stricter determinism and more explicit modeling. The most constrained environment ultimately produced the most stable behavior.


QUANTXT Code Map

Architecture Overview

MAIN
├── INTRO          (splash screen, no deps)
├── BROWSER        (scenario picker UI)
│   └── SCENARIO   (scenario registry + loader + manual input)
├── FILEBRO        (file picker UI, no module deps)
├── SCENARIO       (scenario registry, scenario_count, loader, manual input)
├── MODULES        (M1–M13 stress functions + compute_system_out)
│   └── state.h    (shared types: State, Scenario, Params, SystemOut)
├── SYSTEM         (SYSTEM_engine, init_default_params)
│   ├── MODULES    (pure stress modules, no cycle in v1.11)
│   ├── UTIL       (clip01, LOGF, SMOOTH_STEP, PWL, SPL)
│   ├── HISTORY    (run history ring buffer)
│   └── ENGINE     (OLS scoring engine)
├── DASH           (dashboard UI — BIOS text renderer)
├── QXCALIB        (calibration runner)
│   └── ENGINE
└── ENGINE         (run_engine — pure float scorer)


Module Reference

MAIN.C

Entry point. Menu loop: browse scenarios, manual input, load .TXT file, run calibration.
Calls: run_intro, browse_scenarios, manual_input, browse_txt_files, load_scenario_file_multi, fill_state_from_scenario, compute_system_out, draw_dashboard, qxcalib_run


INTRO.C / INTRO.H

CGA splash screen.
640×200 mono vector logo, grid animation, BIOS tick timing.
No dependencies on other QUANTXT modules.
Exports: run_intro(void)


BROWSER.C / BROWSER.H

Scenario browser UI.
Scrollable list of loaded scenarios; returns selected index or -1 (ESC).
Depends on: state.h, system.h, scenario.h, colors.h
Exports: browse_scenarios(void)


FILEBRO.C / FILEBRO.H

DOS 8.3 file picker.
Scans *.TXT in working directory; scrollable picker returns filename.
Depends on: colors.h only
Globals: file_list[64][13], file_count
Exports: browse_txt_files(char *out_filename), load_txt_file_list(void)


SCENARIO.C / SCENARIO.H

Scenario registry + loader.
Holds built-in scenarios (gfc_2008, debtceiling_2011, today_2026).
Loads external .TXT packs; provides manual input UI.
Depends on: state.h, colors.h
Globals: scenarios[256], scenario_count
Exports: fill_state_from_scenario, manual_input, load_scenario_file_multi


ENGINE.C / ENGINE.H

OLS scoring engine.
Pure function: takes State, returns float score in [0, 1].
19 OLS-calibrated weights (RMSE 0.020, 30-point precedent set).
Depends on: state.h only
Exports: run_engine(const State *st)


MODULES.C / MODULES.H

Stress sub‑modules M1–M13 + compute_system_out.
Each M‑function is pure, deterministic, and returns a normalized [0–1] stress score.
Evaluation order is dependency‑aware: M3 → M1, M5 → M2, M9 → M6.
Now depends only on state.h (circular dependency with SYSTEM removed in v1.11).
Uses the canonical clip01() from util.h (duplicate local version removed).
Exports: M1M13, compute_system_out(const State*, SystemOut*).

Function Inputs Notes
M1 int_rev, M3_score Interest burden
M2 tail_risk, liq_gap, M5_score Liquidity/tail risk
M3 xdate, cbo_deficit X-date + fiscal
M4 usd_reserve_share USD reserve decline
M5 ofr, hy_spread Credit stress
M6 sahm, M9_score, lagged_ai Cycle + AI reflexivity
M7 dxy_mom, oil_price Dollar + oil shock
M9 ai_capex AI capex stabilizer
M10 geopolitical_risk Geo risk
M13 investor_sentiment Sentiment (contrarian)

SYSTEM.C / SYSTEM.H

Full nonlinear macro‑risk engine.
Executes M‑modules in two passes, applies AI‑capex damping, EMA smoothing, regime classification, and PONR probability estimation.
Now depends on: modules.h, util.h, history.h, engine.h — with the previous circular dependency on MODULES removed in v1.11.
Exports: init_default_params(Params*), SYSTEM_engine(const State*, const Params*, SystemOut*).

