Finance · AI markets intelligence
Home Loading Waiting for data

Overview

THELMA Trade OS

Loading operating brief...

Public weights, public filings, daily AI markets signals. Approval-only; no live order execution. Portfolio data is shown as percentages, not share counts or account values.

Today's Decision Stack

Macro regime loading

Loading the current operating brief.

Thelma ranks public-market signals, flow, ownership, macro, and risk into approval-only portfolio targets.

Investor Brief

What Matters Before The Cockpit

Waiting for data

Robert Carver Positioning

Loading live futures book

Action Blotter

0 positions

Add / Trim

Ex-cash weight · current → after move
How to read the Carver sizing line

Example — “Carver forecast +5.5/20 at 66% vol → target 1.0%; buffer ±0.2%”:

  • Forecast +5.5/20 — combined signal (trend + cross-sectional alpha + composite score), scaled so a typical reading is ~10 and capped at ±20. Positive leans long, negative trims toward flat. +5.5 is a mild long.
  • at 66% vol — the name’s annualized volatility. Sizing is volatility-targeted, so higher-vol names get smaller weights for the same forecast.
  • → target 1.0% — the resulting ex-cash portfolio weight: (forecast ÷ 10) × (risk budget per name ÷ vol), floored at 0 (long-only) and capped per name.
  • buffer ±0.2% — a no-trade band around the target. We only trade when the current weight leaves target ± buffer, and then only to the nearest band edge — which cuts needless churn.

Carver Market Read

Rob Carver live futures positioning

Performance Return Proxies

3M / YTD / 1Y current-weight proxy

Today's Trades

0 items

Signal Strength Map

Score vs number of managers

Top Return Drivers

5-day lift and drag

Portfolio Evolution

Current, peer average, and target

Under/Overweight Gaps

Your weight vs focus funds

Study Queue

Positions that need more work

Data Freshness

Source health for this run

Risk Controls

Caps applied before trades publish

Current Portfolio

Weights only, no account values

Trade Feed

Add/trim instructions sized by portfolio weight

Study Queue

Thesis, trigger, and risk

Earnings & Catalysts

Upcoming dates, result markers, and guidance signals

Manager 13F Radar

Loading public filings

Focus Fund Tracking %

Full AI Thesis Core positions

Crowded Focus Positions

Most shared public 13F holdings

Filing Status

Latest public filing date

Insider Form 4 Radar

Loading public insider trades

Key Insider Trades

Actual open-market buys and sales, ranked first

Company Net Read

Buy dollars minus sale dollars by ticker

Recent Form 4 Filings

Source rows from SEC EDGAR

Quant Portfolio Analysis

Simple view first; use the info markers for model details and calculations

Optimizer Proof Dashboard

Decision lift, validation, costs, risk budgets, and pending realized ledger checks

Quant Metric Explorer

Piotroski, Z-score, growth, valuation, return, volatility, and correlation fields

Quant Signal Context

Analyst posture, ranks, market sentiment, valuation, drawdown, and 2-year fundamentals

Top 20 Risk-Adjusted Names

Model favorites after adjusting for volatility, with institutional flow context

Portfolio II

Alpha Vol Target Momentum · weighted fundamentals, quality, growth, Sharpe, and relative Z-score

Current portfolio factor scores

Your holdings vs cross-sectional alpha score

Volatility regime

S&P 500 GARCH / HMM / MSAR

Financial turbulence

Equity turbulence HMM event probability

Alpha rankings

Cross-sectional factor score (robust z)

Factor model weights

Stability-adjusted IC engine

Top portfolio weights

Inverse-volatility blend of top 15 names

Commodity Earnings Cycle

Screen for temporary over-earning and under-earning in commodity-exposed companies

Over-Earning Watch

Peak-cycle profit risk

Under-Earning Watch

Trough-cycle recovery candidates

Macro Regime

Cross-asset signals that gate risk

Cross-Asset Market Map

5-day move by macro proxy

News & Catalysts

Linked sources by symbol and theme

Institutional Flow

Where off-exchange trading pressure looks strongest on names you watch

Flow Buy / Sell Table

Ranked by active flow metric

Watchlist — Most Bought

Highest TRF buy pressure on names you track

Watchlist — Most Sold

Low buy-ratio proxy on names you track

Market — Most Bought

Top volume pool, highest TRF buy pressure (DPI-ranked)

Market — Most Sold

Top volume pool, highest TRF sell pressure

Methodology

How to read these signals

Flow + Signals Portfolio

A simpler watchlist sleeve built from flow, manager, and catalyst signals

Return comparison (1M / 3M)

Flow sleeve return proxy

Buy vs sell flow impact

How TRF direction tilts sleeve weights and sub-sleeve returns

Top 25 Candidates

Approval-only study and sizing context

Market Overview

Model portfolios, backtests, and strategy risks

Allocation

Share counts at configured cash

6-Month Backtest Summary

Sorted by Sharpe ratio

Dynamic Backtest Performance

Growth of $1 vs S&P 500

Call Wall & SPY Vol Signals

Sector ETF distance to call walls, -1 sigma sell rail, and VIX1M/VIX3M stress

Strategy Risks

Dynamic trailing stops for highlighted portfolio

Notes

Methodology

Method & Data Health

Live backend method

Backend Pipeline

What runs before this JSON publishes

Current Source Status

Live inputs used for this run

How Thelma Scores Ideas

Signal-family score first; Research Alpha adds processed fundamentals, sector, scenarios, and Carver context

Risk & Privacy Boundary

Trade sizing and public sanitizer

Audit & Engine Health

Source freshness and publication safety

Source Freshness

Inputs checked for this run

Data Gaps

Items that reduce confidence

Earnings & 13F Calendars

Forward event windows

Earnings Calendar

Manual, SEC, and news-derived markers

13F Filing Calendar

Quarter deadlines and manager status

Research / 3-12m Alpha

Company-first scenarios from live prices, quant, 13F, and macro

Expected-Return Ranks

Risk-adjusted scenario return

Deep Dive

Select a ranked name

Model Portfolio

Approval-ready buy, sell, and hold targets
Target gates and sizing rules

Constraint Counts

First-class target gates

Decision Cards

Action, target, delta, risk, and confidence
Portfolio risk

3–12m Alpha Factor Budget

Portfolio Names & Sizing

Sorted by model % · risk-optimizer weights

Position Detail

Select a name

Factor Exposure

Current portfolio

Return attribution

Model book · not brokerage P&L

Covariance & Crowding

Shared factor risk

Pre-Trade Optimizer

Hedge and sizing checks

Portfolio Construction

Objective, constraints, leverage, and learning
Factor Constraints Enforced inside optimizer
Learning Loop Backlog, attribution, and hindsight

Factor Library

What each factor means

Pipeline & Failure Modes

How to think about the model

Open Model Providers

Active feeds and inactive adapters

Data Coverage

Direct feeds, estimates, and gaps

Trading Engine

Expected-return ranking with paper tracking
Expected-Return Ranks explained

Signal Validation

Did the score work historically?

Factor Contribution

Forward-return evidence by signal family

Expected-Return Calibration

Learned weights replace constants only after outcomes

Model Target Confidence

Confidence gated by historical proof

Factor Quality

IC, premium, turnover, redundancy

Orthogonality Matrix

Are factors independent enough?

Risk Span

Do we cover the main rewarded risks?

Expected-Return Ranks

3-12 month equity candidates

Paper Portfolio

Target changes tracked with proxy fills