Intraday Dy · 20 May 2026 close · t=1045 · Thu 21 May Entry · EXTREME BEAR · UNIVERSE-FILTERED · v9
Intraday Signal Dashboard Seven Sessions in Cash · vol_z Cooling · BNBR Streak Day 23
Tue chfwd backfill: BNBR ran +3.29% (Tue close 152 → Wed high 157) — modest miss — but Wed closed 148. Skip held cash, close-to-close correct again.
Today: 2 picks (BNBR, PACK), only BNBR in universe, combined mult 0.43 still below floor, override fails (ch1d −2.6%).
Regime mixed: regime_score_short−4.13 (barely up from −4.21), regime_score_long−3.83 (deeper from −3.20), but vol_z cooling to 1.36 (from 1.59) — first tentative reversal hint. Still extreme-bear → 0% cash. Seven consecutive sessions sitting out.
Cash sit-out streak
7 sessions
Last entry: Wed 13 May (+0.17%) Holidays: 14, 15 May Skips: 18, 19, 20 May (and counting)
10-day cum: ~+14.1%
⚠ BNBR streak hits day 23, cum −26.7% — vol_z cooling is the first reversal hint
Tue entry chfwd for BNBR (skipped): +3.29% (returnalt1 +2.45%). BNBR closed 152 Tue → Wed high 157 → Wed close 148 (−2.63%). Another modest chfwd miss that ended negative on close, so skip remained correct.
BNBR streak now 23 days, cum gain −26.73% — the drawdown keeps deepening, and prob1 has eroded from 0.74 to 0.68. The name is exhausted.
Regime nuance: regime_score_short flattened (−4.21 → −4.13) and vol_z cooled (1.59 → 1.36) — the first two-part hint that selling pressure may be easing. But regime_score_long deepened to −3.83 and breadth worsened to −0.66. Not a reversal yet — stay in cash, but raise alertness.
Dy ranked picks
2
BNBR + PACK
In universe
1 of 3
BNBR only
Deploy candidates
0
BNBR mult 0.43 < floor
vol_z (cooling)
1.36
down from 1.59 — mild +
Total deploy target
0%
extreme bear binding
📉 Regime extreme but flattening at the short end.regime_score_short ticked up from −4.21 to −4.13 (first non-deterioration in 5 sessions); vol_z cooled to 1.36 (from 1.59). But regime_score_long deepened to −3.83 and breadth to −0.66. All three categorical dims still bear. Extreme-bear tier → 0% mandatory cash, but watch for a turn.
⚪ Signal mix: 1 in-universe, 1 out.
BNBR (rank 1, prob 0.68, IN universe) — fails skip floor and override
PACK (rank 2, prob 0.51, OUT universe) — fell −9.7% Wed (RSI was overheated at 73 Tue)
CITY dropped out of the ranking today. Universe-deployable set remains empty.
Today's ranked picks
Thursday deployment: 0% (mandatory cash, seventh consecutive session).
BNBR combined mult 0.43 below 0.7 floor; override requires last ch1d > 0 (today's was −2.6%). Regime in extreme-bear tier. No deployable in-universe pick.
Tentative reversal hints (vol_z cooling, score_short flattening) are worth watching but not yet a trigger. Stay alert for: (a) signal broadening to ≥3 in-universe, (b) regime_score_short recovering above −3, (c) BNBR strong green day with prob ≥ 0.75.
Opportunity-cost note (out-of-universe). PACK finally cracked −9.7% Wed after running from 252 → 370 (RSI hit 73, overbought). CITY exited the ranking. The out-of-universe momentum names are starting to roll over too — the strength is fading market-wide, which is consistent with the deep-bear regime. Less opportunity cost to lament today.
Universe filter — today's picks
Universe list contains 282 tickers. Hard pre-filter at step 0.
Rank
Ticker
Name
Prob1
ch1d
In universe?