Two-pass evaluation:

Pass 1 (independent): M3, M4, M5, M7, M9, M10, M13
Pass 2 (dependent):   M1(←M3), M2(←M5), M6(←M9)

DASH.C / DASH.H

XT‑style dashboard UI.
Centered 60‑column layout with color‑coded risk bars and a clean BIOS‑text renderer.
Page‑flip recursion removed in v1.11 — draw_dashboard() is now non‑recursive and stack‑safe.
Depends on: state.h, system.h, colors.h
Exports: draw_dashboard(const SystemOut*, const State*)


QXCALIB.C / QXCALIB.H

Calibration runner.
Parses CALIB.TXT key=value precedent blocks, runs run_engine against each, computes SSE/MSE, writes CALIB_RESULT.TXT.
Depends on: engine.h, state.h
Exports: qxcalib_run(precedent_file, result_file, QXCalibResult*)


UTIL.C / UTIL.H

Math utilities.
No QUANTXT dependencies — safe to include anywhere.
Exports:

Function Signature Purpose
clip01 (double x) Clamp to [0, 1]
LOGF (x, mid, k) Logistic sigmoid
SMOOTH_STEP (x, th[], n, k) Hermite multi-threshold step
PWL (x, bp[], slopes[], n) Piecewise linear
SPL (x, knots[], vals[], n, irreversible, *state) Linear spline + optional ratchet

HISTORY.C / HISTORY.H

Run history ring buffer.
Stores up to 256 entries of (M8, M9, M10) per run.
No QUANTXT dependencies.
Exports: history_reset(History*), history_append(History*, m8, m9, m10)


STATE.H

Shared data types — no logic, no dependencies.
Defines Scenario (16 fields + name) and State (19 fields including lagged_ai, infl, unemp, gdp).

COLORS.H

CGA/EGA/VGA color constants.
Standard palette (0–15) + semantic UI aliases (COL_NORMAL, COL_HEADER, etc.).
No dependencies.


Known Issues (Resolved in v1.11)

# File Issue Status
1 SYSTEM.HMODULES.H Circular dependency FixedParams & SystemOut moved to state.h
2 MODULES.C Duplicate clip01() Fixed — Now uses canonical version from util.h
3 DASH.C Recursive draw_dashboard() Fixed — Converted to non-recursive page loop

All major architectural issues resolved in v1.11. The codebase is now clean and stable.


Data Flow (happy path)

startup  
  └─ run_intro()  
  └─ main menu loop  
       ├─ browse_scenarios() / manual_input() / load_scenario_file_multi()  
       │       └─ fills **State**  
       ├─ compute_system_out(State → SystemOut)   [MODULES]  
       │       └─ M3, M4, M5, M7, M9, M10, M13  (pass 1)  
       │       └─ M1, M2, M6                    (pass 2)  
       │       └─ weighted sum → raw_stress  
       │       └─ AI damping → M8  
       │       └─ regime classification  
       └─ draw_dashboard(SystemOut, State)       [DASH]


File Index (v1.11)

File Role Deps (user headers)
state.h Core types (State, Scenario, Params, SystemOut)
colors.h Color constants
util.h/c Math helpers (clip01, LOGF, SMOOTH_STEP, PWL, SPL)
history.h/c Run‑history ring buffer
engine.h/c OLS scoring engine (run_engine) state.h
intro.h/c Splash screen
filebro.h/c File picker UI colors.h
scenario.h/c Scenario registry, loader, manual input state.h, colors.h
modules.h/c M1–M13 stress modules + aggregator state.h
system.h/c Full macro‑risk engine state.h, modules.h, util.h, history.h, engine.h
dash.h/c XT‑style dashboard UI state.h, system.h, colors.h
browser.h/c Scenario browser UI state.h, scenario.h, colors.h
qxcalib.h/c Calibration runner state.h, engine.h
main.c Entry point all of the above

Mathematical Model

Pipeline: x → normalize(x) → Fᵢ → S

Normalization Layer (UTIL.C)

Transform Function
clip01 Bounds inputs to [0,1]
LOGF Log compression y = log(1+a|x|)·sign(x)
SMOOTH_STEP Cubic S-curve s(t) = 3z² - 2z³
PWL Piecewise linear between calibrated breakpoints
SPL Spline-like smoothing across multiple knots

Factor Layer (MODULES.C) — M1–M13

Each module is a pure function. Includes cross-factor interaction terms:

Fᵢ = Σⱼ wᵢⱼ Tᵢⱼ(x̃ⱼ)  +  Σⱼ,ₖ vᵢⱼₖ Tᵢⱼₖ(x̃ⱼ, x̃ₖ)
Module Factor
M1 Inflation pressure
M2 Credit stress
M3 Liquidity conditions
M4 Policy stance
M5 Growth expectations
M6 External shocks
M7 Volatility
M9 Sentiment
M10 Structural risk
M13 Higher-order nonlinear composite

System Layer (SYSTEM.C)

L   = Σᵢ αᵢ Fᵢ                    — weighted linear mix
L'  = SMOOTH_STEP(L; a, b)         — nonlinear shaping
L'' = (L')ᵞ                        — optional convexity
S   = clip01(L'')                   — S ∈ [0,1]

Reflexive feedback: M8 output at period t feeds forward into period t+1, capturing the self-reinforcing dynamics of sovereign stress episodes. Damping functions stabilize the loop.

