Execution Intelligence: A Self-Prompting Trading System for ChatGPT

Institutional logic, modular AI design, and verified execution rules—documented for traders who demand structure.

From screenshot to single, risk-defined trade—no code, no APIs, and no agents.

Most traders drown in conflicting indicators, paywalls, and prompt fatigue. Execution Intelligence (EI) compresses market context—Market/Volume Profile, VWAP (with Sigma deviations), Bookmap-style liquidity, and delta/momentum—into one defendable action.

Built for ChatGPT, EI is a modular, self-prompting engine that remembers session intent, audits its own calls, and outputs a standardized TH.Ticket (Lean, Stop+1t, Abandon rule, RR≥2:1, Targets, Time-Box, BNI sanity, Confidence). Upload screenshots, pick a timeframe mode (Trend, Balance Reclaim, Re-entry), and get an audited ticket in seconds. No black boxes. Just institutional-style execution—inside a chat.

How it works

Screenshot → Synthesis → Ticket

  1. Screenshot Intake
    Provide 1D/1H/5m/1m context (MES or MNQ). The intake layer tags structure, pressure, and flow cues from each image.

  2. Module Chain
    EI routes through Auction Framework, VWAP Matrix™ + Sigma Snap Reversal™, Market/Volume Profile (VPOC/HVN/LVN), Liquidity Engine (heatmap/iceberg behavior), Momentum Delta (initiative/divergence), Indicator Adapter where needed, and Cognitive Guardrails.

  3. Confluence Score
    The router weights agreement—location vs value, slope, reclaim/flip logic, liquidity alignment, and delta confirmation—to favor the simplest valid path.

  4. TH.Ticket
    One position with a named lean (hard gate), defined stop (+1 tick), Abandon rule (time or structure), RR≥2:1, targets, Time-Box (“expected to resolve in 45–120 min”), BNI gate (news/calendar sanity), and a confidence band.

  5. Audit
    Every ticket logs inputs and module verdicts for later review—so you can measure what’s working and prune what isn’t.

Why it’s different

Self-Prompting, not prompt-dependent

EI chains its own prompts under the hood. You don’t babysit mega-prompts; you give it context, it asks for what’s missing (e.g., “choose When-frame”), then it executes the reasoning path. The memory/recall layer preserves the module registry, router map, and ticket schema—so new chats stay consistent without re-teaching PDFs every time.

Modular & deterministic

Each module has an explicit contract. If a signal is unknown, EI says unknown. If confluence is <70%, EI either asks for more data or offers a conditional plan (“valid if reclaim VWAP and hold above σ1 for two 5-min closes”). That’s the difference between trading and storytelling.

Risk & time management (the part most tools skip)

Lean-Against + Time-Box. Every trade must defend a named lean (e.g., prior-day VPOC, LVN shelf, σ2 band). No lean, no trade. The Time-Box sets expectations up front (“balanced range; 2–6h resolution”), helps you manage attention, and avoids profit bleed from zombie holds. If price violates the lean or time-box expires without progress, EI triggers Abandon—protecting both equity curve and headspace.

What markets & tools does it understand?

MES and MNQ first for intraday swings and day-holds. It reads Market Profile structure (value, VPOC, HVN/LVN edges), VWAP + Sigma deviations (Sigma Snap Reversals™, reclaim/flip logic), Bookmap-style liquidity behavior (icebergs, spoof zones), and delta/momentum patterns (initiative, exhaustion). You upload screenshots; EI handles the rest.

Core modules (brief)

  • Auction Framework — Balance → Trend → Reclaim mapping; identifies the auction state and likely path of least resistance.

  • VWAP Matrix™ + Sigma Snap Reversal™ — Deviation bands + re-entry logic for clean fades/reclaims; defines lean and re-test expectations.

  • Market/Volume Profile — Value migration, HVN/LVN shelves, prior-day/IB anchors; converts structure into actionable lean zones.

  • Liquidity Engine — Heatmap-style reading of absorption/sweeps; spot iceberg defense, spoof pulls, and vacuum runs.

  • Momentum Delta — Initiative vs absorption, divergence and pressure thresholds; confirms or vetoes entries.

  • Indicator Adapter — Adapter to translate simple indicators (e.g., stochastic, MTF trendline confluence) into deterministic checks without diluting structure-first logic.

  • Price Discovery Path (PDP) — Converts the confluence picture into a small set of ranked, condition-aware paths; embeds Abandon, flip, and time-box rules directly into the ticket.

  • Cognitive Guardrails — Blocks low-probability impulse trades; enforces “No Lean, No Trade,” RR≥2:1, and re-engagement only after state change.

  • BNI (Bias & News Intelligence) — Quick sanity pass for calendar events and narrative drift so you don’t lean into a headline.

Self-healing memory & recall (why sessions stay consistent)

EI carries a lightweight Self-Healing Memory Recall System that retains module choices, router preferences, and your last valid lean schema. If a chat resets, EI reconstructs the working set from your install PDFs and recent tickets, then verifies against current screenshots before re-using anything. The result: less re-training, more continuity—and no “prompt pack decay.”

Example ticket (abbreviated)

  • State: Balance → Reclaim attempt above prior-day value.

  • Lean: Prior-day VPOC with σ1 VWAP band overlap.

  • Entry: Reclaim + hold two 5-min closes above σ1 with delta confirmation.

  • Stop: σ1 fail +1 tick.

  • Targets: LVN shelf, then HVN edge; trail to σ2 on momentum push.

  • RR: 2.6:1 to T1, 4.1:1 full.

  • Time-Box: 90 minutes; Abandon on failed hold or delta flip.

  • Confidence: 0.76 (VWAP + Profile agreement; liquidity neutral).

Installation (5–10 minutes)

  1. Paste the Main Drive prompt.

  2. Upload the PDF Architecture™ set.

  3. Run the Memory/Recall priming step.

  4. Send screenshots (1m/5m/1h/1D).

  5. Choose mode (Trend, Balance Reclaim, Re-entry). Receive TH.Ticket.

What you get

  • TH Architecture PDFs™ (modules + router flow)

  • Main Drive (Install Sequence) + Memory/Recall routine

  • Quick-start video and how to use the system (10 minutes: from screenshots to ticket)

  • Ticket schema + audit template

  • Lifetime updates to this version

  • Built in tutor layer and “trader jargon” reducer for newer traders

Pricing & license

$147 one-time (Beta). Personal license. Educational decision-support only. Free updates within this version.

FAQ (short)

Is this a bot?
No. It’s a reasoning shell that reads your screenshots and outputs an audited plan.

Does it require APIs or agents?
No. It’s designed for ChatGPT and screenshots—lightweight, portable, and fast.

What instruments?
MES, MNQ first; stocks, commodities (CL, GC, Bonds) and crypto that respect VWAP/profile structures also work.

Win-rate promise?
No promises. The system enforces discipline (Lean, RR≥2:1, Abandon) and avoids low-probability trades—that’s where the edge lives.

Disclaimer

Trading involves risk. Past performance does not guarantee future results. Futures trading involves substantial risk. TradingHouseGPT is an educational decision-support tool, not investment advice.

Continue here —> [FAQs] | [The Blunder] | [Shop]