← Trading Systems
Sierra Chart · ACSIL · C++ DLLs

Sierra Chart DLL Studies — LTVI, Logo & Algo

Custom C++ add-ons for Sierra Chart: orderflow and volume visualization (LTVI), on-chart branding (logo), and ~35 indicators and strategies organized into suite DLLs — from auto-trading around order-book imbalance and key levels to session tools, Kalman smoothing and trade-review utilities — including a licensed member suite shipped to a trading community's paying subscribers.

C++ ACSIL Volume-at-Price Time & Sales GDI+
Sierra logo indicator preview
⚡
Built with agentic AI. Development on this project runs through multi-agent Claude Code workflows — developer → independent-reviewer cycles with evidence gates, batched into unattended runs that execute overnight and resume from committed ledgers. How the agent system works →

What the studies do

A Sierra Chart study is a custom add-on that draws on the chart or runs trading logic. These studies read the live trade stream (Time & Sales) and volume-at-price data to make orderflow visible — and, in the strategy variants, tradeable.

Orderflow & volume visualization (LTVI)

  • Big single prints drawn as numbers or bubbles at the price they hit, so large trades stand out at a glance.
  • Bid/ask imbalance markers per price level or per bar, sized by how extreme the imbalance is.
  • Absorption levels — heavy volume with little price movement — drawn as horizontal lines with expiry and alerts.
  • A GDI+ study renders an embedded image (logo/watermark) scaled onto the chart — no external file needed.

Strategies & indicator families

  • Auto-trading strategies around order blocks, trapped volume and volume imbalance, with per-session (NY/Asia) sensitivity and attached stop/target orders.
  • Structure family ("Unbroken"): swing highs/lows, break-of-structure and retest signals, with filter variants (trend, absorption, RSI divergence).
  • Session & smoothing tools: session high/low and OHLC with correct DST handling; four Kalman-filter studies for delta, volatility, fair value and trend.
  • Workflow utilities: cross-chart signal overlays, a trade-setup jumper, cumulative P&L, end-of-day slippage analysis, signal export/import between studies.

Suite architecture — shipping DLLs to non-developers

The studies are packaged so users never need a compiler, and the developer never touches a member's machine.

One DLL = one suite

  • Each suite builds to a single DLL — one title in Sierra's Add Custom Study dialog, every member study in its dropdown, each in its own namespace.
  • Two-flavor build: a test DLL exposes full tuning inputs plus a "Bake" action that writes settings into a header; the production DLL is a unity shell the study owner compiles inside Sierra Chart itself, with no Visual Studio. Separate license groups keep the dev flavor from ever leaking to members.
  • A hot-reload dev rig releases and re-allows the DLL over Sierra's UDP control interface on every build.

Support built into the product

  • A "Write Diagnostic Report" input dumps one self-contained file a member can email — whichever pipeline stage first reads zero names the cause.
  • Late-added inputs are created at runtime, so a plain DLL swap delivers new features without members rebuilding chartbooks.
  • Ports of platform built-ins are verified bar-for-bar against the original (max difference 0.000000) before replacing them; the order-block replay path was made ~12× faster by keying levels by tick.

Newer study families

Licensed member suite (7 studies)

  • Block Chart — a custom bar builder (inverse flex-renko) that is "correct by construction": bricks complete only from committed highs/lows, so no phantom bars and every close is a traded price.
  • Iceberg — compares traded volume at a price against the next bar's resting depth and bubbles the multiplier.
  • OB Max, Premium, Momentum, Trend — order blocks with the diagnostic-report system, reversal-level retest counting, RSI-extreme coloring, and the parity-verified trend port.

The order-flow research family

  • Ten research studies share one statistical engine that scores order-flow behaviour against context-conditioned baselines instead of global averages — the detection logic itself stays private.
  • The engine guards against the classic false-positive modes of adaptive thresholds (quiet-hours artifacts, warmup instability).
  • Candidate edges were tested against matched random controls and refuted where they failed — the same evidence standard as the quant pipeline.