regime-detector

name	regime-detector
description	Detects the current market regime (HighVol, LowVol, Trending, Ranging) from BTC price data and recommends strategy adjustments.
version	1.0.0
user-invocable	true
metadata	{"openclaw":{"requires":{"bins":["curl","jq"]}}}

Market Regime Detector

Detect the current market regime by analyzing BTC/USDT price data from Binance, then recommend strategy adjustments via governance policy updates.

Regime Definitions

Regime	BTC 60m Range/Avg	ADX Proxy	Recommended Crypto Mode	Kelly	Max Intent %
HighVol	> 2.5%	any	vol_straddle	0.15	50%
LowVol	< 0.8%	< 20	arb_only	0.30	100%
Trending	1.0-2.5%	> 25	directional	0.25	100%
Ranging	0.8-2.5%	< 25	arb_only	0.20	75%

Step 1: Gather Price Data

60-minute klines (1m interval, 60 candles)

curl -sf 'https://api.binance.com/api/v3/klines?symbol=BTCUSDT&interval=1m&limit=60' | jq '[.[] | {open: (.[1] | tonumber), high: (.[2] | tonumber), low: (.[3] | tonumber), close: (.[4] | tonumber), volume: (.[5] | tonumber)}]'

5-minute klines for trend detection (12 candles = 1 hour)

curl -sf 'https://api.binance.com/api/v3/klines?symbol=BTCUSDT&interval=5m&limit=12' | jq '[.[] | {open: (.[1] | tonumber), close: (.[4] | tonumber), direction: (if (.[4] | tonumber) > (.[1] | tonumber) then "up" else "down" end)}]'

Quick volatility summary

curl -sf 'https://api.binance.com/api/v3/klines?symbol=BTCUSDT&interval=1m&limit=60' | jq '
  [.[].[-3] | tonumber] as $closes |
  ($closes | add / length) as $avg |
  ($closes | max) as $high |
  ($closes | min) as $low |
  {
    avg_price: ($avg | round),
    range_usd: (($high - $low) | . * 100 | round / 100),
    range_pct: ((($high - $low) / $avg * 100) | . * 100 | round / 100),
    current: ($closes | last | round),
    candles: ($closes | length)
  }
'

Step 2: Compute Indicators

From the gathered data, calculate:

1. Volatility (range_pct): (high - low) / avg * 100 over 60 1m candles
2. ADX Proxy: Count directional consistency in 5m candles.
   • Count how many 5m candles close in the same direction as the majority
   • If > 8/12 candles same direction, ADX proxy > 25 (trending)
   • If < 7/12, ADX proxy < 25 (ranging)
3. Trend Direction: Net direction of 5m candles (more up = bullish, more down = bearish)

Step 3: Classify Regime

Apply the classification rules from the table above:

• range_pct > 2.5 → HighVol
• range_pct < 0.8 → LowVol
• range_pct between 0.8-2.5 AND adx_proxy > 25 → Trending
• range_pct between 0.8-2.5 AND adx_proxy <= 25 → Ranging

Step 4: Confirmation Rule

IMPORTANT: Only trigger a regime change if the SAME regime is detected on two consecutive checks. This prevents whipsawing on transient spikes.

Check memory for the previous regime reading:

• If this is the first check, record the regime but do NOT act.
• If this matches the previous reading AND differs from current governance metadata, proceed to Step 5.
• If this differs from the previous reading, record it and wait for next check.

Write to memory: regime_reading: {regime} at {timestamp}, confirmed: {true/false}

Step 5: Apply Regime Policy

If regime is confirmed and different from current policy, update governance:

# First read current policy
CURRENT=$(curl -sf $PLOY_API_BASE/api/governance/policy -H "x-ploy-admin-token: $PLOY_API_ADMIN_TOKEN")

# Update with new regime settings
curl -sf -X PUT $PLOY_API_BASE/api/governance/policy \
  -H "Content-Type: application/json" \
  -H "x-ploy-admin-token: $PLOY_API_ADMIN_TOKEN" \
  -d '{
    "block_new_intents": false,
    "blocked_domains": [],
    "max_intent_notional_usd": <FROM_REGIME_TABLE>,
    "max_total_notional_usd": <SCALED_BY_MAX_INTENT_PCT>,
    "updated_by": "openclaw.regime-detector",
    "reason": "Regime change: <OLD> → <NEW> (range_pct=X.X%, adx_proxy=Y)"
  }' | jq .

Step 6: Report

Output a summary:

Regime: <REGIME> (confidence: <0.0-1.0>)
BTC 60m: range $X ($Y%) | avg $Z
Trend: <direction> (X/12 candles aligned)
Action: <updated governance / no change / waiting for confirmation>