FVG indicator: questions and answers

There is a working detector in your browser on the home page: paste candles or pick a real sample window, and it applies the rule. The TradingView (Pine Script v6) and MetaTrader 5 (MQL5) indicators exist as source files that you can read on this site, but they have not been compiled or run on either platform yet, so they are not released.

Take three candles in a row. A bullish gap exists when the low of candle 3 is above the high of candle 1. A bearish gap exists when the high of candle 3 is below the low of candle 1. Only wicks count, the comparison is strict so touching is not a gap, and the gap exists only once candle 3 has closed.

The Pine source draws a gap only on the closing tick of candle 3 and the MQL5 source only reads closed bars, so neither marks a gap on the forming bar. A drawn box is later extended or stopped, and with showFilled off the Pine source deletes a filled box. This describes the code. It has not been observed on a live chart, and no repaint test report exists yet.

minSize (minimum gap height in price units), lookback (bars), showFilled, extendOpen, bullColor and bearColor in both sources, plus alertOnNew in the MQL5 source only. The settings page explains each one and lists the defaults.

Not yet. MT4 uses MQL4, a different language from MQL5, and the MQL5 file does not run there. An MQL4 version is planned and has no date. The rule is short, so the MQL5 source can serve as a guide for porting it.

The source files and the browser detector can be read and used here. A free limited version and a paid one-time licence are planned, but neither is on sale, no price is set and nothing can be bought on this site.

Alerts are early access. The Pine source defines two alert conditions and the MQL5 source has an alertOnNew input, both untested. A Telegram alert feed is planned and depends on the signal scanner supporting it, which it does not yet.

The rule needs only the highs and lows of three consecutive candles, so it can be applied to any instrument and timeframe. Whether the gaps it marks are useful on a market is a different question that this site does not answer.

There is no accuracy figure. The indicator marks a pattern that has already formed and makes no prediction to score. To measure how marked gaps behave afterwards you would need a written definition of the outcome and a long sample that includes the gaps that did nothing. That has not been done here.

The detector is written as a small TypeScript package named fvg-detector, with a readme and a test that checks it against the detector on this site. It is not published to npm yet and its licence has not been chosen.

Educators and indicator authors define details differently: whether a tiny gap counts, whether the gap is measured from wicks or bodies, and what filled means. This site states one definition exactly and exposes the size as a setting. If another tool disagrees, compare its definition with this one before comparing charts.