cyb50-quant/research/dragon/v2/openspec/changes/followthrough-lean-profit-loop/proposal.md

Why

Phase 1 already proved the shadow pending scaffold can be added without drifting RC1. That did not answer the only question that actually matters for this strategy now: after a false veto, which glued weak-rebound paths deserve a fast trend reentry, and is the bottleneck entry timing or the post-entry exit treatment?

The next step should stay lean. We do not need a new platform. We need a repeatable review pack that turns this question into concrete subtype-level evidence.

What Changes

• Add a dedicated review script for glued followthrough profit-loop analysis.
• Reconstruct blocked-candidate observations directly from strategy logic.
• Compare probe reentry trades across:
• subtype lane
• same-close vs next-open execution
• post-entry upside and post-exit followthrough
• Publish CSV + Markdown outputs.
• Keep RC1 default strategy behavior unchanged.

Capabilities

New Capabilities

• followthrough-profit-loop-research: reproducible blocked-candidate and reentry-trade review focused on false-veto recovery.

Unchanged Capabilities

• glued-followthrough-shadow-pending: existing Phase 1 pending scaffold remains intact.
• shadow-branch-observation: RC1 default governance path remains unchanged.

Impact

• Affected code: new review script, new tests, new OpenSpec change pack.
• Runtime impact: offline research/reporting only.
• Strategy impact: none to RC1 default branch.