Formula Analysis

Open of every bar. Also Open.

High of every bar. Also High.

Low of every bar. Also Low.

Close of every bar. Also Close.

Volume of every bar. Also Volume.

Typical price, (H + L + C) / 3.

Number of bars loaded for this symbol.

0, 1, 2 … for each bar.

Bar time as Unix seconds (UTC).

Compact date number: (year − 1900)·10000 + month·100 + day. 1250131 is 31 Jan 2025.

Filter = DateNum().ge(1250101);

Calendar year of each bar.

Month 1–12.

Day of month 1–31.

0 = Sunday … 6 = Saturday.

1–366.

Ticker of the symbol being processed.

Company or asset name.

Sector, where known.

Industry, where known.

Listing market code: US, IDX, TSE, CRYPTO …

Currency the prices are quoted in.

Long entry signal. Read at the bar's close and filled on the next bar (trade delay 1) at the open by default.

Buy = Cross(C, MA(C, 50));

Long exit signal.

Sell = Cross(MA(C, 50), C);

Short entry signal. Needs direction 'short' or 'both' in the settings, or SetOption("AllowShort", true).

Short exit signal.

Price to buy at instead of the open/close. Clamped into the bar's high–low range.

BuyPrice = HHV(H, 20).ref(-1); // breakout stop order

Price to sell at.

Price to short at.

Price to cover at.

Which bars appear in an Exploration. Without Filter only the last bar of every symbol is listed.

Filter = RSI(14).lt(30);

Caption shown above the chart.

Draw a line on the chart. style: line, histogram, dots, area, step. pane: "price" overlays the candles, any other name opens a separate pane, "auto" decides by scale.

Plot(MA(C, 20), "MA 20", { color: "orange" });
Plot(RSI(14), "RSI", { pane: "RSI" });

Mark bars on the chart. shape: arrowUp, arrowDown, circle, square. position: aboveBar, belowBar.

PlotShapes(GapUp(), "arrowUp", "Gap", { color: "teal", position: "belowBar" });

Add a column to the Exploration table. format: 1.2 (two decimals), "pct" (0.12 → 12%), "int", "money", "date". color: a colour name, a series (positive green, negative red) or ColorIf().

AddColumn(ROC(C, 20), "20-day %", { format: 1.1, color: ROC(C, 20) });

Text column. Pass a string, an array with one entry per bar, or a function of the bar index.

AddTextColumn(SectorName(), "Sector");

Default sort of the Exploration table.

SetSort("Score");

Per-bar colour list for AddColumn / AddTextColumn.

Write to the formula log (first 200 lines). Same as _TRACE.

A number the user can change with a slider without editing the code.

var n = Param("Period", 20, 5, 200, 1);

An on/off switch.

var useVol = ParamToggle("Volume filter", true);

A choice from a list. Returns the chosen text.

var avg = ParamList("Average", "MA|EMA|WMA", "EMA");

A parameter the optimiser and walk-forward test vary. In every other run it takes its default (or the value set in the parameter panel).

var fast = Optimize("Fast", 20, 5, 50, 5);

Simple moving average.

Exponential moving average, seeded with the simple average of the first period.

Linearly weighted moving average.

Double exponential moving average.

Triple exponential moving average.

Hull moving average.

Wilder's smoothing (used inside RSI and ATR).

Adaptive average with a per-bar smoothing factor between 0 and 1.

Volume-weighted moving average of the close.

Relative Strength Index of the close.

RSI of any series.

EMA(fast) − EMA(slow) of the close.

Signal line: EMA of MACD.

MACD − Signal.

Slow stochastic %K.

Stochastic %D.

Williams %R, 0 to −100.

Commodity Channel Index.

Money Flow Index.

Percent change over the period, in percent (5 means +5%).

series − series n bars ago.

Aroon up.

Aroon down.

Average True Range.

ATR as a percent of the close.

True range of every bar.

Average Directional Index.

+DI.

−DI.

Upper Bollinger band.

Lower Bollinger band.

EMA + mult·ATR.

