🧐Bias

A single score that captures direction and conviction, from one wallet to an entire cohort.

Bias is HyperTracker’s way of summarizing positioning into something you can compare instantly. It combines two ideas: which side a wallet is leaning toward, and how much size that wallet is putting behind the view relative to its perp equity.

The result is a bias score that works at every level. You can read it for one wallet, or aggregate it across a cohort to see whether a segment of traders is meaningfully long, meaningfully short, or simply not committed.

Bias is descriptive. It is designed to help you spot trends, crowding, and leverage risk. It does not predict price by itself.

Wallet Calculation

Perp Bias is calculated in three steps, using a weighted approach that separates direction from conviction.

1) Net Directional Bias

This measures pure direction, independent of position size.

Range: -1 means fully short, +1 means fully long

2) Exposure Ratio

This measures how much positioning is being taken relative to available perp equity.

Interpretation: Higher values mean the wallet is taking larger positions relative to its equity.

3) Final Bias Score

This combines direction and exposure into a single score.

Range: Values beyond -1 or +1 indicate high leverage positioning, since exposure is exceeding equity.

Example 1

Wallet (Equity: $100,000, Long: $1,000, Short: $0)

Example 2

Wallet (Equity: $500, Long: $1,000, Short: $0)

---

Cohort Calculation

Cohort bias is computed by aggregating position totals across all wallets in the cohort, then applying the same bias formula used at the wallet level. This keeps the score consistent and comparable across cohorts of different sizes.

Steps

• Sum total cohort LONG across wallets

• Sum total cohort SHORT across wallets

• Aggregate total cohort Perp Equity

• Apply the same three step bias calculation

This approach highlights both direction and conviction at the cohort level. A cohort can only show strong bullish or bearish bias if it is both leaning in one direction and allocating meaningful exposure relative to its combined equity.

Position Age Filter

When a timeframe filter is active (24h, 7d, or 30d), it controls which wallets qualify for the cohort, not which individual positions are counted within a qualifying wallet. This filter does not apply when all is selected; in that case, all wallets with open positions are included.

Wallet eligibility:

• A wallet is included if it has at least one open position whose open time falls within the selected timeframe.

• A wallet is excluded entirely if it has zero positions opened within the timeframe.

For included wallets, the full current state is used, not just the positions that triggered eligibility:

• Sum Open Value = the current total open notional of all open positions on that wallet, regardless of when they were opened.

• Sum Perp Equity = the wallet's full current perp equity.

Inclusion is all-or-nothing. A wallet either qualifies (at least one position in the timeframe, meaning all its open value and equity are counted) or it does not (nothing is counted). There is no partial inclusion.

Cohort Exposure is then calculated as Sum Open Value / Sum Perp Equity, computed over qualifying wallets only.

Why this matters: Using the Position Age filter to gate wallet eligibility (rather than to filter individual positions) ensures that cohort exposure reflects how leveraged the qualifying wallets actually are. If only newly-opened positions were summed, a wallet with large older positions would appear far less leveraged than it truly is, distorting the cohort signal.