DEV Community

Anders
Anders

Posted on • Originally published at zarq.ai

We Backtested Our Crypto Ecosystem Score — Here's What Predicted Crash Protection (p < 0.001)

During the 2025–2026 crypto crash, our top-quintile tokens lost 26% while bottom-quintile lost 70%.

We didn't cherry-pick this result. We tested across 3 time windows, 355–412 tokens each, with proper out-of-sample methodology. Here's what we found.

What Is the Vitality Score?

ZARQ rates 15,000+ crypto tokens on a 0–100 "Vitality Score" measuring ecosystem health across 5 weighted dimensions:

  1. Ecosystem Gravity (20%) — protocol count, TVL, stablecoin presence on the token's chain(s)
  2. Capital Commitment (20%) — TVL retention, yield pool density
  3. Coordination Efficiency (15%) — DeFi category diversity, audit coverage
  4. Stress Resilience (25%) — NDD stability, crash probability, drawdown behavior
  5. Organic Momentum (20%) — TVL trend, price trend, rating trend

Grades range from S (≥85, exceptional) through F (<25, minimal).

The Backtest: 3 Windows, No Look-Ahead Bias

We reconstructed a historical Vitality proxy at 3 past dates using only data available at that time, then measured forward returns:

Window Score Date Forward Period Tokens
A Jan 2024 Jan 2024 → Jan 2025 (12 months) 355
B Jan 2025 Jan 2025 → Jan 2026 (12 months) 363
C Jul 2025 Jul 2025 → Feb 2026 (crash, 8 months) 412

For each window, we split tokens into quintiles by Vitality Score and measured median returns.

The Results

Window C (the crash window) — statistically significant

Quintile N Vitality Score Median Return
Q1 (TOP) 82 52.5–67.7 -26.1%
Q2 82 43.6–52.3 -48.8%
Q3 82 38.3–43.5 -55.4%
Q4 82 33.3–38.2 -56.1%
Q5 (BOTTOM) 84 26.7–33.3 -70.4%

Q1–Q5 spread: +44.3%. Perfectly monotonic (4/4 steps). t-statistic: 3.35, p-value: 0.0008.

Windows A and B

Window A (bull market): Q1–Q5 spread +9.3%, NOT statistically significant (p=0.556). The model doesn't predict upside well.

Window B (bear market): Q1–Q5 spread +27.1%, perfectly monotonic (4/4), but NOT significant (p=0.392) due to high variance.

Which Dimension Matters Most?

We tested each dimension independently. The answer is clear:

Dimension Window B Spread Window C Spread
Stress Resilience +66.1% +52.5%
Organic Momentum +1.1% +6.2%
Capital Commitment -1.9% +3.2%
Ecosystem Gravity -11.4% -8.3%
Coordination Efficiency -18.5% -8.6%

Stress Resilience dominates. Ecosystem Gravity and Coordination Efficiency actually show negative spreads — they measure ecosystem quality but don't predict returns.

Being Honest About Limitations

  • Only Window C is statistically significant. Windows A and B show the right direction but aren't conclusive alone.
  • The model predicts downside protection better than upside performance.
  • Sample sizes are 355–412 tokens per window — large for crypto research but small in absolute terms.
  • Survivorship bias: tokens that died during windows are excluded from the analysis.
  • Crypto returns are fat-tailed and non-normal — p-values should be interpreted cautiously.
  • Past performance does not guarantee future results.

How to Use It

Check any token's Vitality Score:

GET https://zarq.ai/v1/vitality/ethereum
Enter fullscreen mode Exit fullscreen mode

Response:

{
  "token": "ethereum",
  "vitality_score": 65.0,
  "grade": "B",
  "dimensions": {
    "ecosystem_gravity": 97.6,
    "capital_commitment": 53.3,
    "coordination_efficiency": 80.8,
    "stress_resilience": 58.9,
    "organic_momentum": 40.0
  },
  "confidence": 100,
  "interpretation": "Developing ecosystem with solid fundamentals"
}
Enter fullscreen mode Exit fullscreen mode

Or check it alongside Trust Score and crash probability:

GET https://zarq.ai/v1/check/ethereum
Enter fullscreen mode Exit fullscreen mode

The Bottom Line

Vitality Score is most valuable as a crash protection indicator. Tokens with high ecosystem quality — especially high Stress Resilience — lost significantly less during the 2025–2026 crash. The signal is consistent across all 3 windows but only statistically significant in the crash window.

This doesn't mean high-Vitality tokens will go up. It means they're less likely to collapse when the market turns.

Links:


Data and analysis by ZARQ — the trust layer for the machine economy.

Top comments (0)