Learn
Every number on Pulse comes from a fixed, documented calculation over market or SEC data. This page explains each term, the exact formula and thresholds Pulse uses, where the data comes from, and what it can’t tell you.
Pulse is a quantitative view of U.S.-listed stocks. It measures things. It does not predict them.
How many stocks are rising, falling, above their trend lines, or making new highs and lows, rolled into the Pulse Index.
For each stock: how active it is versus its own recent normal (Heartbeat), how tightly it is trading (Compression), and where it closed in the day’s range.
A fixed set of rule-based labels (Compressing, Tight Coil, Possible Breakout, Strong Close) that the Screener sorts by.
Annual figures companies report to the SEC, such as revenue, net income, cash flow and balance sheet, plus ratios Pulse calculates from them.
Every Pulse reading is stored by trading day, so you can see how a stock’s signals and the whole market have changed over time.
Pulse surfaces measurements and rule-based signals. It is not investment advice, and no signal predicts or guarantees a price move or its direction.
The Market Pulse page describes the whole market, not one stock. It is recalculated after every screener scan (about every 30 minutes during market hours) and stored once per trading day for history.
The set of stocks every market-wide percentage is calculated over.
The Screener and Market Pulse deliberately cover different sets of securities.
A 0–100 score for how broadly stocks are participating in the market right now. 50 is balanced.
Pulse Index = 0.25 × Breadth
+ 0.20 × % above 20-day SMA
+ 0.20 × % above 50-day SMA
+ 0.15 × New-High/Low balance
+ 0.20 × Strong-Move balance
Breadth = (advancing % − declining % + 100) ÷ 2
New-High/Low bal. = (new 20D high % − new 20D low % + 100) ÷ 2
Strong-Move bal. = (% up > 2% − % down > 2% + 100) ÷ 2
Result is clamped to 0–100 and rounded to one decimal. A component with no data counts as a neutral 50.
A plain-language band for the Pulse Index. It is fixed thresholds, not a judgment.
| Pulse Index | Label |
|---|---|
| below 20 | Broad Weakness |
| 20 to under 40 | Weak Participation |
| 40 to under 60 | Mixed Market |
| 60 to under 80 | Broad Participation |
| 80 and above | Broad Strength |
Pulse has no “Bullish / Neutral / Bearish” market classification. Those words appear only as a single stock’s daily direction.
Stocks in the breadth universe whose latest price is above (advancing) or below (declining) their previous close. Exactly unchanged stocks count as neither.
The share of stocks whose latest close is above their simple moving average of the last 20 (or 50) daily closes. A stock with fewer than 20 (or 50) days of history is left out of that percentage entirely, not guessed.
New high: today’s high is at or above the highest high of the last 20 sessions (today included). New low: today’s low is at or below the lowest low of those 20 sessions, and it isn’t also a new high.
The middle daily % change across the breadth universe, so half of stocks did better and half worse. Unlike an average, a few extreme movers can’t pull it around.
The share of stocks in six daily-change buckets: below −5%, −5 to −2%, −2 to 0% (unchanged included), 0 to 2%, 2 to 5%, and above 5%. Up > 2% / Down > 2% are the two tails the Pulse Index uses.
How many stocks carried each screener status on each day. Counts cover the whole scanned universe, not just the breadth universe.
Today’s Pulse Index, % above 20D/50D and setup counts, compared with the stored reading from about 30 days, 90 days and 1 year earlier.
Sentences filled in from the numbers above by fixed templates, with no AI involved. For example, the share of stocks up more than 2% reads “broad, expansive participation” at 25% or more, “moderate” at 10% or more, and “relatively narrow” below that.
A short written summary produced by a language model from the Market Pulse numbers. It is the only AI-generated content on Pulse.
Each stock is scored from its own daily bars. Every score compares the stock to its own recent behavior, so a quiet stock and a volatile one can both read “normal”. Scores are computed on the screener’s window, the last 130 calendar days of bars, and need at least 51 trading days of history.
Every stock gets exactly one status, checked in this order; the first match wins:
The stock is trading in an unusually narrow, overlapping range while activity stays low. It is Pulse’s strongest compression reading.