EMA − mult·ATR.

Parabolic SAR.

Supertrend line. .dir is +1 in an uptrend, −1 in a downtrend.

var st = Supertrend(10, 3);
Buy = Cross(st.dir, 0);

Low above the previous high.

High below the previous low.

Inside bar.

Outside bar.

On-balance volume.

Chaikin Money Flow.

Value n bars ago. Offsets must be negative or zero — a positive offset would read the future and is refused.

var yesterday = Ref(C, -1);

1 on the bar where a moves from at or below b to above b.

Bar-by-bar choice.

1 where every argument is true.

1 where any argument is true.

1 where a is false.

1 from a bar where a is true until b becomes true.

Removes repeated a signals until b appears. Buy = ExRem(Buy, Sell) keeps the first buy of each run.

1 for n bars after a was true.

Bars since the condition was last true.

Value of series at the nth most recent bar where condition was true.

Highest value over the period, including this bar.

Lowest value over the period.

Bars since the highest value of the period.

Bars since the lowest value of the period.

Highest value since the first bar.

Lowest value since the first bar.

Highest value since condition was last true.

Lowest value since condition was last true.

How many of the last n bars were true.

Rolling sum.

Running total.

Population standard deviation.

(series − MA) / StDev.

Rolling Pearson correlation.

Slope of the least-squares line per bar.

End value of the least-squares line.

Start value of the least-squares line.

Percent of the previous n values below today's (period up to 500).

Absolute value.

Larger of two values per bar.

Smaller of two values per bar.

Square root.

Natural log.

Base-10 log.

e to the power x.

x to the power y.

Round per bar.

Round down.

Round up.

−1, 0 or 1.

Replace empty values (warm-up bars, missing data).

1 where the value is empty.

a + b. Same as a.add(b).

a − b.

a × b.

a ÷ b, empty where b is zero.

Turn a number or a plain array (one value per bar) into a series.

Fiscal-year figure as a step series. Each year's numbers start on the filing date (or 90 days after the year end), so a backtest never sees a report before it was public. Keys: revenue, gross_profit, operating_income, net_income, eps, equity, assets, liabilities, cash, shares, cfo, capex, fcf, dividends, gross_margin, operating_margin, net_margin, roe, roa, debt_to_equity, current_ratio, revenue_growth, eps_growth.

Filter = F("roe").gt(0.15).and(F("revenue_growth").gt(0.1));

Price ÷ split-adjusted EPS. Empty when earnings are negative or the report is in another currency without a stored rate.

Price ÷ book value per share.

Price ÷ revenue per share.

EPS ÷ price.

Free cash flow per share ÷ price.

Dividends paid per share ÷ price.

Close × split-adjusted shares.

Split-adjusted earnings per share.

Close of the symbol's market index (S&P 500 for US, IHSG for IDX, Bitcoin for crypto …) aligned to its bars.

Ticker of that index.

Close ÷ benchmark, rebased to 100 at the first bar.

Rolling beta of daily returns against the benchmark.

Run part of the formula on hourly, daily, weekly, monthly, quarterly or yearly bars built from the chart's bars (on 5-minute or hourly charts, "daily" gives the previous day's values), then bring the result back to every daily bar. Each day sees only the last finished period, so there is no look-ahead: a weekly value appears on the first bar of the next week. Plot, AddColumn and Rank must stay outside.

var weeklyUp = TimeFrame("weekly", function () { return C.gt(MA(C, 10)); });
Buy = And(weeklyUp, Cross(C, MA(C, 20)));

Another symbol's O, H, L, C or V aligned to this symbol's dates (last known value on each bar). Write the ticker as plain text so it can be loaded before the run; up to 6 per formula. Use "~NAME" for one of your composites.

var spy = Foreign("SPY");
var riskOn = spy.gt(MA(spy, 200));

One of your saved composites. C is the combined value, V is how many symbols went into it that day.

var breadth = Composite("ABOVE200");
Filter = breadth.gt(60);

Build your own index across every symbol in an Explore or Scan run. method: avg, sum, median, min, max, or pct (share of symbols where the value is true, 0–100). Saved under your account after the run and readable from any formula with Composite() or Foreign("~NAME").