Outputs

Output Description
S Stress score ∈ [0,1]
Regime Low / Medium / High Risk
PONR Point-of-no-return crisis threshold probability

Calibration

Calibrated against 60 scenarios spanning the full US economic cycle — from early expansion through severe crisis and back through recovery — using FRED macroeconomic series and market data.

Metric Value
Scenarios (initial) 30
Mean error (initial) 0.135473
Scenarios (current) 60
Total error (current) 0.228393
Mean error (current) 0.003807
Error reduction ~35× improvement
Data sources FRED + market data

At mean error 0.0038 on a 0–1 scale, the model is hitting targets with sub-percent accuracy across the full economic cycle.


SCENARIO

Scenarios are in the following format. The goal will be to soon unify the Scenario and Calibration data formats.

#name int_rev debt_gdp usd_reserve_share cbo_deficit xdate sahm tail_risk liq_gap ofr hy_spread dxy_mom oil_price ai_capex geopolitical_risk investor_sentiment

managed_norm_1 22 70 65 3 90 0.2 1.2 1.0 0.95 600 5 80 1.0 0.5 0.4
managed_norm_2 23 72 66 3.2 88 0.18 1.1 1.05 0.96 580 4 82 1.1 0.45 0.42
managed_norm_3 21 68 64 2.8 92 0.22 1.3 0.98 0.94 620 6 78 0.9 0.55 0.38

Input Vector — 19-Field Sovereign Risk Schema

Field Description Range
int_rev Interest as share of revenue 0.0 – 1.0
debt_gdp Debt as share of GDP 0.0 – 2.0
usd_reserve_share USD share of global reserves %
cbo_deficit CBO projected deficit % GDP
xdate Days to debt ceiling breach 0 – 360
sahm Sahm rule recession indicator 0.0 – 1.0
tail_risk Tail risk index 0.0 – 1.0
liq_gap Liquidity gap index 0.0 – 1.0
ofr OFR financial stress index 0.0 – 1.0
hy_spread High yield spread index 0.0 – 1.0
dxy_mom DXY momentum index 0.0 – 1.0
oil_price Crude oil price USD/barrel
ai_capex AI capital expenditure index 0.0 – 5.0
lagged_ai Prior period AI capex 0.0 – 5.0
geopolitical_risk Geopolitical risk index 0.0 – 1.0
investor_sentiment Investor sentiment index 0.0 – 1.0
infl Inflation rate 0.0 – 1.0
unemp Unemployment rate 0.0 – 1.0
gdp GDP growth rate -0.1 to 0.1

Calibration File Format

# Scenario Name
int_rev=0.32
debt_gdp=0.95
usd_reserve_share=57.0
cbo_deficit=11.0
xdate=20
sahm=0.40
tail_risk=0.90
liq_gap=0.85
ofr=0.90
hy_spread=0.88
dxy_mom=0.70
oil_price=60.0
ai_capex=0.15
lagged_ai=0.20
geopolitical_risk=0.75
investor_sentiment=0.15
infl=0.02
unemp=0.12
gdp=-0.05
target=0.92

Scenarios are separated by blank lines. Comments begin with # or ;.


Build

# QUANTXT — Build Instructions

## Requirements

| Tool | Version | Notes |
|------|---------|-------|
| [Open Watcom C](https://github.com/open-watcom/open-watcom-v2/releases) | V1.9 | Primary toolchain |
| DOS or DOSBox | Any | For running the compiled binary |

> **DOSBox** is the recommended environment for modern systems (Windows, macOS, Linux).  
> Download: https://www.dosbox.com/

