Block Trade Detector

Block trades appear on consolidated tape/time & sales. Free: Yahoo Finance time & sales (limited), broker platforms. Paid: Bloomberg, Refinitiv, Trade Alert provide better analytics. TSX provides marketplace data.

Common thresholds are 10,000+ shares or $100,000+ value. Some use higher thresholds for large-cap stocks (25,000+ shares). The key is that it's too large for typical retail trading.

No. Many blocks are rebalancing, crosses, or non-directional. Focus on: larger blocks, clear direction (at ask/bid), clusters of blocks, and context that makes sense. Use as confirmation, not sole signal.

A cross is a pre-arranged trade where a broker matches a buyer and seller (often at midpoint). It's less directional because it's negotiated, not one side urgently buying/selling in the market.

In Canada, trades are reported in real-time to the consolidated tape. You see blocks almost immediately after execution. Some negotiated/upstairs trades may have slight delays.

Compare to ADV: Block / ADV = significance %. Above 1% is notable, above 5% is very significant. Also check premium/discount to VWAP - larger premiums indicate more urgency.

A cluster is multiple blocks in the same direction within a short window (30-120 min). Clusters matter because they suggest sustained institutional program rather than one-off trade. More predictive than single blocks.

At ask price = buying (buyer paid up). At bid = selling (seller accepted lower). Also check: uptick/downtick, above/below VWAP. Combine signals for higher confidence classification.

Yes, definitely. Block data shows stock positioning; options show derivatives positioning. When both align (buy blocks + call buying), signals are stronger. Divergence warrants investigation.

Dark pool blocks are reported but the venue is hidden until execution. Same analysis applies - size, direction, premium. Compare lit vs dark block activity; heavy dark may indicate larger programs.

Features: net block direction, size-weighted premium, cluster score, persistence. Target: forward returns. Model: regression or ML. Rank stocks by block score; construct long-short if desired. Walk-forward validate.

Signs: consistent block sizes (algorithmic), regular time intervals, same direction over days, similar price levels. When you identify a program, estimate its total size and monitor for completion.

Opening/closing blocks: often passive (rebalancing). Give more weight to midday blocks. Also track timing vs events - pre-announcement blocks may be informed (or coincidence). Build time-of-day factors.

Map block activity periods to subsequent 13F filings. Persistent buy blocks in Q1 should correspond to increased ownership in Q1 13F (reported in May). This validates whether your block signals predict actual positioning.

Build composite signal: Block score + Options flow score + Dark pool score. Weight by predictive power (backtest). Require multiple confirmations for high-conviction trades. Track which combinations work best.