AddToComposite(C.gt(MA(C, 200)), "~ABOVE200", "pct");
AddToComposite(ROC(C, 20), "~MOM", "median");
Filter = 1;

Where this symbol ranks on the value against every other symbol on the same day: 1 = highest. Works in Analysis runs (the server runs the formula twice); on a single chart it is blank.

var r = Rank(ROC(C, 120), "mom");
Buy = r.le(10);
Sell = r.gt(30);

Same ranking as a percentile: 100 = best on that day, 0 = worst.

How many symbols had a value for that ranking on each day.

Value of a line you drew on this symbol's chart, on every bar. A horizontal line is constant; a trendline starts at its first point and extends right; a Fibonacci drawing returns the retracement level for ratio (0 = the second point, 1 = the first, 0.618 = the golden ratio). Blank on symbols without a drawing of that name. Drawings are saved with your account, so scans and backtests see them too.

var support = Study("Line 1");
Buy = Cross(C, Study("Trend 1"));
Sell = C.lt(support);

Raise an alert when the condition is true. kind: buy, sell, info, warning. Explore shows every alert in the range; a Daily scan records alerts on the newest bar, keeps a history and sends an in-app notification. At most 10 per formula.

AlertIf(Cross(C, MA(C, 200)), "Crossed above the 200-day average", "buy");

1 on bars that complete the pattern. 33 patterns: Doji, LongLeggedDoji, DragonflyDoji, GravestoneDoji, SpinningTop, Hammer, HangingMan, InvertedHammer, ShootingStar, BullMarubozu, BearMarubozu, BullBeltHold, BearBeltHold, BullEngulfing, BearEngulfing, BullHarami, BearHarami, Piercing, DarkCloudCover, TweezerBottom, TweezerTop, BullKicker, BearKicker, MorningStar, EveningStar, ThreeWhiteSoldiers, ThreeBlackCrows, ThreeInsideUp, ThreeInsideDown, ThreeOutsideUp, ThreeOutsideDown, RisingThree, FallingThree. Reversal patterns (hammer, stars, tweezers, belt holds) require the matching 5-bar move before them. Every one also exists as Cdl + name, e.g. CdlHammer(). Run the Candle patterns analysis to see which ones actually worked on your list.

Buy = And(CdlBullEngulfing(), C.gt(MA(C, 200)));
Sell = CdlBearEngulfing();

The list of pattern names with their bars and bias (bullish, bearish, neutral).

Run one of your saved formulas first, so its functions and variables can be used here — build your own indicator library once and reuse it. Matches the exact title of a formula you own; up to 8 includes, nested at most 3 deep. Library code should define values and functions, not Buy/Plot.

include("My indicators");
Buy = Cross(C, MyTrendLine(C, 20));

Remember a value (number, text or series) after this run ends. In an analysis the runtime that ran one symbol keeps it for the next symbols it runs, so it can count or collect across symbols — but the order of symbols is not fixed; use Rank() or AddToComposite() for cross-symbol values you rely on.

StaticVarSet("runs", StaticVarGet("runs", 0) + 1);

Read a remembered value, or the fallback if there is none.

Forget one remembered value, or all of them when called without a name.

Write a label above (or below, position: "below") the bars where the condition is true, or at one bar number (negative counts from the end). Pass a series as the text to print its value at each bar.

PlotText("breakout", Cross(C, HHV(H, 20).ref(-1)));
PlotText(RSI(14), Cross(30, RSI(14)), null, { position: "below", color: "orange" });

Draw your own candles (or bars with style: "bars") — e.g. a spread, a ratio or Heikin-Ashi computed in the formula. Goes to its own pane unless pane: "price".

var r = C.div(Benchmark());
PlotOHLC(O.div(Benchmark()), H.div(Benchmark()), L.div(Benchmark()), r, "vs index");

A horizontal reference line at a fixed level in a pane (e.g. 30 and 70 for RSI).

