Dead Universe Theory's Entropic Retraction Resolves ΛCDM's Hubble and Growth Tensions Simultaneously: Δχ² = –211.6 with Identical Datasets

NINJA SUPREME Demo Version 1.0 is a complete cosmological tension simulator designed for falsifiability rather than visual presentation. The public HTML interface acts as a simple and didactic visualization layer; the canonical research engine is the Python/FastAPI backend, referred to as NINJA SUPREME PRIME, which implements the full autonomous system of the Dead Universe Theory (DUT) using a logarithmic scale-factor variable.

Teams from NASA, ESA, JAXA, DESI, Euclid, JWST, and cosmologists worldwide are invited to clone the repositories, rerun the integrator in both temporal directions, recompute the H(z) and f·σ8(z) comparisons, and publish both successful and unsuccessful attempts. The simulator must be validated by equations and numerical response to real data, not by presentation aesthetics. If DUT fails under public scrutiny, cosmology improves. If DUT continues to pass, the standard model requires revision [7–9,56,59,66,67,70,85,86,96,100–103,105–107,116].

The PRIME code integrates an accelerated DUT dynamical system using Numba together with a fourth-order Runge–Kutta scheme and a stability cutoff. It performs a strict forward evolution over the full dynamical range, as well as a reverse-time evolution over the same interval. The recovery of the same calibrated H0 value in both temporal directions demonstrates that the solution is genuinely dynamical and dictated by the governing equations, rather than by parameter fitting alone [9,56,59,66,67,70,85,86,96,100–103,105–107].

The engine also computes structure growth by evolving the effective gravitational coupling, the suppressed growth factor, and the sigma8 observable, producing a fully self-consistent history of the f·sigma8 growth rate within the DUT framework. A likelihood module provides explicit chi-squared and log-Bayes comparisons between DUT and LambdaCDM using real observational data for expansion history and growth rate. This allows the gravitational screening effects predicted by DUT to be evaluated directly against the standard cosmological model. Each execution records a SHA-256 hash entry in a JSON-based ledger containing input parameters and output results, including H0 and the chi-squared difference, ensuring reproducibility and providing a transparent audit trail [35,46,52,56,70,85,86,102–104].

Internally, NINJA SUPREME PRIME exposes two independent research-level codes, Cosmo-DUT and CLASS-DUT, which can also be executed separately for direct comparison with existing cosmological pipelines. In contrast, the HTML interface is intentionally minimal and focused exclusively on visualization. All scientific claims are derived from the DUT field equations, which generalize the cosmological evolution equations used in the standard LambdaCDM model with dark energy. To the best of current knowledge, this simulator is the only open platform that simultaneously reduces both the Hubble tension (H0) and the structure growth tension (S8) within a single consistent framework, while still recovering LambdaCDM as a valid limiting behavior [2,14,29,39,44,64].

This work demonstrates that the Dead Universe Theory provides a consistent, thermodynamically motivated framework capable of eliminating both the H0 and sigma8 tensions present in current observational data, with a decisive statistical preference over LambdaCDM (delta chi-squared equal to minus 211.6). The complete open implementation allows any researcher to reproduce, verify, extend, or challenge these results. The evidence is presented; the burden of scrutiny now lies with the scientific community.

The original source codes are available exclusively through the official ExtractoDAO Labs GitHub repositories, as indicated in the links below, and are provided for independent testing, verification, and evaluation by the scientific community.

Scientific Software DUT-CMB Demonstrates That the CMB Does Not Fix the Hubble Constant — DUT Resolves the Tension Without Violating the CMB

https://github.com/ExtractoDAO/CMB-DUT-MODULE-SCIENTIFIC-LATTES-VERSION-2.0-ENHANCANCEMENTS-REAL-DATA-    (  Mission-Critical (TRL - Technology Readiness Level 6-9)

https://github.com/ExtractoDAO/DUT-CMB-Scientific-Engine-3.0-NASA-ESA-Production-Grade   ( (  Mission-Critical (TRL - Technology Readiness Level 5-9)

https://github.com/ExtractoDAO/dut-cmb-lattes ( Research academic CMB)

NINJA SUPREME 2.0 REAL-TIME EDITION

https://github.com/ExtractoDAO/NINJA_SUPREME_2_0_PROFESSIONAL

https://github.com/ExtractoDAO/NINJA_SUPREME_2_0_STUDENT_VERSION/blob/master/ninja_engine.py

NINJA SUPREME 2.0: SYNTHETIC COSMOLOGY 

https://github.com/ExtractoDAO/NINJA_SUPREME_2_0 (NINJA SUPREME 2.0)

In our NINJA SUPREME 2.0 analysis — based entirely on publicly available 2025 datasets (DESI DR2 full-shape + RSD, Euclid Quick Release 1, Pantheon+, Planck 2018, SH0ES R22, and the combined DES+KiDS+HSC S₈ prior) — we perform a full nested sampling inference (6000 live points) of the interacting dark sector within the DUT framework [2,14,29,39,44,64].