---

## Quick Build (Makefile)

If you have Open Watcom installed and the repo cloned, the simplest path is:

```bat
wmake -f QUANTXT.mk

This compiles all modules and links QUANTXT.EXE in one step.


Manual Build

Step 1 — Compile each module

Run wcc on each .C file. Replace C:\WATCOM\h with your Watcom include path if different.

wcc BROWSER.C  -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc DASH.C     -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc ENGINE.C   -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc FILEBRO.C  -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc HISTORY.C  -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc INTRO.C    -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc MAIN.C     -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc MODULES.C  -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc QXCALIB.C  -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc SCENARIO.C -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc SYSTEM.C   -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml
wcc UTIL.C     -i="C:\WATCOM\h" -w4 -e25 -zq -od -d2 -fpi87 -bt=dos -fo=.obj -ml

Step 2 — Link

wlink name QUANTXT sys dos op st=16384 op m op maxe=25 op q op symf FILE BROWSER.obj,DASH.obj,ENGINE.obj,FILEBRO.obj,HISTORY.obj,INTRO.obj,MAIN.obj,MODULES.obj,QXCALIB.obj,SCENARIO.obj,SYSTEM.obj,UTIL.obj

Compiler Flag Reference

Flag Meaning
-ml Large memory model — required
-fpi87 Inline 8087 FPU instructions — required
-bt=dos Target real-mode DOS
-w4 Warning level 4
-e25 Stop after 25 errors
-zq Quiet mode
-od Disable optimization (debug-safe)
-d2 Full debug info

Linker Flag Reference

Flag Meaning
sys dos Real-mode DOS executable
op st=16384 Stack size 16 KB — required
op m Generate map file
op maxe=25 Stop after 25 errors
op q Quiet mode
op symf Generate symbol file

Running on Modern Systems (DOSBox)

  1. Install DOSBox
  2. Launch DOSBox and mount your build directory:
mount c C:\path\to\QUANTXT
c:
QUANTXT.EXE
  1. The splash screen will load, followed by the main menu.

On a real IBM PC/XT the calibration routine will pause briefly while loading 60 scenarios — this is expected.


Tested Environment

Component Spec
Compiler Open Watcom C V1.9
Target OS DOS 6.22
Hardware IBM PC/XT or compatible
FPU Intel 8087 (or DOSBox emulation)
RAM 256 KB minimum
Display CGA

Requirements:

  • Watcom C/C++ (wcc / wlink)
  • Real-mode DOS (or DOSBox for modern systems)
  • Memory model: large (-ml)

System requirements:

  • IBM PC/XT or compatible (8087 for Math CoProcessor acceleration)
  • 8087 recommended for full precision
  • 256 KB RAM minimum
  • CGA graphics
  • DOS 3.3+

Usage

Key Action
1 Browse scenario files
2 Manual Entry of Data
3 Load Scenario File
4 Run Calibration File*
ENTER Run engine
ESC Quit
  • Screen will pause for a few seconds on an IBM PC as it loads 60 sets of calibration data in to the model

Development History

Phase Technology Purpose
I YAML Parameter schema; model contract definition; PTJ-Weimar benchmark
II Python 3 Mathematical validation; scenario runner; output format lock-in
III C89 / 8087 Production engine; deterministic execution; calibration hardening

The flat-file TXT format served as the binding data contract across all three phases. AI Scalped scenarios feed directly into QUANTXT.EXE without transformation.


Why C89 on XT Hardware

The constraint is the point.

An engineer who can build a calibrated, nonlinear, multi-factor sovereign risk engine within the constraints of a 1981 coprocessor understands the computation — not just the tools. No runtime. No garbage collector. No dynamic loader. No abstraction noise between the mathematics and the metal.

The 8087's 80-bit extended precision and real-mode determinism produced tighter calibration than Python floating point. The constraints proved that a big idea could be ported to one of the earliest iterations of home PC.

If it works on an IBM XT, it works anywhere.


File Structure

QUANTXT/
├── MAIN.C          — Orchestrator
├── ENGINE.C        — Computational hub
├── MODULES.C       — M1–M13 factor modules
├── SYSTEM.C        — Aggregation + regime classification
├── SCENARIO.C      — Scenario loader → STATE struct
├── FILEBRO.C       — File browser + I/O
├── DASH.C          — ASCII dashboard renderer
├── QXCALIB.C       — Calibration engine
├── HISTORY.C       — Scenario state tracking
├── UTIL.C          — Math transforms (clip01, LOGF, SMOOTH_STEP, PWL, SPL)
├── STATE.H         — Core struct definitions
├── ENGINE.H        — Engine interface
├── MODULES.H       — Module interfaces
├── SYSTEM.H        — System layer interface
├── QXCALIB.H       — Calibration interface
