Run census

Momentum runs: 31 (~31 per session)
Run length distribution: {2: 3, 3: 1, 4: 4, 6: 2, 7: 2, 8: 3, 9: 1, 13: 2, 15: 1, 17: 4, 18: 4, 20: 4}
Runs lasting <= 2 candles: 3 (9.7%)
Direction at start: {'DOWN': 18, 'UP': 13}
Boundary structure at start: {'none': 31}

Self-consistency

Self-consistency (labeled direction vs the run's own realized drift):
  mismatched runs: 16/31 (51.6%) | by run length: {'>2': 15, '<=2': 1}

Predictiveness

Predictiveness of run-start direction (signed forward return):
  +5 candles: hit rate 35.5% over 31 runs, average signed return -0.00031
  +10 candles: hit rate 48.4% over 31 runs, average signed return -0.00010
  by starting state (+10 candles):
    STARTING_DOWN      n=  11 hit= 45.5% avg=-0.00011
    STARTING_UP        n=   6 hit= 50.0% avg=+0.00028
    CONSOLIDATING      n=   5 hit= 40.0% avg=-0.00070
    DECELERATING_DOWN  n=   4 hit= 50.0% avg=-0.00034
    ACCELERATING_UP    n=   3 hit= 33.3% avg=+0.00014
    DECELERATING_UP    n=   2 hit=100.0% avg=+0.00049
  by boundary trigger (+10 candles):
    non-structural     n=  31 hit= 48.4% avg=-0.00010

Boundary churn

Boundary churn:
  direction reversals between consecutive runs: 17
  flip-flops (reversal after a run of <= 2 candles): 1
  flip-flops by the reversing run's boundary trigger: {'non-structural': 1}

Change rate

Absolute change_rate at run start (vs MOMENTUM_MIN_CHANGE_RATE=0.0001):
  DOWN           n=  18 median=0.000197 below threshold: 22.2%
  UP             n=  13 median=0.000246 below threshold: 23.1%

Anchor angle

No anchor-angle observations to measure.

Velocity quality

Kinematic velocity quality (recency-weighted velocity / noise, #642), signed +10-candle forward return:
  |q| 0.0-0.1    n=   81 hit= 66.7% avg=+0.00042
  |q| 0.1-0.2    n=   77 hit= 51.9% avg=+0.00006
  |q| 0.2-0.3    n=   89 hit= 65.2% avg=+0.00022
  |q| 0.3-0.4    n=   79 hit= 54.4% avg=+0.00011
  |q| >= 0.4     n=   53 hit= 49.1% avg=+0.00013

Occurrence census

Trend occurrences: 160 (~160.0 per session)
Occurrence length distribution (candles): {1: 148, 2: 6, 4: 1, 5: 1, 32: 1, 33: 1, 40: 1, 60: 1}
Occurrences lasting <= 5 candles: 156 (97.5%)
Direction at start: {'UP': 88, 'DOWN': 72}
Starting state: {'BULLISH_EMERGING': 88, 'BEARISH_EMERGING': 72}

Self-consistency

Self-consistency (labeled direction vs the occurrence's own realized drift):
  mismatched occurrences: 8/160 (5.0%)

Predictiveness

Predictiveness of occurrence-start direction (signed forward return):
  +10 candles: hit rate 45.3% over 150 occurrences, average signed return +0.00008
  +30 candles: hit rate 53.7% over 134 occurrences, average signed return +0.00031
  by starting state (+30 candles):
    BULLISH_EMERGING     n=  77 hit= 68.8% avg=+0.00086
    BEARISH_EMERGING     n=  57 hit= 33.3% avg=-0.00044

Occurrence churn

Occurrence churn:
  direction reversals between consecutive occurrences: 12
  flip-flops (reversal after an occurrence of <= 5 candles): 8

