By: Old West Solutions
|
A terminal for traders who would rather know than guess. Test what you are about to do against the conditions that actually occurred — then act on what the record tells you, not on a feeling.
II/Execution
Interactive Brokers execution, embedded directly into the workflow.
Wallet view · account context
Margin and buying power refresh on every quote tick — execution sizing stays inside live account limits.
Options chain · live contract pricing
SPY · Apr 18 · 480.12 · +0.34%| Strike | Bid | Ask | Last | Spr |
|---|---|---|---|---|
470 | 12.10 | 12.45 | 12.40 | 0.35 |
475 | 8.65 | 8.92 | 8.85 | 0.27 |
480 | 5.46 | 5.72 | 5.60 | 0.26 |
485 | 3.04 | 3.22 | 3.15 | 0.18 |
490 | 1.54 | 1.69 | 1.62 | 0.15 |
Quotes shown before order submission for price-aware execution. ATM row pinned for the working strike.
Quote latency
14 ms
Route
SMART
Pre-trade checks
passed
III/The model
One normalized market state.
One supervised model.
Every decision measured against what actually happened.
NVDA
$104.75Model probability
42.0%
Expected return
+0.0%
Confidence
High
Powered by TradingView Lightweight Charts™
Model decision
09:42:00 ET
Prediction, adjusted for what the model gets wrong.
Model
-2.5%
Data
-0.6%
Regime
-0.1%
Execution
-2.1%
Why the model moved
51.0% raw → 42.0% final
Current regime
Trend continuation
Expected vs realized
Model error is visible by design. The system earns trust by showing where estimates diverged from outcomes.
Model calibration
- 35% predicted50% observedn=10
- 45% predicted40% observedn=10
- 55% predicted63% observedn=8
- 65% predicted89% observedn=9
- 75% predicted8% observedn=8
- 85% predicted8% observedn=8
Historical validation
Train 2019-01-02–2023-12-29 · Val 2024 · Test 2024+
Every prediction has a past.
Every model decision is tied to the exact point-in-time feature state used to produce it.
Prediction is only the first number.
General.Exchange measures what the model expected against what actually happened. Every signal is calibrated, stress-tested, and evaluated against point-in-time historical data before it becomes part of the decision layer.
Supervised learning · Point-in-time data · Walk-forward validation
IV/Inference
From market state to trade signal — only when the probability is worth acting on.
No signal
68.6%
Threshold 70.0%
Not enough edge to surface.
Inference trace
Real-time inference
Market state is evaluated first. Only context relevant to the current regime enters the decision layer.
Context without the weight
RAG retrieves only the information relevant to the current setup. Cached context keeps inference light, fast, and repeatable.
Selective signals
Calibrated probability, expected return, regime, and execution constraints determine whether a candidate survives.
Traceable decisions
Every inference carries its market snapshot, retrieved context, model version, and final decision.
V/Research Library
Build, fork, and govern strategy research in one institutional library.
Folders
Mean Reversion
14 strategies
Momentum Breakout
9 strategies
Earnings Plays
22 strategies
Workstation
Momentum Breakout · 9 strategies
Gamma Squeeze Fade
TSLA
Premium Harvest
SPY
Earnings Straddle
AAPL
Vol Contraction Play
QQQ
VI/Newspaper
The market newspaper for decision-makers.
Today's lead story
Futures hold gains as commodity curves tighten and macro liquidity rotates back into risk.
Our morning desk read shows participation broadening across cyclicals while index downside hedging fades. The notable change is concentrated in commodities-linked transport and energy proxies.
US 10Y
4.18%
-4 bps
WTI
$79.42
+1.7%
DXY
103.1
-0.3%
Sections
Market Pulse
Who added risk, who reduced.
Flows & Positioning
Institutional pressure map by sector.
Volatility Desk
Regime shift watchlist and catalysts.
Commodities Tape
Term structure and basis stress scan.
Market narrative
Session-level context on participation, pressure, and structural changes that matter now.
Flow intelligence
Institutional prints, options flow, and liquidity shifts translated into an actionable brief.
Execution context
Where conditions are favorable or hostile before risk is committed to the tape.