This post-mortem was written from the summarized fill log. The full order log (54 fills, firm-reconciled; audit math in derive_campaign.py) corrects three facts cited below:
The narrative below is kept as written for the historical record. The corrected, full-log replay is at SMCI_20260611_VCET_VISUAL_REPLAY.
You had the right thesis (VWAP −0.5 SD reversion to 2-day VWAP / 4 SD target). You also FOUND the right entry (12:24 above +1 SD after the BB tightened). The v1.1 overlay would have done two things you didn't have today:
Result: you scalped a +$2.50 trade for +$0.025 per share and then chased the runner higher with smaller size. The conversion happened; you missed most of it.
You churned SMCI for ~4.5 hours. The biggest single mistake: 3:43-3:53 sequence at 12:30-12:36 where you bought the winner, sold 50% at +$0.04-0.05, then sold the rest at break-even, then chased back higher at 29.86 and stopped out at 29.815. The trade went from 29.56 entry to 32.00 close ($2.44 / +8.3%).
| Time | Action | Price | Shares | Inferred VCET reading |
|---|---|---|---|---|
| 09:53:19 | B SMCI | 29.54 | 50 | RED tape No fresh compression; chasing reversal off opening lows. |
| 09:53:26 | B SMCI (add) | 29.57 | 50 | Adding to a RED-tape entry. +3 min fade would have fired by 09:57. |
| 09:57:24 | S SMCI | 29.35 | 100 | -$0.21/share = -$21. Overlay would have shown +3 min fade label at 09:56 (close below entry). |
| 10:01:27 | B SMCI (re-buy) | 29.42, 29.35 | 90 | Same RED gate. Re-chasing. No CBO LONG yet. |
| 10:10:44 | S SMCI | 29.10 | 100 | -$0.25/share = -$25. Wrong-side: CBO probably broke SHORT (price below prior channel low). Overlay would have fired "wrong-side CBO - reduce / exit" within 5 min of 10:01 entry. |
| 10:33:09 - 10:46:37 | 3 re-entries | 29.41, 29.37, 29.50 | ~500 | Each entry inside the same RED-tape rotation. Bands flat-to-narrow but no compression-breakout pattern yet. |
| 10:46:13 - 10:46:37 | B SMCI (chase highs) | 29.41/29.49/29.50 | 300 | Chasing the 10:30 high - this is the “Chase / Extended” pattern. Setup mode would have shown UNKNOWN. Gate still RED. |
| 10:49:55 | S SMCI | 29.34 | 300 | -$0.15/share x 300 = -$45. Largest single loss of the morning. The 10:46 entry was THE chase the dashboard would have screamed RED at. |
| 11:13 - 11:42 | 4 more scalp attempts | ~29.37 - 29.43 | ~400 | Rotation-scalp behavior - getting into the rotation but exiting too fast. Overlay would have shown ROTATION sustained. |
The 09:53 - 11:40 window was extended-gap volatility with no compression-breakout signature. The chart overlay's dashboard would have shown Pre-trade gate: RED the entire morning. Every entry in this window was the kind of trade the Quality Gate explicitly says to pass or reclassify. The fill log shows you knew this was wrong - you kept stopping out and re-entering, which is the textbook "I want to be in this stock right now" trap from the Quality Gate doc.
Your narrative: "the BB got really tight... bought the 12:24 candle as it got above 1sd of vwap... all the MA alignment here a fash late was triggered... this is a reversion to the prior day vwap play."
That description maps exactly onto the true_pull_in_to_trend mechanism signature from the v1.1 plan:
| Signature element | What you saw | VCET v1.1 reading |
|---|---|---|
| Entry state: COMPRESSION or ROTATION | 2+ hours of consolidation (11:00 - 12:20); bands tight ("BB got really tight") | COMPRESSION at entry-5, transitioning |
| Band width compresses then expands | BB tight then expanding above | widthExpanding = TRUE at entry; width arrow ↑ on dashboard |
| VWAP slope accelerates in direction | VWAP turning up after sustained flat | slope crossing slopeFlatThresh upward |
| CBO break in trade direction | Close above prior consolidation high | CBO LONG fires - aqua triangle below the 12:24 bar |
| Price above +1 SD VWAP | Your entry trigger | distFromVwapSD ~+1.0 to +1.5 |
| State transition to TREND_UP within 15 min | Trend kicked in within 10 min of entry | EARLY_EXPANSION → EXPANSION_CONFIRMED → TREND_UP |
You bought 100 @ 29.585 at 12:24:52 + 100 @ 29.56 at 12:24:58 (total 200 shares around 29.57 avg).
Correction (2026-06-12, full order log): the 12:24 event entry was actually 100 shares wavg 29.570 (20+20 @ 29.585 + 60 @ 29.56); position peaked 120 sh only after the 12:30 re-adds. The exact 12:24→12:36 campaign P&L is −$16.79. This page's 200-share narrative came from the summarized fill log and is kept as written for historical context — see the corrected visual replay.
Then:
| Time | Action | Price | Shares | What VCET timer would have shown |
|---|---|---|---|---|
| 12:24:52-58 | Entry bar | 29.585 / 29.56 | 200 | Manual entry time = 12:24. Manual entry mode ON. |
| 12:27 (+3 min) | — | ~29.59 | — | +3 min label: "no early failure" (close ABOVE entry price) |
| 12:29:11 | S SMCI (1st partial) | 29.59 | 50 | +4.5 min - between checkpoints. You took +$0.005 / share (basically scratch). |
| 12:29:13 | S SMCI (2nd partial) | 29.60 | 50 | Took +$0.015 / share. You exited the 12:24 winner before the +10 checkpoint. |
| 12:30:23 | B SMCI (re-add higher) | 29.705 | 100 | Re-entering 14 cents higher than original. You're CHASING what you just sold. |
| 12:30:42 | B SMCI (more) | 29.725 | 20 | Adding more 17 cents above original. |
| 12:32:19 | S SMCI | 29.59 | 120 | Sold lower than re-add price. The 12:30 re-adds netted −$0.13 / share × 120 = -$15. |
| 12:32:20 | SS SMCI (SHORT!) | 29.61 | 120 | Flipping short - this is the dashboard's WRONG-SIDE warning territory. Setup mode would have been showing TREND (candidate) LONG; you went short on a normal post-breakout retest. |
| 12:32:26 | B SMCI (cover short) | 29.615 | 120 | Quick cover. Lucky scratch on the short - if SMCI had ripped from here you'd have been hammered. |
At +10 min from 12:24 entry, SMCI would have been around 29.71-29.75 area. Conditions checked:
At +15 min (12:39), SMCI was approaching 29.86 (where you actually re-entered and got stopped). Conditions still HOLD. Label fires "15 min: confirmation HOLDS" GREEN. Dashboard "Management" cell: HOLD CANDIDATE (trail wide).
At +25 min (12:49), SMCI was extending higher. State would have flipped from EXPANSION_CONFIRMED to TREND_UP. Label fires "25 min: TREND CONVERSION" GREEN.
The overlay would have explicitly told you to hold, three times.
After you exited the original 12:24 setup, you re-entered SMCI 5 more times:
| Time | Action | Price | Inferred state | Mechanism class |
|---|---|---|---|---|
| 12:36:16 | B 100 | 29.86 | EXPANSION_CONFIRMED long | trend_continuation_entry - state was already in expansion; chasing a confirmed move. |
| 12:36:22 | S 100 | 29.815 | (6 sec later) | Quick scratch. The chart overlay's dashboard would show "HOLD CANDIDATE" on entry; you ignored. |
| 13:43:00 | B 20 | 30.37 | TREND_UP already established | trend_continuation_entry (smaller size signaling reduced conviction) |
| 13:46:33 | S 20 | 30.373 | (3 min later) | Scratch. CBO had already fired LONG and was being retested. |
| 14:01:10-11 | B 40 | 30.625 | TREND_UP | trend_continuation_entry |
| 14:03:54-57 | S 40 (mixed venues, including SS) | 30.40 | (3 min later) | -$0.22 / share = -$9. Stopped out on normal post-breakout retest, then SHORTED. |
| 14:06:06 | B 20 (back to long) | 30.31 | TREND_UP (retest holding) | Whipsawed. |
The dashboard's state cell would have shown TREND_UP from 12:50 onward - you weren't entering a Pull-In setup, you were entering a continuation. The 4th mechanism class from the Mechanism Validation Plan: trend_continuation_entry - different management entirely:
You correctly recognized you weren't on a Pull-In anymore - the smaller size proves it. But you treated each continuation entry like a fresh scalp, exiting on noise. The overlay's "Management" cell would have shown HOLD CANDIDATE (trail wide) on each one if you'd been in manual entry mode.
| Rule | What the dashboard cell says | Action |
|---|---|---|
| If Pre-trade gate = RED | RED | Pass. The morning chop today was 10 RED-tape entries. The dashboard cell would have screamed this every time. |
| If +3 min checkpoint = "no early failure" | green label below bar | Do NOT scratch. The 12:29 partials were premature - the trade was passing the +3 min check fine. |
| If +10/+15 = "trend conversion candidate" AND CBO LONG was on entry bar | green label + dashboard "HOLD CANDIDATE" | Do NOT exit on a normal retest. 12:30-12:32 was a textbook breakout retest. The overlay would have said hold. |
You described the 12:24 setup as: "reversion to the prior day vwap... above 1sd of vwap... fash late triggered... all the MA alignment."
That's the true_pull_in_to_trend mechanism. The Mechanism Validation Plan §4 names it precisely:
entry state ∈ {ROTATION, COMPRESSION, EARLY_EXPANSION}
AND cbo_break_dir == trade_direction first confirmed in +5..+15 (POST-entry)
AND band_width(+5) >= band_width(0) * 1.5
AND slope(+5) - slope(0) >= +5°
AND expansion_score crosses >= 50
AND state -> EXPANSION_CONFIRMED / TREND matching dir within 15 min
=> true_pull_in_to_trend (HIGH confidence)
management: trend-hold; trail wide
You had ALL of these. The overlay would have lit up "trend conversion candidate" GREEN at +10 min. That's the trade you keep making and keep exiting.
You said in the chat: "that black square is where i want the hot key to get my fills of vwap - .5 and risk what ever and sell % at 2 @ and 4 SD and leve e runners 2 SD 2 day vwap imena"
That's the canonical Pull-In setup. The v1.1 overlay validates it:
So when (and only when) the dashboard shows GREEN gate + state in {COMPRESSION, ROTATION} + you're near VWAP −0.5 SD, fire the hotkey. The overlay tells you whether to push the button.
The chart overlay would not have prevented the morning hours of churn unless you respected the RED gate cell. Where it would have changed your trade is the 12:24 setup: it would have shown CBO LONG firing on the breakout bar, state transitioning to EARLY_EXPANSION, the +3 min checkpoint saying "no early failure," and the +10/+15 checkpoints saying "trend conversion candidate" GREEN. You scalped out 4-8 minutes after entry, missing 80% of a $2.50 trade. The overlay says HOLD when CBO + compression-precursor + slope-acceleration align. Tomorrow, watch the dashboard's "Management" cell - if it says HOLD CANDIDATE, do not exit on a normal retest. That is the single behavior change today's trade points to.