Next filter
Signal trajectory — last 8 sessions + holidays
Date
Period
Ranked
In uni
Deployed
Realized
Notable
8 May
t=1039
13
5
35-69%
+0.95% / +4.18%
MORA +20% missed
11 May
t=1040
1
1
0%
0%
BNBR chfwd +1.82%
12 May
t=1041
8
4 → 3
35%
+0.17%
BNBR +19.2% missed
13 May
t=1042
3
1 → 0
0%
0%
BNBR chfwd −0.57% (correct skip)
14-15 May
—
HOLIDAYS — Kenaikan Isa Almasih + cuti bersama
18 May
t=1043
1
1 → 0
0%
0%
BNBR chfwd +8.07% missed
19 May
t=1044
3
1 → 0
0%
0%
BNBR chfwd +3.29% missed
20 May
t=1045
2
1 → 0
0%
TBD
BNBR closed 148 (−2.6%)
Audit thought. The strategy has been in cash for ~7 trading sessions. BNBR has had three chfwd opportunities we passed on (+19.2% on 12 May, +8.07% on 18 May, +3.29% on 19 May). But on each occasion BNBR's close was negative — meaning hold-to-close with the size discipline would have lost. The override rule fires on prior-day positivity which BNBR keeps failing.
The structural question worth answering separately: is the BNBR-only-in-universe pattern a sign that the universe filter is too narrow for the current regime, or that the model's broader ranking is gaming the IDX small caps that aren't tradeable? Worth a one-time review of the 282-stock universe.
Tuesday t=1044 → Wednesday t=1045 chfwd
Rank
Ticker
Universe
Mon close
chfwd
returnalt1
Status
#1
BNBR
IN
Rp 152
+3.29%
+2.45%
Skipped — modest miss, closed −2.6%
BNBR Wednesday session breakdown
Metric
Value
Notes
Tue close (entry basis)
Rp 152
—
Wed open
Rp 148
gap −2.63% (big-down bucket → ×1.25)
Wed high
Rp 157
+3.29% chfwd
Wed low
Rp 142
−6.6% intraday drawdown
Wed close
Rp 148
−2.63% from Tue close
Streak
23 days
cum −26.73%
Counterfactual scenarios (if we had deployed BNBR @ 8%)
+3% TP filled (high 157)
+0.24%
−3% stop hit (low 142)
−0.24%
Hold to close 148
−0.21%
0% (actual)
0.00%
Marginal either way. Wed low at 142 (−6.6% from entry) would have tripped the stop before a +3% TP filled, depending on path. Skip stays correct.
Universe-cost on the day
PACK chfwd (Tue→Wed, out of uni)
+9.73%
CITY chfwd (Tue→Wed, out of uni)
+9.09%
But both rolled over Wed (PACK −9.7%)
momentum fading
Out-of-universe names had one more up-day then PACK cracked −9.7%. The market-wide momentum is fading — consistent with the deepening regime. Less to lament now.
10-day rolling cumulative
Entry day
Cum
Note
Through Wed 13 May entry
~+14.1%
last realized deploy
Mon-Tue 18-19 May (0%)
~+14.1%
idle
Wed 20 May (0% deploy)
~+13.5%
window dropped a winner day
Thu 21 May (0% planned)
~+13.5%
window now mostly idle; cum decaying slowly as good days roll off
Regime indicators — Tuesday vs Wednesday
Mixed signals — short-end flattening, long-end deepening.regime_score_short ticked up slightly from −4.21 to −4.13 (first non-deterioration in 5 sessions). vol_z cooled to 1.36 (from 1.59). But regime_score_long deepened to −3.83 and breadth to −0.66. Tentative bottoming hints, no confirmation.
Dimension
Tuesday 19 May
Wednesday 20 May
Δ
market_condition
bear
bear
same
regime_5h (tactical)
bear
bear
same
regime_25h (strategic)
bear
bear
same
regime_score_short
−4.210
−4.133
+0.08 (flattening)
regime_score_long
−3.196
−3.827
−0.63
vol_z
1.59
1.36
−0.23 (cooling)
breadth_score
−0.629
−0.664
−0.04
Reading the turn signals.
Two of the four reversal precursors I flagged Tuesday have now appeared: vol_z peaked (1.59 → 1.36) and regime_score_short flattened (−4.21 → −4.13). The missing pieces: regime_score_long is still deepening (−3.83) and breadth_score hasn't bottomed (−0.66).
So this is an early, partial bottoming signature, not a confirmed turn. The textbook sequence is vol_z peak → breadth bottom → score recovery. We have step 1; steps 2-3 are pending. Stay in cash but be ready to redeploy quickly if breadth turns up Thursday.
Regime tier — sizing decisions
Tier
regime_score_short
Multiplier
Today?
