The headline number versus the detail underneath

[Shadow Trader]

An algorithm reads "payrolls 250k versus 180k expected" and trades within milliseconds. The market spikes. A human reading the same release sees the unemployment rate rose, wage growth missed, and prior months were revised down. The second move goes the other way.

This pattern, an initial reaction followed by reversal, shows up regularly. Headline-reading algorithms act on the first number. Slower participants, and some faster ones with more sophisticated parsing, act on the broader picture.

For a scalper, the implication is that the first move and the second move often have different causes. Chasing the first is a bet on the headline algos. Trading the second is a bet on the detail.

It also explains why prices can whip in both directions within a minute of a release, long before the "real" trend establishes itself.

Different data has different structure. Employment reports have many components. Inflation data has headline and core. Retail sales have a control group. Learning which component the market watches most in the current environment is part of the job.

How the headline algorithm works matters. Many trading systems parse the number from a news feed and compare it with the consensus. The decision takes microseconds. The details, such as revisions, components, and wage data, take longer to parse, so they arrive in the second wave.

Humans enter the picture even later. A discretionary desk may take thirty seconds to a few minutes to form a view and act. Their flows often drive the larger and more lasting move.

For a short-term trader, that suggests a timing structure. The first print is a reaction by machines. The next few minutes are a transition. The move after that has more information behind it.

Not every release follows this, but the pattern is common enough to be worth testing on the specific releases traded most. Marking the three phases on a chart for a dozen events often shows a consistent shape.

Which release has shown the largest gap between the first reaction and the one that lasted?