What it means
Market sentiment describes the tone of views about a stock or the wider market. Sentiment analysis tries to measure that tone from sources such as news reports, investor commentary, or surveys. It asks what people appear to think—not whether their views are correct or what the stock will do next.
The result depends on the source, the time window, and what is being classified. A news headline about a company’s earnings may express a different tone from investor comments about its share price. Sentiment can also change quickly as new information arrives.
How it is calculated
There is no single standard market sentiment formula. One simple approach classifies each relevant item as positive, negative, or neutral, then compares the positive and negative counts:
Net sentiment = (P − N) / (P + N + U)
Pis the number of items classified as positive.Nis the number classified as negative.Uis the number classified as neutral.
The result is a unitless score from −1 to +1, provided at least one item is counted. Zero means the positive and negative counts are equal; it does not necessarily mean that investors are indifferent. If no items qualify, the score is undefined, not zero.
Other methods weight items by recency, source, or strength of language. Those choices change the result, so a sentiment score is meaningful only alongside its method and sample.
Worked example
Suppose a hypothetical review finds 20 relevant news items about one stock during a defined period. A classifier labels 12 positive, 5 negative, and 3 neutral.
- Count the items:
P = 12,N = 5, andU = 3. - Subtract negative from positive:
12 − 5 = 7. - Divide by all classified items:
7 / (12 + 5 + 3) = 0.35.
The net sentiment score is +0.35. That means positive items outnumber negative items in this sample. It does not mean the stock has a 35% expected return, a 35% chance of rising, or a 35% chance of any other trading outcome. A different selection of sources or classification rules could produce a different score.
Why it matters before you trade
Sentiment can help identify the narrative surrounding a proposed stock trade. For example, a recent earnings announcement may explain why coverage has become more positive or negative. Reading the underlying items helps distinguish company-specific information from a broad market mood.
Sentiment is only one part of a pre-trade check. News may already be reflected in prices, and a strongly worded article says little about the cost of executing an order. Liquidity, bid-ask spread, potential market impact, volatility, and basket concentration address different questions. A headline count should not substitute for those checks.
How PreTrAIde uses it
PreTrAIde performs pre-trade analysis on a user-built stock basket before any order is placed. Its data enrichment includes recent news per ticker and analyst ratings. Its AI agents reason about each stock and the basket using that information alongside quantitative scores; the product facts do not specify a standalone sentiment score or a positive-versus-negative headline-count formula.
The analysis returns a GO, NO-GO, or MODIFY verdict for the basket and each line, with a model-estimated success probability, risk assessment, suggested execution strategy, and plain-language reasoning. Those verdicts are decision support, not recommendations. PreTrAIde does not place orders; users trade through their own broker.
Common mistakes
- Treating tone as a price forecast. Favorable coverage does not establish the direction or size of a future move.
- Ignoring the denominator. A score based on five items is not supported by the same amount of text as one based on 500.
- Confusing analyst ratings with all investor opinion. Ratings represent analysts’ assessments, not a complete measure of market sentiment.
- Overlooking selection and timing. Repeated reports of one event can dominate a sample, while older items may no longer reflect current discussion.
- Assuming a classifier understands every context. Negation, sarcasm, and mixed news can lead to misleading labels.