Skip to main content

    Stock scoring explained: how stock scores are built, what they measure and how to read them

    Stock scoring turns several measurements, such as momentum, trend, volatility, quality and valuation, into one number by ranking stocks against each other. Here is how scores are built, using a published methodology as the example.

    Peakwise Research · · Updated · 4 min read

    Contents

    Stock scoring condenses several measurements of a stock, such as how it has moved, how much it swings, how profitable it is and how it is valued, into a single number, usually by ranking each stock against the others on a list and averaging the ranks with weights. A score is a fast way to see where a stock stands. It is also easy to overrate, because one number hides which factors drove it and what was left out. This guide takes a scoring system apart step by step, using the published Peakwise methodology as the example.

    Step 1: Choose the factors

    Every scoring system starts with a list of what it measures:

    Factor familyWhat it captures
    MomentumPrice change over recent months
    TrendPrice compared with longer-term averages
    RiskVolatility and drawdown
    QualityProfitability and balance sheet strength
    ValuationWhat you pay for earnings, sales or cash flow

    The choice shapes the result. A score heavy on momentum favours recent winners, one heavy on valuation favours cheap stocks. No score is neutral.

    Step 2: Put the factors on one scale

    Raw numbers cannot be added: a P/E of 25 and volatility of 30 percent have different units. So most systems turn each factor into a rank or percentile within a list, from 0 (worst) to 100 (best).

    That makes the score relative. It tells you how a stock compares with the others on the list, not whether it is good in absolute terms. If the whole market falls, the best-ranked stock can still score high.

    Step 3: Weight and combine

    The ranks are averaged with weights. The published Peakwise weighting, for a fixed list of 50 large S&P 500 companies:

    PartWeightInput
    Momentum30%Price change over 3, 6 and 12 months
    Trend20%Price versus its 200 day average
    Volatility15%How much the price swings, calmer ranks higher
    Drawdown15%Largest fall from a peak over 12 months
    Quality10%Net profit margin and debt relative to equity
    Valuation10%Earnings yield and free cash flow yield

    When ratios are missing, quality and valuation drop out and the rest are re-weighted. The full detail is on the methodology page. With 80 percent of the weight on price behaviour, this score mainly describes how a stock has traded, with a smaller check on fundamentals. Read it alongside a fundamental view, for example from the AI stock analysis.

    Step 4: Label the ranges carefully

    Labels make tables easier to scan. Peakwise uses "Very strong profile" for 80 to 100, "Strong profile" for 65 to 79, "Mixed profile" for 50 to 64, "Weak profile" for 35 to 49 and "Very weak profile" below 35. The labels describe the range and recommend nothing.

    Step 5: Update on a schedule

    Price-based scores need fresh prices. The Peakwise score updates once per trading day after the close from end of day prices and shows the date of the data. Fundamental inputs change only when companies report.

    How to read a score well

    • Read the parts. Two stocks scoring 70 can be very different underneath: one strong on momentum and weak on valuation, the other the reverse.
    • Watch the direction. A slide from 75 to 55 over a month means the stock weakened relative to the list. Find out why.
    • Remember what is missing. Lawsuits, product launches and management changes appear only once they show up in prices or reports.
    • Do not confuse past and future. A score describes data that already happened.

    Score vs analyst rating

    Rule-based scoreAnalyst rating
    SourceFormula on dataOpinion of a research firm
    ConsistencySame inputs, same resultVaries by analyst
    Forward-lookingNoYes, includes forecasts
    Conflicts of interestNone by designPossible

    When a score helps most

    • Building a shortlist from a large list, see AI stock screener.
    • Checking whether a stock you hold has weakened against peers.
    • Comparing similar companies quickly before a deeper look with how to compare two stocks.

    Each of the 50 scored companies has its own page in the stock section, for example Microsoft or Visa, with the score and its parts, and the AI stock finder lets you sort the whole list.

    This guide is for education and is not investment advice.

    Frequently asked questions

    Most scores measure several factors, such as momentum, volatility, profitability or valuation, convert each into a comparable scale, often a rank within a list, and combine them with weights. A good methodology states the inputs and weights.

    It depends on the scale. In a relative score from 0 to 100, a high number means the stock ranks above most others on the list for those inputs. It does not mean the stock will rise.

    Price-based scores usually update daily after the close. Inputs from financial statements change when companies report.

    No. An analyst rating is a published opinion. A rule-based score is computed from data with a fixed formula and expresses no opinion.

    Compare stockswith the AI stock score

    The score compares price history and ratios across a fixed list of 50 large S&P 500 stocks in one number from 0 to 100.

    Free for 7 days. Not investment advice.