├── UTIL.H          — Utility function declarations
├── COLORS.H        — CGA color constants
├── QUANTXT.MK      — Watcom makefile
├── CALIB.TXT       — 60-scenario calibration dataset


Status

  • Core engine — stable
  • 60-scenario calibration — mean error 0.0038
  • YAML → Python → C89/8087 evolution complete
  • CGA QXGraph visualization operational
  • Multi-scenario TXT loader
  • Header architecture hardened (debugging sprint complete)
  • Multi variable scenario simulations used for calibrating the dataset, and more to be performed over time.
  • Developing Forecast modeling on sequential daily data
  • Improving Regime transition validation on real observed path
  • Current US environment live stress assessment
  • QXGraph multi-parameter overlay + zoom

Author

ArchitectureLabs Systems engineer. 16 years industrial operations experience. Building quantitative tools at the intersection of macro-financial analysis and constraint-driven systems engineering.


Changelog

QUANTXT v1.11 — Stability & Architecture Cleanup Release

Release Date: May 2026

Focus: Architecture cleanup, dependency fixes, dashboard stability, and UI alignment improvements.

Overview Version 1.11 is a structural and stability-focused update that resolves long‑standing architectural issues, removes circular dependencies, stabilizes the dashboard renderer, and improves the visual layout of the XT‑style interface. This release does not change model logic — it strengthens the foundation the engine runs on.

Key Fixes & Improvements

  1. Removed SYSTEM.H ↔ MODULES.H Circular Dependency Centralized all shared types (State, Scenario, Params, SystemOut) into state.h.

modules.h and system.h now depend only on state.h, eliminating recursive include chains.

Build is now deterministic under Watcom and easier to maintain.

  1. Removed Duplicate clip01() Implementation Deleted the shadowed static version inside modules.c.

The engine now uses the canonical implementation from util.h.

Prevents inconsistent clipping behavior across modules.

  1. Eliminated Recursive draw_dashboard() Call Removed the page‑flip recursion that previously caused stack growth.

Dashboard rendering is now linear, predictable, and safe.

  1. Replaced Direct VRAM Writes With BIOS Text Rendering Removed fragile CGA memory writes that caused screen bleed and instability.

Switched to _outtext()‑based rendering for full compatibility across DOSBox, CGA, EGA, and VGA.

Dashboard is now flicker‑free and stable.

  1. Centered Bargraph Layout Introduced DASH_BAR_WIDTH and DASH_BAR_OFFSET for clean horizontal centering.

Bars now align visually with the dashboard header and numeric columns.

Produces a more balanced and readable XT‑style UI.

QUANTXT v1.12 — Stability, Precision & Calibration Release

Release Date: May 21, 2026

Focus: Floating-point precision corrections, calibration engine upgrade, file browser improvements, and UI stability fixes.

Precision Fixes

  • Corrected float → double throughout ENGINE.C, ENGINE.H, and QXCALIB.H for correct 8087 FPU precision. State fields were already double; ENGINE was the remaining inconsistency.
  • Removed erroneous (float) casts in ENGINE.C weight accumulation.
  • QXCalibResult total_error and mean_error promoted to double.

Calibration Engine Upgrade

  • Calibration option now opens a file browser to select any .TXT calibration dataset rather than being hardcoded to CALIB.TXT.
  • In-progress screen with blinking status message shown during run so the engine does not appear frozen on real hardware.
  • Full Phase 1 diagnostic report now written to CALIB_RESULT.TXT: aggregate metrics (SSE, MSE, RMSE), bias analysis, regime breakdown across 5 stress bands, and per-scenario detail table with flagged outliers.
  • Post-run results displayed on screen with spacebar paging.

Browser Fixes

  • BROWSER.C: fixed viewport scroll bug where scenarios beyond row 25 were not visible. Added top/sel coupled scrolling, PgUp/PgDn, scroll indicators, and scenario counter.
  • FILEBRO.C: same viewport fix applied. Added scroll indicators, PgUp/PgDn, file counter, and empty directory message.

General

  • Apache License 2.0 header added to source files.
  • Version references updated to v1.12 throughout.

QUANTXT v1.12 — May 21 2026

QUANTXT is licensed under the Apache-2.0 License. See LICENSE for details.

About

QUANTXT is a fully self‑contained macro‑risk modeling engine built in ANSI C89 and designed to run on real DOS hardware, DOSBox, and retro‑compatible toolchains. It combines a deterministic 13‑factor stress model, a nonlinear system engine, and an XT‑style dashboard UI into a compact, high‑performance analytics suite.

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages