Issue 1,935 – Copyright, 2025, by Wyckoff AM Trader (New Version)
FROM THE DESK OF WYCKOFF TRADER (Read the Terminology at the bottom if you need clarification on WB-isms)
– 🐂 The Bull – Represents structural resilience, bid stacking, and market strength. If liquidity holds, the bull carries the weight
– 🐻 The Bear – Symbolizes liquidity exhaustion, price vulnerability, and market weakness. If bids vanish, the bear drags the market down.
– 👻 “GHOST – If the market doesn't have breath, it will fall due to the
lack of structure under the pressure of constant changing of belief.” – Wyckoff Trader
On Wednesday, the indexes remained buoyant thanks to a near-2% jump from Nivida (NVDA) to a record high and the first company to reach $4 trillion in valuation. Nvidia’s contract chip maker TSMC (TSM) posted a Q2 revenue of $31.9 billion, crushing market forecasts. The company posted a 38.6% rise in year-over-year sales. Everyone is singing the praises of the great one!
Check below this email to see a glimpse in the AI Logic Stack. That is what is running in RAM on those Nvidia’s chips.
Wyckoff Trader
P.S.
If you have not watched this video you might want to
Probably the best video on the big picture of AI and/in your future https://www.youtube.com/watch?v=iE1UwQj1kjo
Now let's get to the markets
📈 Today’s tape printed a textbook markup sequence under Composite control—effort matched result, structure held, and accumulation deepened without spill. Price tested conviction early with a soft dip, but operator support defended the spine and turned midday volume into targeted campaign lift. From open to last hour push, every print aligned with cause, rejecting noise and validating Wyckoff logic with precision
Big picture perspective
⚔️ Geopolitical Tensions, 🏆 Trade Tensions and Tariff Deadlines, 🛢️ US Fiscal Health Concerns,
– Monthly: Upthrust out of the last four months from last year. Closing at the high.
– Weekly: Holding the last two weeks high. Today may take price up higher or give it all away.
– Daily: Lifting support with closes at the highs.
It was a clock on day from turn to turn. As you can see below, the spill sealed at the open and then drifted down to the AM LOW. A bid entered in and the Composite Man lifted offers up to the lunch high. Then air got thin with thin trade. The close was hinting it could not hold the highs a price started to lose its structure at the closing prints.
🕒 SERIES S2L 🕒
|—————————————————————-|
| PREV DAY | HIGH | 3:55 | late | 6309 | 58,000 |(+22) 85m |
|—————————————————————-|
| OPEN | 6310 | 15,600 | (-09) |
|—————————————————————-|
| TURN | H/L | SEAL | BIAS | HNDLE | VOL | DELTA MIN |
|—————————————————————-|
| Spill | up | 9:30 | early | 6310 | 15,600 | (00) 00m |
|—————————————————————-|
| AM | LOW | 10:30 | center | 6299 | 13,300 | (-11) 60m |
|—————————————————————-|
| MID AM |
|—————————————————————-|
| lunch | high | 1:40 | center | 6331 | 9,000 | (+32)140m |
|—————————————————————-|
| mid pm | low | 2:55 | late | 6318 | 4,100 |(-13) 75m |
|—————————————- ———————–|
| LAST HOUR | high | 3:40 | late | 6323 | 48,700 |(+45) 105m |
|—————————————————————-|
PROSE AND THE NUMBAHS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Good morning. Today is SERIES S3L. I'm leaning bullish with search for structure. With all-time highs fever. the PEONS initiative drive could see some pull back into the spine of 6299 up to 6310 handle.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🌍 After Hours Globex: – That's the dog (CASH) but what did the tail (PEONS) do? —WB
1. Tone Last Night: (Liquidity)
🦘🐂 After a weak LAST HOUR HIGH the Aussies try to recover half of the lost gain and fail.
🏮 🐂 Asains take over and throw momma from the train. Loosing all the CASH gain. A bid enters at the 6287 and holds up to the 6302 handle.
🏛️ 🐂 PEONS hole the bid for a bit and then drop it lower.
2. Large Size: 🦘🐟 (no tuna), 🏮🐟 (21 tuna), 🏛️🐟 (0 tunas)
Absence of size = absence of conviction. Liquidity flow was skeletal—designed to shape expectation, not execute dominance.
3. 7:00 AM / Premarket: 💲 (0 tunas) Price is trading around Wednesday's low looking for a hero.
4. News: – 📰 (Treasury Statement at 2:00. All time EST)
5. Cycle Bias: 🏛️ 🐂 💲 Looking good, Louis! Looking good! All time high fever. Let's what that got today. Stalled at the gate. Price is trading at the Daily S2 pivot of 6280 handle. We just came off the weekly R2 handle. Time for a little pullback.
