The SAVE Score Explained: Sentiment, Analyst Consensus, Valuation, and Earnings Quality

Sentiment moves markets. Everyone knows this. What almost no one does is measure it systematically.

Most "sentiment indicators" are noise. Reddit post counts, Twitter mentions, put/call ratios — useful signals in isolation, but rarely combined into a coherent, calibrated measure.

The SAVE score is our attempt to do this properly.

What SAVE Stands For

S — Sentiment: the directional lean of current market opinion. Are analysts, institutions, and market participants net positive or net negative on this stock right now? Sentiment captures momentum in perception before it shows up in price.

A — Analyst Consensus: the trend and direction of professional analyst coverage. Are ratings upgrading or downgrading? Is earnings forecast consensus rising or falling? Analyst behavior is a leading indicator that retail participants often miss — particularly when consensus is quietly shifting.

V — Valuation: the quantitative comparison between the stock's fundamental fair value and its current market price. A stock trading well below its multi-model consensus fair value scores positively here. The gap between price and value is the margin of safety — and it is the core of everything Equity Rank calculates.

E — Earnings Quality: the reliability and trend of the earnings base. Consistent beat rates, low revision volatility, and improving forward guidance reliability all raise a stock's earnings quality signal. A cheap stock with deteriorating earnings quality is often cheap for a reason.

These four components are combined into a single score from 0 to 100, where 50 represents neutral — the baseline expectation.

Why Four Components?

Each component captures a distinct information layer. Sentiment without supporting analyst consensus can be retail noise. Analyst consensus without valuation context can lead toward expensive consensus favorites. Valuation without earnings quality can trap a portfolio in value stocks with deteriorating fundamentals.

Combining all four reduces the likelihood that any single noisy signal dominates the output.

The Accuracy Question

Design-stage simulation figures for the SAVE score were retired from publication in August 2026; the platform now publishes only live factor diagnostics with stated statistical significance.

A few important caveats:

• This is based on simulation and historical modelling, not live trading results
• Directional accuracy means "did the price move in the direction the score predicted" — it does not predict magnitude or timing
• The real backtest against live market data begins October 2026

Transparency about model performance means publishing live measurements rather than simulation numbers. The methodology page carries the platform's daily factor diagnostics — including unflattering readings — with their statistical significance stated. Markets are not perfectly predictable, and no reading is a guarantee.

How It Adjusts the Fair Value

The SAVE score modifies the base fair value through a calibrated sentiment overlay:

• A neutral score (50) adds zero adjustment — the base valuation stands
• Scores above 50 lift the fair value estimate proportionally
• Scores below 50 discount it proportionally
• The adjustment is capped so sentiment alone cannot swing the fair value by more than a defined percentage — the fundamentals still anchor the number

This design means a stock with strong fundamentals and strong sentiment gets a higher adjusted fair value. A stock with strong fundamentals but negative sentiment gets a lower one — reflecting that the market's current perception is a real drag on convergence timing, even if the fundamentals are sound.

The Independence Insight

What makes the combined model work is that the SAVE score and the Innovation Score are measuring genuinely different things.

Sentiment is what the market thinks right now. Innovation is what a company is building for the future. These two signals have a correlation of r = -0.005 — essentially zero. They're independent.

When independent signals are stacked additively, the combined accuracy is higher than either alone — but only if they're truly independent. Correlated signals just double-count the same information.

This is why hedge funds have stacked independent signals for decades. The math works. We built Equity Rank to bring the same approach to retail investors.

Combined SAVE + Innovation accuracy figures from design-stage simulation were retired in August 2026 — live diagnostics on the methodology page are the published readings.

Explore the full methodology at equity-rank.com/methodology

Simulation-based results. Not a guarantee of future performance. Real-world backtest begins October 2026.