Momentum-trend alignment

Momentum-trend alignment (momentum run starts, signed +5-candle forward return):
  with-trend     n=  25 hit= 32.0% avg=-0.00041
  counter-trend  n=   5 hit= 60.0% avg=+0.00029
  no-trend-yet   n=   1 hit=  0.0% avg=-0.00102

Trades: 21 | wins 9 (42.9%) | net +238.31
Positions (round trips): 21 | wins 9 (42.9%)

Session and cohort net PnL

  qqq_2026-04-09     trades=  21 wins=  9 net=   +238.31  [training]
  training   ( 1 sessions): trades=  21 wins=  9 net=   +238.31
  all        ( 1 sessions): trades=  21 wins=  9 net=   +238.31

Net PnL by session regime

  RANGE          ( 1 sessions): trades=  21 wins=  9 net=   +238.31

By exit path

  STOP_MARKET    n=  12 wins=  0.0% pnl=   -375.32
  TAKE_PROFIT    n=   9 wins=100.0% pnl=   +613.63

By direction

  DOWN           n=  11 wins= 45.5% pnl=    +98.29
  UP             n=  10 wins= 40.0% pnl=   +140.02

Hold duration in candles (all closes): {0: 7, 1: 8, 2: 2, 3: 1, 5: 1, 10: 2}
Hold duration in candles (stop-outs):  {0: 7, 1: 2, 2: 1, 10: 2}
Longest hold: 10 wall-clock minutes | cross-session holds: 0 (must stay 0 with the session flush enabled)

Round-trip friction / 1R: median=0.13 p90=0.22 (#612 target: <= 0.20)
Stop distance as fraction of entry price: p10=0.00016 median=0.00028 p90=0.00044
Entry risk/reward ratio: min=1.55 median=2.16 max=4.62
Maximum favorable excursion before stop-out (in R): median=0.00 p75=0.84 | >=0.5R 33.3%, >=1R 16.7% of stopped trades

MAE/MFE in R (closed-candle excursions vs initial risk)

  winners n=   9 | MAE median=0.00R p75=0.00R p90=0.00R | MFE median=2.02R p75=2.88R
  losers  n=  12 | MAE median=1.06R p75=1.21R p90=1.27R | MFE median=0.00R p75=0.88R

Win rate by entry-confidence decile

  0.4-0.5: n=   7 wins= 42.9%
  0.5-0.6: n=  11 wins= 36.4%
  0.6-0.7: n=   3 wins= 66.7%

Entry expectancy by 30-minute session bucket (minutes since open)

     0-30   n=  2 wins=  50% net=    +0.83 meanR=+0.98
    30-60   n=  3 wins=  33% net=    -5.97 meanR=-0.24
    60-90   n=  1 wins= 100% net=   +55.99 meanR=+2.38
    90-120  n=  1 wins=   0% net=   -22.55 meanR=-1.18
   120-150  n=  3 wins=  33% net=   +90.09 meanR=-0.21
   180-210  n=  2 wins=  50% net=    +0.25 meanR=+0.12
   210-240  n=  2 wins= 100% net=   +85.14 meanR=+2.40
   240-270  n=  1 wins=   0% net=   -18.00 meanR=-1.24
   270-300  n=  3 wins=  67% net=  +119.02 meanR=+1.03
   300-330  n=  2 wins=   0% net=   -35.31 meanR=-1.25
   330-360  n=  1 wins=   0% net=   -31.19 meanR=-1.13

Churn checks (both should stay at 0 after the 2026-07-08 fixes)

  REDUCE fills on in-profit positions: 0 (of 0 REDUCE fills)
  Same-direction re-entries within 3 candles of a stop-out: 0 (of 12 stop-outs)

Bracket vs realized outcome

Filled candidates: 21
Win-direction agreement rate: 81.0%
Realized-R mean absolute error: 0.761

Regime

Regime: RANGE
Signed efficiency: 0.0648
Mean absolute return: 0.0003
Flip density: 0.5399