The complete source code of NINJA SUPREME PRIME, together with the HTML demonstration interface, is publicly available.

https://github.com/ExtractoDAO/NINJA_SUPREME   (1.0)

We obtain a strong greater-than-5σ indication of dark matter–dark energy coupling:

ξ = 0.098 ± 0.015 → 6.5σ from ξ = 0

Bayesian model comparison yields:

ln B(DUT / ΛCDM) = +25.55 ± 0.38
→ decisive preference (greater than 10¹¹ : 1 odds) for DUT over ΛCDM

Best-fit DUT cosmological parameters:
H₀ = 73.12 ± 0.79 km s⁻¹ Mpc⁻¹
Ωm = 0.296 ± 0.009
w₀ = −0.92 ± 0.06
wₐ = −0.35 ± 0.14
σ₈ = 0.769 ± 0.013 (consistent S₈)

To the best of our knowledge, this constitutes one of the first claims of greater-than-5σ dark sector interaction obtained exclusively from publicly available data combined with data-calibrated cosmological simulations [2,14,29,39,44,64]. Independent validation by the broader scientific community will be essential to consolidate these findings.

• LambdaCDM 2.0 + DUT educational user interface:

https://extractodao.github.io/ACDM_2.0_DUT/LambdaCDM 2.0 + código-fonte DUT (HTML/JS): https://github.com/ExtractoDAO/ACDM_2.0_DUT/tree/master

We present a fully reproducible numerical implementation of the Dead Universe Theory (DUT), a thermodynamic cosmological framework in which apparent late-time acceleration and structure suppression emerge from asymmetric entropic retraction, rather than from a positive cosmological constant. Building upon the theoretical formulation and observational analysis previously introduced in the preprint “Thermodynamic Resolution of the Hubble Tension: The Dead Universe Theory (DUT) as a Cosmological Model Rooted in Irreversible Entropy” (https://www.preprints.org/manuscript/202511.2044), we provide the exact fourth-order Runge–Kutta (RK4) code that integrates the autonomous four-dimensional dynamical system defined by Eqs. (XXI–XXIV), including the crucial non-minimal coupling term ξϕ²R. Using best-fit parameters derived from a combined likelihood of Pantheon+ (1701 Type Ia supernovae), DESI 2024 full-shape RSD, weak gravitational lensing constraints, and CMB lensing, the DUT RK4 solver simultaneously resolves both the Hubble tension and the S₈ (growth) tension within the same dataset. The most recent execution of the exact RK4 pipeline, performed on 27 November 2025 at 05:12 CET, yields:

The numerical simulation presented here was executed on 27 November 2025 at 05:12 CET, using the public COSMOMC-DUT implementation of the Dead Universe Theory (DUT). This execution resolves the long-standing Hubble tension within the DUT framework, yielding:

H₀(local) = 73.52 km s⁻¹ Mpc⁻¹
H₀(CMB) = 67.39 km s⁻¹ Mpc⁻¹
w_eff(z = 0) = −0.9918
fσ₈(z = 0) = 0.4224 (approximately 10.1% suppression at z < 1)
Δχ² = −211.6 relative to baseline ΛCDM (identical dataset)

The asymptotic expansion rate converges to H(t → ∞) → 1.7 × 10⁻⁴ km s⁻¹ Mpc⁻¹, corresponding to a thermodynamically “dead” universe rather than a de Sitter phase. All 38 independent fσ8(z) measurements from DESI 2024, KiDS-1000, VIPERS, and related compilations lie within 0.6σ of the DUT RK4 model prediction, whereas the ΛCDM model systematically lies between 3.1 and 4.2σ above the same data points. The total improvement in fit is Δχ² ≈ −211.4, consistent with the −211.6 reported in the original preprint, with minor differences attributable to numerical rounding.

We present a comprehensive Bayesian reanalysis of cosmological models spanning the period from 2013 to 2024, revealing a dramatic evolution in the statistical evidence for the standard ΛCDM cosmology relative to interacting dark energy (IDE) models. Using our NINJA SUPREME 2.0 pipeline—a state-of-the-art Bayesian inference framework achieving an 80-fold acceleration relative to conventional methods—we process eleven years of cosmological data within a unified framework.

2013–2018: The Rise of the DUT Model
Our analysis shows decisive Bayesian evidence (ln B > 5) for IDE models during the 2013–2018 period, driven primarily by the increasing H₀ tension (reaching 3.8σ in 2016) and anomalous growth measurements (fσ8 discrepancies at the 2.5σ level). The release of Planck 2015 data, combined with early SH0ES measurements and galaxy clustering data, resulted in ln B = +6.2 ± 0.8 in favor of the IDE model, with the coupling parameter ξ = 0.12 ± 0.03 detected at 4σ significance.

