TradeSync: making the preview more important than the execute button

Context#

TradeSync turns funding, open interest, cumulative volume delta, price structure and technical indicators into inspectable directional scores. The difficult part is not producing a long or short label. It is preserving enough evidence and state around that label for a person to understand what the system believes, what could invalidate it and whether an execution path is allowed to proceed.

One scoring and risk pipeline#

The public Phase 3D record documents a shared tradesync_core library for scoring, risk and symbol normalisation. Services import that canonical logic, while an offline replay runner sends historical JSONL events through the same calculate-score, enhanced-score and RiskGuardian sequence. That avoids the dangerous version of a backtest: one that quietly evaluates different rules from the live-facing services.

Preview before authority#

The execution contract gives preview its own first-class endpoint. A preview returns a proposed plan, an allow-or-block risk verdict, machine-readable block codes and execution-risk evidence such as spread, estimated impact and market depth. The point is to make the decision inspectable before any executor is considered—not to use a confirmation dialog as decoration around an opaque order request.

A truthful cockpit is part of the safety system#

A March 2026 public audit removed hard-coded flow metrics and fake event lines, replaced a static dry-run badge with backend-derived state, separated DEMO from PAPER and LIVE, and disabled Copilot and source-upload surfaces that did not exist yet. Those fixes matter because a trading dashboard that cannot distinguish mock positions from connected paper operation is not merely unfinished; it gives the operator the wrong mental model.

Current boundary#

TradeSync remains an engineering and research system. Its public README describes the cockpit as mock/dry-run, states that no real venue keys are configured, and marks wallet authority and later AI surfaces as future work. The repository contains execution-shaped contracts and services, but this site does not convert those shapes into a claim of live trade execution, autonomous authority or financial performance.

Next proof#

The next credible milestone is not a louder autonomy claim. It is a reproducible paper evaluation where the same versioned scoring and risk rules produce a reviewable preview, a blocked-or-allowed result and an outcome record—without silently crossing into live execution.