Renko Charts

A Renko chart is a price-driven representation of the market: instead of printing one bar per unit of time, it prints one brick every time price travels a fixed distance. How many bricks form in an hour is irrelevant; what matters is displacement. The name comes from the Japanese *renga* (brick), and the construction is deliberately lossy — a Renko series discards time, most intra-move noise and, in the classic form, volume, keeping only a stepwise record of net movement. Strategies built on Renko are therefore not strategies about candles at all. They are strategies about a filtered version of the price series, and most of what determines their behaviour lives in how that filter is defined.

## How bricks are built

Pick a brick size — say 20 points. A new brick is appended only when price closes more than 20 points beyond the top of the last up brick or the bottom of the last down brick. A 19-point move draws nothing; a 65-point move prints three bricks at once, all at the same instant in real time. Reversals are asymmetric: because a brick of the opposite colour must start from the far edge of the current one, flipping direction generally requires about twice the brick size. That asymmetry is what produces Renko's long unbroken runs of one colour, and it is the whole trading premise: moves smaller than the reversal threshold are suppressed, so trend-following logic survives noise that would shred it on a time chart. The cost is a structural, non-negotiable lag equal to the reversal distance at every turn. Bricks have no wicks, and the price path taken inside a brick is recorded nowhere.

## Main variants

Traditional Renko fixes the brick in points, ticks or currency and never changes it. ATR Renko derives the brick from a volatility estimate, so bricks widen in fast markets and narrow in quiet ones. Percentage Renko sizes bricks as a fraction of price, which is the usual choice for instruments spanning orders of magnitude. Median or "smoothed" Renko rebuilds brick bodies from averaged values, producing cleaner charts whose levels no longer correspond to traded prices. Adjacent families — range bars, Kagi, Line Break, Point & Figure — share the price-driven axis but use different reversal rules.

The logic layered on top is usually one of three shapes: colour-flip entries on the first brick of a new direction, confirmation entries after N consecutive bricks, or hybrid designs where Renko supplies the trend filter and a conventional chart supplies the trigger.

## What typically differentiates implementations

Brick size dominates everything else, and it is rarely reported as a ratio to the instrument's volatility, which makes settings non-comparable across markets. Next is grid alignment: two traders using the same brick size but a different anchor price or start date get bricks at different boundaries and therefore different signals from the same data — an under-discussed source of variance. Then the definition of the trigger price (closes only versus highs and lows), the resolution of the source data used to build the bricks (ticks versus one-minute bars), whether exits are managed on the brick series or on the underlying instrument, and whether the same brick size is used for entry, stop and target.

## Common mistakes

The most frequent is tuning brick size until the historical equity curve looks good; brick size is the single most overfittable parameter in the family. Others: ignoring costs relative to brick size, when a round turn's spread plus commission can consume a meaningful share of one brick; forgetting that a reversal gives back roughly two bricks before the new signal exists; reading indicator periods as durations when they are now distances; assuming a brick is a tradeable move rather than a recorded one; and using volatility-scaled bricks in a backtest without checking whether historical bricks were redrawn when the parameter updated.

## How to evaluate and backtest a version

Treat Renko as a signal generator, not as a market. Build the bricks from the finest data available, take signals from them, then simulate fills, stops and drawdown on the underlying time or tick series — most platform backtesters treat a brick as a bar whose high and low equal its body, which erases intra-brick excursions and flatters both stop placement and maximum drawdown. Replay bar by bar to confirm no historical brick changes after the fact. Sweep brick size across a range and look for a plateau of acceptable results rather than a peak, and repeat the sweep for grid offset to see how much of the result is alignment luck. Map results back to calendar time, since brick counts say nothing about exposure, holding period or financing. Finally, test across volatility regimes and instruments: a design that only works in one trending sample has not been separated from that sample.

## Decoded versions on this page

Three implementations are decoded individually and are worth comparing along the axes above: **Renko Offset Indicator** (TheStopHunter), which touches grid alignment; **Renko Charts Strategy** (PythonIA), a programmatic implementation; and **AI Backtesting Trading Strategy (Renko Charts, Bitcoin Perpetual Futures)** (The Good, The Bad And The Bitcoin), which applies the concept to crypto perpetuals. Each page states the rules, parameters and assumptions as presented in the source video.

Strategies in this concept (4)

Frequently asked questions

Do Renko charts repaint?

It depends on the variant. With a fixed brick size built from closing prices, a printed brick does not change afterwards — although the current, unconfirmed brick appears and disappears in real time while price oscillates below the threshold, which is not the same as a stable signal. Variants that derive brick size from a rolling volatility measure can redraw historical bricks when that measure updates, so the chart you study today may not be the chart that existed live. Smoothed or median Renko additionally draws bodies at averaged values that were never traded.

How should I choose the brick size?

There is no universal value, and the choice is a trade-off with two hard boundaries. The brick must be large enough that spread, commission and slippage on a round turn are a small fraction of it; and small enough that the roughly two-brick reversal lag does not return most of the move you captured. Express the brick as a fraction of the instrument's average range so settings are comparable across markets, test a range of values rather than one, and prefer a region where results are stable over an isolated best value.

Can I backtest a Renko strategy in a standard backtester?

Only with care. Most platform backtesters feed the brick series to the engine as if bricks were ordinary bars whose high and low equal the body extremes. Because the real price path inside a brick is never recorded, stops and targets placed within a brick resolve far too favourably and maximum drawdown is understated. The more defensible setup is to generate signals from the bricks but simulate execution, stops and equity on the underlying time or tick series, with an explicit cost model.

How is Renko different from Heikin Ashi or range bars?

Heikin Ashi is still time-based: one candle per period, with smoothed open/close values. Renko removes time entirely. Range bars are also price-driven but close after a fixed high-to-low range and have no asymmetric reversal rule, so they flip direction as easily as they continue. That reversal asymmetry — needing roughly double the brick size to change colour — is what distinguishes Renko and what produces both its noise filtering and its lag.

Do standard indicators work on Renko charts?

They compute without error, but their meaning changes. A 14-period moving average becomes 14 bricks, which is a distance rather than a duration, so its effective lookback stretches and contracts in calendar time as volatility changes. Oscillators reach extremes quickly during a run of same-coloured bricks, since the input series is monotone by construction. Volume and time-based studies are the most distorted, because bricks can print several at once or none for hours.

Does Renko suit every market and session?

The construction is instrument-agnostic, but the premise is not: it needs sustained directional movement large relative to trading costs. Wide-spread or strongly mean-reverting instruments penalise it, because the reversal threshold is paid repeatedly with no trend to recover it. Markets that trade continuously add a separate issue — costs such as funding accrue by clock time while Renko ignores clock time, so holding duration has to be tracked outside the chart to evaluate the strategy honestly.

All strategy concepts · Explore all strategies