6. Honing (turns): Spill up, AM LOW, MID AM HIGH, lunch low, mid pm high, LAST HOUR LOW.
7. Previous Day's Numbers:
┌──────────────────────────── Session Summary ───────────────────────────┐
│ Weekly High 6335.50 Weekly Low 6296.00 │
│ Weekly Settle 6307.75 Final Settle 6320.25 (+12.50) │
│ Open 6307.75 Candle Body 12.50 Bullish │
│ High (10:07) 6335.50 Upper Wick 0.00 │
│ Low (14:23) 6296.00 Lower Wick 11.75 │
│ Session Range 39.50 Midpoint 6315.75 │
│ Volume 758,441 contracts Weekly Δ 39.50 │
└────────────────────────────────────────────────────────────────────────┘
8. Floor Trader Pivots:
🧱 Zones of Control Weekly Pivot (6242.00)
🔺 Bull Zone
R2 (6352.75)
R1 (6336.50)
⚖️ Decision Zone
Pivot (6316.75)
🔻 Bear Zone
S1 (6300.50)
S2 (6280.75)
Dates: 📅 July 10 Strawberry Moon Buck Moon p.m. Fall Solstice (Equinox) 📅 September 22 2:19 p.m. (all times Est)
********************* Terminology *********************
"Looking good, Louis! Looking good!" – Bullish (high confidence)
"Looking good, but…" – Neutral (cautious optimism)
"Things just got ugly, Louis. Very ugly." (Black Swan)
"It's looking real bad for the bulls, Louis. Real bad." – Bearish (pessimistic)
"It's looking real bad for the bears, Louis. Real bad." – Bullish (optimistic)
🐟 (Minnows)
⏳ Lifespan: 1–2 years
💧 Liquidity Impact: Momentary—tiny, rapid entries and exits
🔄 Interpretation: Their short biological life mirrors *fleeting market presence*. They appear as *blips* on the liquidity radar—numerous but not lasting.
🐟 (Sardines)
⏳ Lifespan: 3–4 years
💧 Liquidity Impact: Mid-range, directional waves
🔄 Interpretation: Like their lifespan, they sustain presence a bit longer—*waves of flow* with follow-through. Not permanent, but *tactically significant*.
🐟 (Tuna)
⏳ Lifespan: 15–40 years
💧 Liquidity Impact: Deep, anchoring, strategic
🔄 Interpretation: Their long biological endurance maps to *long-dwelling liquidity*. When they enter, they sit, influence, and stabilize. They're duration-defined presences**, just like their species.
🧖 "Dropping the Soap"
🏀 "Running with the ball"
👵 "momma got thrown from the train!"
📈 Up Turn – Higher highs and higher lows.
📉 Down Turn – Lower highs and lower lows.
↔️ Chop Turn – Bulls and Bears equally balanced.
💥 dropped it like it was hard (no bid stacking, no liquidity, hitting the bid)
🏀↓ smack it down or 🏀↑ bid it up (indicates a potential change in trend)
🧽 adding liquidity back into the market
Probably the best video on the big picture of AI and/in your future https://www.youtube.com/watch?v=iE1UwQj1kjo
On Wednesday, the indexes remained buoyant thanks to a near-2% jump from Nivida (NVDA) to a record high and the first company to reach $4 trillion in valuation. Nvidia’s contract chip maker TSMC (TSM) posted a Q2 revenue of $31.9 billion, crushing market forecasts. The company posted a 38.6% rise in year-over-year sales.
Let’s simulate how a 12,288-dimensional vector—the internal representation of a token in a large transformer model like GPT-3—evolves across layers. This is like watching a thought crystallize as it passes through the mind-machine. 🧠✨
🧩 Step-by-Step Evolution of a Token Vector
Let’s say the token is "Willis" from your earlier example. Here's how its vector might transform:
🔹 Layer 0: Embedding Layer
"Willis" is mapped to a learned embedding vector of shape 1 × 12,288.
This vector encodes static semantic features: name, proper noun, maybe pop culture associations.
🔹 Layer 1–4: Shallow Attention Layers
The vector begins interacting with nearby tokens like "talkin'", "bout", "Guy".
Attention heads shift weights to emphasize dialogue structure and speaker attribution.
The vector starts to encode contextual meaning: not just “Willis,” but “Willis being addressed.”
🔹 Layer 5–8: Mid-Level Abstraction
The model starts recognizing semantic patterns: this is a quote, possibly iconic.
The vector now encodes tone (skeptical, humorous), intent, and dialogue dynamics.
Positional encoding blends in—“Willis” is at the end of a sentence, which matters.
🔹 Layer 9–12: Deep Reasoning Layers
The vector begins to reflect cultural associations: sitcom, Gary Coleman, catchphrase.
Attention heads may activate long-range dependencies—linking "Willis" to "Guy" as speaker-target.
The vector now holds latent knowledge: “this is a famous line,” “respond with wit.”
🔹 Layer 13–Final: Output Construction
The vector is transformed into logits—probabilities for the next token.
It influences the model to generate a response like:
“Ahh, iconic! That phrase is pure 80s TV gold…”
🧠 Visualization Analogy 📈📉
Imagine each layer as a lens:
Early layers sharpen the image.
Mid layers add color and depth.
Late layers interpret the scene and decide what to say next.
Each transformation is a matrix multiplication + nonlinearity, reshaping the vector’s meaning space.
Want to simulate this with a different token—like a metaphor or byte-encoded mystery? Or visualize how attention heads shift across layers for "Willis"? I can sketch that next.
Comments are closed