Research Methodology

Methodology maintained by Mr. Semi

This page explains what the site's indicators mean, how often they update and where judgment or data limitations can affect the result.

Research framework

The model organizes public information into semiconductor value-chain nodes. Evidence is classified by the node it directly affects, its direction, timing and source. Node observations are then compared across connected upstream and downstream stages so that a demand or supply change can be traced through the chain.

Update schedule

• Core node judgments: recalculated once every 24 hours.
• Price and candlestick data: refreshed every 10 minutes when the relevant market data is available.
• Flash news: scanned hourly. A scan can complete without publishing a new item if no relevant, sufficiently distinct information is found.
• Displayed timestamps: shown to the minute so readers can distinguish the latest scan from the latest published item.

Inputs and processing

Inputs include public news, company and regulatory disclosures, market prices, revenue or operating indicators, and macroeconomic series. Software normalizes the data. AI-assisted classification and summarization help map text to nodes and remove duplication. An AI-generated summary can omit context or make a classification error, so linked primary evidence should be reviewed before relying on it.

What the scores mean

Industry-activity scores are comparative research indicators, not forecasts of a security's return. Market-expectation measures describe observable pricing or expectation signals and can diverge from fundamentals. A high score does not mean that an asset is inexpensive, and a low score does not by itself mean that it should be sold.

Limitations and corrections

Public data can be delayed, revised, unavailable or inconsistent across regions. A lack of new evidence is not evidence that conditions are unchanged. Relationships in the value-chain map simplify a real system with feedback loops, inventories and long lead times. Material errors should be corrected in the underlying record and reflected in the next published analysis.

Investment disclaimer

The research is general information for education and investigation. It is not investment advice, a solicitation, or a recommendation to buy or sell any security. Users remain responsible for independent verification and investment decisions.