What it means
Research analysts publish ratings to summarize their views on a stock. Common labels include buy, hold, and sell, but firms may use terms such as outperform, neutral, or underperform. A label needs context: the analyst’s stated benchmark, time horizon, and rating definitions determine what it means. For example, “outperform” may describe a stock’s expected performance relative to an index rather than an expectation that its price will rise.
A rating is an opinion, not an order instruction. It can change when an analyst revises assumptions or new information becomes available. Different analysts can reasonably reach different conclusions about the same company.
How it is calculated
An individual rating is assigned through an analyst’s research process; there is no universal formula that converts company data into a buy, hold, or sell label. A consensus rating summarizes multiple analysts’ ratings. One possible summary maps labels to numbers and takes an average:
Consensus score = (sum of rating scores) / N
- Consensus score is the arithmetic mean of the included numerical scores; it has no monetary unit.
- Rating score is the number assigned to each label under the chosen scale—for example, buy = 1, hold = 2, sell = 3. The direction depends on the scale.
- N is the number of included analyst ratings.
This calculation is only a summary convention. Data providers may use different labels, mappings, weighting methods, or rules for including ratings, so their consensus figures may differ. A mean also hides disagreement: it does not show whether every analyst chose hold or whether analysts were split between buy and sell.
Worked example
Suppose, hypothetically, five analysts rate the same stock. Three assign buy, one assigns hold, and one assigns sell. Use the example scale buy = 1, hold = 2, and sell = 3.
- Convert the five labels to scores: 1, 1, 1, 2, and 3.
- Add them:
1 + 1 + 1 + 2 + 3 = 8. - Divide by five ratings:
8 / 5 = 1.6.
The consensus score is 1.6 on this three-point scale, not a 1.6% expected return or a probability. Calling it “buy” would require a stated rule for translating averages back into labels. The individual ratings remain important: one of the five analysts assigned sell.
Why it matters before you trade
A rating can highlight an analyst’s interpretation of earnings, valuation, or company risks. Before using one in a pre-trade check, look at when it was published, whether it changed, what the label means at that firm, and whether the analyst explains the underlying assumptions. A consensus figure may be stale if its component ratings were issued at different times.
Keep a rating separate from a price target: the rating is a categorical opinion, while a price target is an analyst’s estimated future price at a specified horizon. Neither tells you what your order will cost to execute. Bid-ask spread, liquidity, and market impact are separate considerations, as are position size and concentration within a basket.
How PreTrAIde uses it
PreTrAIde includes analyst ratings in the data used to enrich each stock. Its AI agents reason about each stock and the whole basket using quant scores, news, and analyst ratings. In a Pre-trade analysis, the user enters stocks, a buy or sell side, quantities, and execution conditions before placing any order. PreTrAIde then provides GO, NO-GO, or MODIFY verdicts, a model-estimated success probability, a risk assessment, a suggested execution strategy, and plain-language reasoning.
A rating is one input, not a standalone verdict. PreTrAIde’s verdicts are decision support, not recommendations; users place any trades themselves with their own broker. The success probability is not a guarantee or a forecast of returns.
Common mistakes
- Treating “buy” as an instruction or as a guaranteed positive return.
- Assuming different firms’ rating labels, benchmarks, and time horizons mean the same thing.
- Mistaking a consensus score for a percentage, or overlooking disagreement behind an average.
- Reading an old rating without checking whether later earnings or news changed its context.
- Treating a favorable rating as evidence that a particular order will be inexpensive to execute.