Heartbeat < 55 and Compression ≥ 80 (state “Tight”).Ranges, volatility and volume are contracting versus the stock’s 50-day baseline.
Heartbeat < 55 and Compression ≥ 60 (and below 80, which would be Tight Coil).A compressed stock just expanded sharply on heavy volume with a decisive close.
The stock finished the day near its high.
Close Strength ≥ 80, for a stock that isn’t Tight Coil, Compressing or a Possible Breakout.None of the four conditions above applies. This is most stocks on most days.
How active the stock is today compared with its own recent normal. It is not bullish or bearish.
Heartbeat = 100 × ( 0.35 × RelRange′ + 0.30 × RelVolume′
+ 0.20 × |Change %|′ + 0.15 × (1 − today’s overlap) )
RelRange′, RelVolume′ : 0.5× → 0, 3.0× → 1 (linear, clamped)
|Change %|′ : 0% → 0, 6% → 1
today’s overlap : see Range Overlap
How much the stock’s recent trading has contracted versus its 50-day baseline.
Compression = 100 × ( 0.30 × ATR contraction
+ 0.20 × range contraction
+ 0.25 × Range Overlap (5D)
+ 0.15 × volume reduction
+ 0.10 × persistence )
contraction of a ratio r (ATR14÷ATR50, avg range 14D÷50D,
avg volume 14D÷50D): r ≤ 0.6 → 1, r ≥ 1.3 → 0
persistence = consecutive days (max 10) with ATR14 ÷ ATR50 < 1, ÷ 10
Where the stock closed within the day’s high–low range.
Close Strength = (Close − Low) ÷ (High − Low) × 1000 = closed at the day’s low, 100 = at the high. Shown as N/A when High = Low.
The pill next to a company name, from Heartbeat (HB), Compression and yesterday’s Heartbeat: Tight HB < 55 & Comp ≥ 80 · Compressing HB < 55 & Comp ≥ 60 · High Energy HB ≥ 75 · Cooling HB 55–74 and down ≥ 8 from yesterday · Expanding HB 55–74 · Waking Up HB 30–54 and up ≥ 8 · Resting otherwise.
Only the day’s price change: above +0.15% is Bullish, below −0.15% is Bearish, anything in between is Neutral. It is independent of Heartbeat and is not a forecast.
Consecutive trading days (today included) the stock has been in the Compressing or Tight state. 0 means it isn’t compressed today.
Markers on the company price chart: Possible Breakout days, the first day of a Tight Coil, a Compression run, or a High Energy run, days with Relative Volume ≥ 1.5×, and Strong Close days. The table shows the median close-to-close return 5 and 20 trading days later. These are small samples, and past outcomes don’t predict future ones.
All windows count trading days (sessions), and each window includes today unless noted.
Today’s volume compared with the stock’s recent average.
Relative Volume = Today’s volume ÷ average volume of the last 14 sessions1.0× is normal, 2.0× is double, 0.5× is half.
How big today’s high–low spread is compared with the stock’s recent days.
Range % = (High − Low) ÷ previous close × 100 Relative Range = today’s Range % ÷ average Range % of the last 14 sessions
The typical size of a daily move in dollars, gaps included.
True Range = max( High − Low, |High − prev close|, |Low − prev close| )
ATR(14) = Wilder smoothing of True Range:
seed = mean of first 14 TRs,
then ATRₜ = (ATRₜ₋₁ × 13 + TRₜ) ÷ 14Recent volatility compared with longer-term volatility.
ATR vs 50 = ATR(14) ÷ ATR(50)Below 1× is quieter than its baseline (shown as “Volatility Below Baseline” in Quick Check); above 1× is more active. This ratio also drives Compression.
How much each day’s range sits inside the previous day’s range, averaged over 5 days.
today’s overlap = length of (today’s [Low, High] ∩ yesterday’s [Low, High])
÷ today’s (High − Low) → 0 to 1
Range Overlap 5D = average of today’s overlap over the last 5 sessions
It is divided by today’s range on purpose. An “inside day” (fully within yesterday’s range) scores 100%, and a day that engulfs yesterday scores low.SMA(N) = average of the last N daily closes. Market Pulse uses 20 and 50. The company page adds 200 when at least 200 bars exist. “Above 20D” means the latest close is above SMA(20).
The highest intraday high and lowest intraday low over the last 20, 50 or 252 sessions. 252 sessions is the 52-week high/low. “From 52W High” = Close ÷ 52W High − 1.
Plain averages of daily volume over the last 14 and 50 sessions. The 14-day average is the denominator of Relative Volume.
Close-to-close: 20D Return = Close ÷ Close 20 sessions ago − 1, and likewise for 50D and 1Y (252 sessions). The price chart’s “Period Return” runs from the first to the last close shown.
standard deviation of daily log returns × √252 over the selected chart range, shown as an annualized %. It needs at least 10 returns.
(Latest price − previous close) ÷ previous close. During market hours the latest price is the current, partial day’s bar.
A deterministic 0–100 summary of the measurements on this page, calculated in Python with no AI. It describes the current setup; it is not a buy, sell or hold recommendation.
Pulse Score = 0.25 × Technical Setup + 0.20 × Momentum / Trend
+ 0.10 × Volume + 0.20 × Financial Quality
+ 0.10 × Balance Sheet + 0.10 × Financial Trend
+ 0.05 × Market Environment
Each component is 0–100. Technical: screener status (an up-breakout scores high, a down-breakout low), Compression, Close Strength and position in the 20-day range. Momentum: price vs 20/50-day SMA, 20- and 50-day returns, new 20-day highs or lows, and the share of up days. Volume: relative volume confirmed by price action, liquidity, and the 20-day vs 50-day volume trend. Financial Quality: margins, positive cash flow and earnings, and return on assets. Balance Sheet: equity, leverage (on a separate scale for banks and insurers) and cash. Financial Trend: up to 4 consecutive fiscal years. Market: the Pulse Index and % above 20/50-day SMA.Price, daily change, state, direction, and the four headline readings: Heartbeat, Compression, Close Strength and Relative Volume. They are identical to the Screener’s values for the same session.
A factual snapshot with no combined score: price vs 20/50/200-day SMA, the Pulse readings, 20D/50D returns, and distance from the 52-week high and low. Each chip is marked as price or pulse.
Daily closes with 20/50/200-day SMAs, volume bars, and optional Pulse event markers. On long ranges each point can combine several sessions (last close, average daily volume). The statistics above the chart always use every day.
The supporting measurements (Relative Range, ATR, ATR vs 50, Range Overlap, Compression Duration), the Heartbeat history chart, and a table comparing today with the low, high and median of the last 30 and 90 sessions.
Raw OHLCV for the latest session and statistics computed directly from it: averages, rolling highs/lows, SMAs and returns.
Five stocks, each the top name in a different category (movers, Pulse setups, company size). Only liquid operating companies qualify: price ≥ $5, ≥ $25M average daily dollar volume, ETF venues excluded. Categories rotate daily. This is not a recommendation list.
Company financials are annual figures reported to the SEC in 10-K filings, in U.S. dollars. Pulse shows every stored fiscal year (up to 10) and never fills a gap. A missing value shows as —, never as zero.
Taken directly from each company’s XBRL-tagged 10-K data. Companies tag the same idea differently, so Pulse accepts a short, ordered list of tags per field. The first one reported for that year is used, and the tag is shown under the value.
| Field | Accepted SEC (us-gaap) tags, in order |
|---|---|
| Revenue | Revenues → RevenueFromContractWithCustomerExcludingAssessedTax → SalesRevenueNet |
| Net income | NetIncomeLoss → ProfitLoss − NetIncomeLossAttributableToNoncontrollingInterest → NetIncomeLossAvailableToCommonStockholdersBasic |
| Operating income | OperatingIncomeLoss |
| Total assets | Assets |
| Total liabilities | Liabilities |
| Stockholders’ equity | StockholdersEquity → StockholdersEquityIncludingPortionAttributableToNoncontrollingInterest |
| Cash & equivalents | CashAndCashEquivalentsAtCarryingValue → CashCashEquivalentsRestrictedCash… → CashAndDueFromBanks |
| Operating cash flow | NetCashProvidedByUsedInOperatingActivities → …ContinuingOperations |
| Capital expenditures | PaymentsToAcquirePropertyPlantAndEquipment → PaymentsToAcquireProductiveAssets → …AndIntangibleAssets → PaymentsForCapitalImprovements |
Banks and insurers usually report no single “revenue” or “operating income” line, so those fields are correctly blank for them.
