ElementWatch

Methodology

← Back to app
Methodology

ElementWatch methodology

ElementWatch isn't a chart library — it's a decision-grade ranking. Every signal carries a strength (Watch / Medium / High / Critical), a confidence band, an operator action, and a watch-next set. Here's how those are built.

What we track

  • Signal strength — weighted from 7-day move magnitude + supply tightness + policy events
  • Confidence — High / Medium / Low based on source reliability and freshness
  • Operator action — AI-generated on top of structured context (Pro tier)
  • Watch-next — 3–5 monitoring items the operator should track this week
  • Tracking mode — "price" for benchmarks with spot data; "signal" for policy / allocation-driven materials
  • 4-phase refresh — Phase 2 writes live prices immediately so AI-timeout never wipes the price layer

Who this is for

Anyone evaluating ElementWatch for governance, audit, or internal decision-grade use.

Decisions ElementWatch supports

  • Validating that an ElementWatch signal is fit for your decision
  • Mapping our schema to your internal risk register
  • Building reproducible procurement-risk reporting

Primary sources