Plot(RSI(14), "RSI", { pane: "RSI" });
PlotGrid(70, { pane: "RSI", color: "red" });
PlotGrid(30, { pane: "RSI", color: "green" });

Shade the area between two series — a channel, a cloud or a band.

PlotBand(BBandTop(C, 20, 2), BBandBot(C, 20, 2), "Bollinger", { color: "#8b5cf6" });

Tint the price pane behind every bar where the condition is true, e.g. to show regimes.

PlotBackground(C.lt(MA(C, 200)), "red");

Color each candle, usually with ColorIf.

SetBarColor(ColorIf(C.gt(MA(C, 50)), "green", "gray"));

When more symbols signal than there are free positions, the highest absolute score is bought first.

PositionScore = ROC(C, 120);

Negative = percent of equity (−20 is 20%). Positive = an amount of money per trade.

PositionSize = -20;

type: loss, profit, trailing, nbar. mode: percent, point, bars. The amount can be a series — its value on the signal bar is used, so ATR-based stops work. A gap through the stop fills at the open. If one bar touches both the stop and the target, the stop is assumed first.

ApplyStop("loss", "percent", 8);
ApplyStop("trailing", "point", ATR(14).mul(3));
ApplyStop("nbar", "bars", 20);

Override a setting from inside the formula: InitialEquity, MaxOpenPositions, CommissionPercent, CommissionMinimum, CommissionPerShare, SlippagePercent, AllowShort, TradeDelay, TradePrice, RoundLot, PositionSizePercent, PointValue, MarginDeposit, TickSize, MaxScaleIns, ScaleOutPercent, BacktestMode, RebalanceBars, WorstRankHeld.

SetOption("MaxOpenPositions", 5);
SetOption("CommissionMinimum", 1);
SetOption("PointValue", 50);
SetOption("MarginDeposit", 12000);

How much to buy on each entry. mode percent: % of equity. value: money per trade. shares: a number of shares or contracts. risk: risk amount % of equity on the stop distance (in price points) — the size is chosen so hitting the stop loses exactly that much. Amount can be a series.

SetPositionSize(1, "risk", ATR(14).mul(2));
ApplyStop("loss", "point", ATR(14).mul(2));

While a trade is open, add another position of the normal size on bars where this is true (pyramiding). Limited by MaxScaleIns (default 3) and free cash. The entry price becomes the average.

ScaleIn = Cross(C, HHV(C, 20).ref(-1));

While a trade is open, sell ScaleOutPercent (default 50%) of it on bars where this is true. Each partial sale is listed as its own trade with reason scale-out.

ScaleOut = C.gt(Ref(C, -10).mul(1.1));
SetOption("ScaleOutPercent", 33);

rotation ignores Buy and Sell: every RebalanceBars bars it ranks all symbols by PositionScore, holds the top MaxOpenPositions and sells a holding once its rank is worse than WorstRankHeld. Negative scores are shorted when shorts are allowed.

SetBacktestMode("rotation");
SetOption("MaxOpenPositions", 5);
SetOption("RebalanceBars", 21);
PositionScore = ROC(C, 120);

Record a value on the signal bar of every trade. The backtest report adds a Trade metrics tab comparing winners with losers and splitting trades into groups by this value — a quick way to find a filter.

TradeMetric("RSI at entry", RSI(14));
TradeMetric("Distance from 200d %", C.div(MA(C, 200)).sub(1).mul(100));

Comparisons. JavaScript cannot overload > on series, so write C.gt(MA(C, 20)) instead of C > MA(C, 20).

Arithmetic on series or numbers.

Combine conditions.

Same as Ref(a, -n).

Same as Cross(a, b).

Value at bar i; negative counts from the end (a.at(-1) is the last bar).

Value on the last bar.

Bar-by-bar custom calculation. Use .v for the raw Float64Array when writing loops.

var streak = C.map((v, i) => i > 0 && v > C.v[i - 1] ? 1 : 0);