risk-mgmt

Risk Intelligence — concentration, consistency, anomalies in one score

How HHI / CV / z-scores translate into a single Portfolio Risk Score you can act on

6 min · intermediate

What you'll have when finished

  • Understand what your Risk Score (0–100) is actually measuring
  • Read HHI values without needing to remember the table
  • Spot whether you have a venue / pair / chain / strategy concentration problem
  • Distinguish "conviction trade" anomalies from "impulse trade" anomalies

Before you start

  • A high Risk Score is BETTER (lower risk). The numerical scale is intentionally inverted from "scary" intuition.
  • CEX-only metric. DEX bots use your wallet directly so the "drift" concept doesn't apply here.

Walkthrough

  1. The Portfolio Risk Score (hero card) — what 0-100 means

    One number that summarizes your trading risk profile. HIGHER = LOWER RISK. **75-100 STRONG** — Well-diversified across venues / pairs / strategies. Consistent activity. No anomalies. Keep doing what you're doing. **55-74 MODERATE** — Some concentration or moderate variance. Probably fine for most users. Worth reviewing the sub-scores. **35-54 ELEVATED** — One or more risk dimensions are flagging. Check the sub-scores to see which. **0-34 HIGH RISK** — Concentrated, inconsistent, or anomalous in a way that warrants action. The score is a transparent weighted average of four sub-scores. There's no AI black box — every number traces back to a published formula. Click the Methodology tab to see the math.

    Success criteria: Read your overall score · Identified which band you're in

  2. The four sub-scores — diagnose where the risk is

    Right side of the hero card shows four mini-cards: **Concentration** — derived from HHI across exchanges + pairs. Lower HHI = better diversification = higher sub-score. **Engagement** — active days as a fraction of the window. Passive/sporadic = lower. Daily-active = higher. **Consistency** — derived from CV (coefficient of variation) of daily volume. Steady daily volume = higher. Wildly swinging = lower. **Anomaly-free** — fraction of trades within 2σ of your typical size. Outliers drag this down. If overall score is low, ONE of these is usually the culprit. Find which and address it.

    Success criteria: Identified which sub-score is your weakest

  3. HHI reading without a cheat sheet

    HHI (Herfindahl-Hirschman Index) measures concentration. The number is between 0 and 10000. **0-1500 MINIMAL/LOW** — Highly diversified. No single venue/pair/chain dominates. **1500-2500 LOW** — Well distributed. **2500-5000 MODERATE** — Some concentration. Usually fine but worth knowing. **5000-8000 HIGH** — One venue/pair clearly dominates. Review carefully. **8000+ EXTREME** — Almost everything in one bucket. Single point of failure. The Overview tab shows HHI for Exchange, Pair, Chain, and Strategy separately. Often one is fine and another is extreme — that's the one to address.

    Success criteria: Read all 4 HHI numbers · Identified which dimension is most concentrated

  4. Concentration tab — visual proof

    Visualizes the four HHI dimensions: **Volume Distribution by Pair** (horizontal bar) — top 10 pairs by USD volume. The first bar being >50% of the total is a clear signal. **By Exchange (CEX)** (pie) — where your CEX activity lives. Single dominant exchange = key-risk concentration. **By Chain (DEX)** (pie) — Ethereum / Base / Polygon / Arbitrum / BSC / etc. Heavy gas chains should be a minority of swap count unless every trade is large. **By Strategy** (pie) — CEX / Smart Trade / Algo / DEX surfaces. Diversification across surfaces matters too.

    Success criteria: Spotted at least one over-concentrated dimension

  5. Anomalies tab — conviction trade or impulse trade?

    Lists every trade whose size was >2 standard deviations from your typical size. These trades dominate your P&L exposure. For each anomaly, ask: - **Was it planned?** A pre-decided large conviction trade is fine. Journal it for future review. - **Was it impulse?** FOMO into a pumping coin, revenge after a loss, or "this time is different" — these are the trades that blow up accounts. A high anomaly count is not automatically bad. But you should be able to NAME each one before it disappears off the recent list.

    Success criteria: Reviewed your anomaly list · Could name the reason for each

  6. Recommendations tab — what to actually do

    Plain-English flags the page generated based on your activity. Examples: - "63% of your volume is on Coinbase. If you lose API access, you lose access to most of your activity." - "85% of your volume is in ETH/USDC. Diversifying into 2-3 uncorrelated pairs reduces shock risk." - "You traded on 6 of 30 days. If you're trying to maintain a strategy, consider scheduled execution (Algo Orders TWAP/DCA)." Each recommendation has a specific concrete action. Pick the one you can act on this week.

    Success criteria: Picked one recommendation to act on this week

  7. Methodology tab — read the formulas

    Every metric has an open formula. The math isn't hidden behind "AI confidence scoring" or proprietary indices. If you want to know: - Why HHI = 5400 → it's the sum of squared market shares × 10000 - Why CV = 1.8 → standard deviation ÷ mean of daily volume - Why Risk Score = 62 → weighted average of the 4 sub-scores Click Methodology. Everything is there. Compare against your own calculation in a spreadsheet if you want.

    Success criteria: Skimmed the methodology tab once

What's next

If concentration is the issue: - Route some trades through a second exchange ([Settings → Connections](/settings)) - Add 2-3 new pairs to your active list - Switch some DEX activity to Base/Arbitrum if you're Ethereum-heavy If consistency is the issue: - Use Algo Orders → DCA mode for scheduled execution - Set Smart Trade position caps so trade size stops being arbitrary If anomalies are the issue: - Journal every >2σ trade in [Trading Journal](/journal) - Consider a "trade-by-trade approval" pause if impulse trades dominate