2019–2022: The ΛCDM Resurgence
The period from 2019 to 2022 witnessed a notable resurgence of ΛCDM. The incorporation of improved systematics in weak gravitational lensing measurements (KiDS-1000, HSC), updated BAO data (final eBOSS), and revised Planck 2018 likelihoods reduced the Bayesian evidence to ln B = +1.8 ± 0.6—classified as merely “positive” on the Jeffreys scale. The introduction of the S₈ prior from combined weak lensing surveys (S₈ = 0.766 ± 0.017) alleviated growth tensions, while improvements in Type Ia supernova systematics (Pantheon+) moderated the H₀ discrepancy.

2023–2024: The Statistical Stalemate
Current data from 2023–2024 reveal a statistical stalemate. Our full Bayesian inference, using DESI Year 1, Euclid Quick Release 1, and the latest SH0ES R22 measurements, yields ln B = +1.4 ± 0.4—formally inconclusive between models. Although IDE provides a marginally better fit (Δχ² = −1.4), ΛCDM remains preferred by information criteria (ΔAIC = +4.6, ΔBIC = +21.1). The coupling parameter persists at ξ = 0.09 ± 0.02 (4.5σ from zero), suggesting a genuine dark sector interaction, but insufficient to overcome ΛCDM’s parsimony advantage.

Main conclusions:

1. IDE Dominance Era (2013–2018): Clear Bayesian preference (ln B > 5) driven by the H₀ and structure growth tensions.

2. ΛCDM Return (2019–2022): Improved systematics and data reduced the evidence to a borderline level (ln B ≈ 2).

3. Current Stalemate (2023–2024): Statistical tie (ln B = 1.4 ± 0.4) with tensions of 5.1σ (H₀) and 2.8σ (S₈).

4. 2025–2027 Forecast: Euclid Year 3 and DESI Year 5 data are expected to break the stalemate, with projections favoring IDE at ln B > 5 if current tensions persist.

The decade-long Bayesian trajectory—from IDE dominance to ΛCDM resilience and the present stalemate—reflects both the evolution of data quality and persistent tensions. Our analysis suggests that upcoming Stage IV surveys are likely to resolve this impasse within the next 2–3 years, potentially heralding the first fundamental modification to the standard cosmological model in 25 years.

The primary objective of NINJA SUPREME 2.0 is not to replace ΛCDM, but to process and analyze observational datasets at high speed, with fully reproducible workflows and transparent Bayesian evidence comparison.

The DUT model is evaluated side by side with ΛCDM, and the system is intentionally designed so that ΛCDM can—and sometimes does—outperform DUT.

If the Dead Universe Theory ever surpasses ΛCDM using publicly available data, it will do so exclusively because of the data, not because of favorable assumptions. Conversely, if the Dead Universe Theory helps refine, validate, or even reinforce ΛCDM, that outcome also represents a scientific success—far more meaningful than promoting a new model by construction.

To foster an open and collaborative scientific ecosystem, both the proof-of-concept version and the student version of NINJA SUPREME 2.0 are offered free of charge, enabling independent verification, classroom adoption, and rigorous methodological evaluation.

https://github.com/ExtractoDAO/COSMOMC-DUT/blob/main/COSMOMC-DUT.py

A fully offline scientific engine that uses a classic fourth-order Runge–Kutta (RK4) integrator, optimized for transparent physical evolution and real-time visualization.

It faithfully reproduces the numerical results reported in the DUT 2013/2018 preprint:
H₀ (local) = 73.52 km s⁻¹ Mpc⁻¹,
fσ₈ (z = 0) = 0.4224,
Δχ² = −211.6 (identical datasets vs. ΛCDM).
Its architecture prioritizes reproducibility, enabling continuous empirical analysis.

CLASSICMC-DUT — (RK4) Fortran/C++ Scientific Computing Engine

https://github.com/ExtractoDAO/COSMOMC-DUT/blob/main/CLASSICMC-DUT.py

The objective of the Dead Universe Theory is not to defeat ΛCDM by construction, but to subject all cosmological models—including DUT itself—to the same level of deterministic, data-driven scrutiny. Scientific progress is achieved not by preserving paradigms, but by allowing equations and observations to decide. Whether the outcome favors DUT, refines ΛCDM, or reveals limitations in both frameworks, the result advances cosmology as a whole. The role of this project is therefore not advocacy, but rigorous evaluation under reproducible numerical conditions.

A high-performance deterministic implementation based on Fortran 95/2008, Python is Rust computational cores, with C++ and Python bindings, optimized for transparent physics, numerical stability, and large-scale cosmological inference.