How Synthetic Exposure calculates Exposure Scores

An Exposure Score is the signed Pearson correlation coefficient between a crypto asset and a stock over the past 90 days. It ranges from −1.00 (perfect inverse relationship) through 0.00 (no linear relationship) to +1.00 (perfect positive relationship).

A score of +0.78 means the two assets have had a strong positive price-movement relationship over the window. A score of −0.62 means an inverse relationship — when one rose the other tended to fall. A score near 0.00 means no consistent linear relationship was detected.

This is not a forecast. An Exposure Score reflects past price-movement similarity, not guaranteed future behaviour.

• Adjusted daily closing prices for the selected stock (Marketstack)
• Daily crypto prices sampled at a consistent UTC timestamp (CoinGecko)
• Default lookback: 90 calendar days
• Only dates on which the stock has a valid close are used
• Minimum 45 aligned observations required to compute a score

1. Sort prices oldest → newest
2. Align stock and crypto PRICE series by date first (inner join) — a Monday return computed independently would span Friday→Monday for stocks but Sunday→Monday for crypto under the same date label, so alignment must happen before returns, not after
3. Compute daily log returns on the aligned price slices:
   return_t = ln(price_t / price_{t-1})
4. Calculate Pearson correlation r across aligned observations — the result is the Exposure Score, stored without any clamp or transform:
   exposure_score = r    (range: −1.00 to +1.00)

Scores are always displayed with an explicit sign: +0.78, −0.62, 0.00. They are never shown as percentages. A score does not represent ownership, guaranteed exposure, or an equivalent percentage of price movement.

Edge thickness in the Exposure Graph reflects magnitude (|r|). Colour reflects direction: violet for positive relationships, amber for inverse relationships.

Historical Exposure Scores recalculate Tuesday and Friday after US market close. Live Exposure Scores update after every completed 30-minute market interval while the US market is open, and freeze (never recalculated) while it is closed — crypto prices keep updating continuously in the background. Scores are pre-computed and cached — they are not recalculated during a page request.

Data freshness is shown on every score card and graph node.

• The 90-day window may not capture long-term structural relationships
• Correlation can change rapidly with market regime shifts
• The asset universe is curated, not exhaustive — 8 stocks × 30 crypto free, 20 × 100 for verified $SynthEx holders
• Stock market holidays reduce observation counts
• Scores reflect price-movement similarity, not ownership, market-cap exposure, or guaranteed equivalent performance