Bull / Neutral
≥ −0.5
×0.90 to ×1.10
No
Mixed bear
−1.0 to −0.5
×0.65 to ×0.75
No
Full bear
−2.5 to −1.0
×0.50
No
Deep bear
−3.0 to −2.5
×0.30
No
Extreme bear (today)
≤ −3.0
0% mandatory
YES
Thursday's recommended deployment (Thu 21 May entry)
✗ 0% deployment — full cash (seventh consecutive session)
✗ Only 1 in-universe pick (BNBR), combined mult 0.43 below 0.7 floor
✗ Override fails — BNBR ch1d Wed −2.63% (override needs prior ch1d > 0)
✗ Extreme-bear tier still binding (regime_score_short = −4.13)
⚬ Partial reversal hints (vol_z cooling, score_short flattening) — raise alertness but not yet a trigger
✓ Maintain ~+13.5% 10-day cum cushion
✓ Re-engage triggers: breadth_score turns up, OR regime_score_short recovers above −3, OR signal broadens to ≥3 in-universe
Sizing pipeline (v8 — no changes)
0. Filter to universe list (step 0 hard pre-filter)
1. Inverse-Volatility base
2. × streak_multiplier (fresh ×1.20, stalled ×0.85)
3. × regime_multiplier (extreme ×0% / deep ×0.30 / full ×0.50 / mixed ×0.65–0.75 / neutral ×0.9 / bull ×1.0)
4. × prob1_multiplier (prob < 0.65 → ×0.85)
5. × DOW adjustment
6. × vol-target overlay (target 2.5%)
7. × gap_multiplier (at next-day open)
8. Cap any single position at 20%
9. Skip if cumulative multiplier < 0.7
10. Override (discretionary): prob ≥ 0.75 + last ch1d > 0 + in universe → allow 5-8%
What's still working
✓ 7-session cash holding kept 10-day cum near +13.5%
✓ Override rule kept us out of BNBR each day it closed red (Wed close −2.6%)
✓ Universe filter holding; PACK rolled over −9.7% Wed (we never chased it)
✓ Extreme-bear tier still binding; first reversal hints appearing (vol_z cooling)
Watch Thu 21 May 09:00
1. New t=1046 ranked list — any signal broadening to ≥3 in-universe?
2. breadth_score — this is the missing reversal piece; a turn up confirms bottoming
3. vol_z — continued decline toward 1.0 reinforces the cooling signal
4. regime_score_short — recovery above −3 down-tiers to deep bear (×0.30)
5. BNBR — streak 23 days; a strong green day + prob recovery could finally reset it
6. Action item: review the 282-stock universe scope when regime calms
★ Gap-aware sizing — dormant.
No deploy candidates means the gap rule has no application. BNBR Wed gap was −2.63% (big-down bucket, would be ×1.25) — but the skip floor binds regardless.
|gap| > 2% lift
+62%
7.70% vs 4.75% mean chfwd
Mild gap-up [0,+0.5%]
3.78%
vs 5.90% baseline
Best cell
9.78%
big gap-up × low vol
BNBR Wed gap
−2.63%
open 148 vs prev 152 (big-down)
Direction matters?
No
big up ≈ big down
BNBR Thursday gap scenarios (override path)
Extreme-bear regime is binding — override path is effectively closed until score_short recovers. Documenting for completeness.
Thu open scenario
Open price
Gap
Override decision
Big down
≤ Rp 145
≤ −2.0%
Skip (override fails on red day)
Mild down to flat
Rp 146–148
−1.4% to 0%
Skip
Mild up zone
Rp 148.5–149
0% to +0.5%
Skip (mild-up penalty)
Modest up
Rp 149–151
+0.5% to +2.0%
Skip (extreme bear binding)
Big up
≥ Rp 152
≥ +2.5%
Consider 3-4% probe ONLY if score_short recovers above −3 AND breadth turns
Wednesday-morning operational checklist
09:00 — Pre-open observation (Thu 21 May)
1. Pull t=1046 ranked list — verify expectation (likely BNBR-only or narrow)
2. Check breadth_score: a turn up is the missing reversal confirmation
3. Check vol_z: continued decline toward 1.0 reinforces cooling
4. Default action: full cash
5. If signal broadens to ≥3 in-universe → reconsider down-tier deployment
6. If BNBR alone but with strong gap + score_short recovery + breadth turn → consider 3-4% probe only
Caveats & what to watch
✗ Partial bottoming signature (vol_z peak + score_short flat) but breadth_score not yet turned — wait for confirmation
✗ Three BNBR chfwd misses now (12, 18, 19 May) — but each closed red, so skips were correct
✗ Out-of-universe momentum (PACK) finally rolled over −9.7% — strength fading market-wide
✗ Be ready to redeploy quickly if breadth turns up Thursday — the turn could be sharp from these depths