“FY2025” means the annual period that ended in 2025.
Free Cash Flow = Operating Cash Flow − Capital ExpendituresCapEx is stored as the positive cash paid. FCF is shown only when both values exist for that year. It isn’t computed if CapEx has an unexpected negative sign.
Ratios and growth rates Pulse derives from the reported values. Each row shows its inputs, its formula, and a factual reading.
| Metric | Formula | Shown as “n/m” or “—” when |
|---|---|---|
| Revenue / Net Income / Operating Income / Asset / Equity YoY | (This FY − prior FY) ÷ |prior FY| | the exact prior year isn’t stored, or the prior value is zero or negative |
| Revenue 3Y CAGR | (Revenue FY ÷ Revenue FY−3)1/3 − 1 | FY−3 missing, or either value ≤ 0 |
| Operating / Net / Operating-cash-flow / FCF margin | value ÷ Revenue | revenue missing or ≤ 0 |
| Return on Assets | Net income ÷ year-end total assets | assets ≤ 0 |
| Return on Equity | Net income ÷ year-end equity | equity ≤ 0 (common after large buybacks) |
| Liabilities / Assets, Equity / Assets, Cash / Assets | value ÷ total assets | assets ≤ 0 |
| Liabilities / Equity | total liabilities ÷ equity | equity ≤ 0 |
| FCF YoY | (FCF − prior FCF) ÷ prior FCF | prior FCF ≤ 0 |
“Essentially flat” in a reading means a change smaller than ±1%. Returns use year-end balances, not averages.
Daily open/high/low/close/volume bars for every stock, from Alpaca’s SIP feed, the consolidated tape across all U.S. exchanges. On Pulse’s data plan, recent data is delayed by about 15 minutes. The stock universe (symbol, name, exchange, tradability) also comes from Alpaca.
IEX is a single exchange, not the consolidated market, so its volumes and prices can differ from the full tape. Pulse does not use IEX data for any price, volume or signal. It is mentioned here because the separate State Ledger project stores IEX bars, and Pulse deliberately doesn’t read them.
The SEC’s public database of company filings. Company profiles (name, industry code, headquarters) and annual financial facts come from its submissions and Company Facts APIs.
Pulse doesn’t download SEC data itself. A companion project, State Ledger, collects and cleans it; Pulse reads it read-only, matches tickers to companies through the SEC’s own ticker list, and caches results for an hour. If that source is unavailable, the rest of the page still works.
Pulse’s own store of daily price bars, the latest screener reading per stock, intraday Market Pulse snapshots, and the daily history of every stock’s Pulse readings and of the market-wide index.
The AI Market Brief is written by an OpenAI language model from Pulse’s computed numbers. Nothing else on Pulse is AI-generated.
Daily OHLCV bars: at least about 5 years for the stock universe, and up to about 20 years for stocks that have been looked up individually. This drives the price chart, moving averages and 52-week figures.
Each stock’s Pulse reading per trading day (Heartbeat, Compression, Close Strength, status, Relative Volume, ATR ratios and more), recomputed with the same screener method on the bars available that day, with no look-ahead. Stored from early 2022, and a new session is added each trading day after the close (about 4:30 pm ET).
One end-of-day Market Pulse reading per trading day (charts: 30D to ALL), plus intraday snapshots taken after each screener scan (about every 30 minutes during market hours).
A price bar is what happened. A Pulse observation is Pulse’s measurement of that day. Readings are tagged with a methodology version so results from different formulas are never mixed.
Recent IPOs and spin-offs have fewer bars. A Pulse reading needs 51 trading days, the 200-day SMA needs 200, and 52-week figures need 252. SEC history depends on how many annual 10-K filings the company has.
Screener and Market Pulse: about every 30 minutes during market hours. Daily history: once after each close. SEC financials: when State Ledger re-imports filings (annual data changes a few times a year).
Pulse is an analytics tool, not investment advice. Nothing on Pulse is a recommendation to buy or sell any security.