Are Renko Backtests Too Good to Be True?

Are Renko Backtests Too Good to Be True?

By Rami Alame (Akylles) | Trade Feeld

Level: Intermediate | Instruments: Forex, Indices, Crypto

Yes, Renko backtests can be too good to be true when they treat chart-generated brick levels as executable market prices. Renko charts simplify price movement, but that simplicity can hide the sequence, timing, and trading costs that determine whether a strategy was actually tradable. A strong-looking result is not automatically invalid. It needs testing against underlying market data, realistic fills, and signals that were genuinely available at the time. This article is trading education only, not financial advice.

1. Understand what a Renko brick represents

A Renko chart draws a new brick when price moves far enough under its construction rules. Unlike a time-based candle, a brick does not represent a fixed duration. Several bricks may appear during one sharp move, while a quiet period may produce none.

Renko brick construction is a transformation of market data, not a record of completed trades at every displayed level. That distinction is the starting point for evaluating any strategy built on it.

Construction varies by platform and configuration. Important settings include:

• Fixed brick size versus a size derived from volatility, such as average true range.
• Closing-price inputs versus high-low or more granular inputs.
• The timeframe and depth of the underlying data.
• Reversal requirements and treatment of gaps.
• Whether unfinished, projected bricks are displayed.

Two charts of the same instrument can therefore show different brick sequences without either containing a simple calculation error. They may be using different information or rules.

Record those settings before testing. Platform documentation, available through providers such as TradingView, is the place to verify how a particular implementation builds historical and real-time bricks. Do not assume every Renko engine behaves identically.

2. Separate attractive chart geometry from executable prices

A backtest can become misleading when it buys or sells at a synthetic brick open or close simply because that value appears on the chart.

Suppose an underlying market jumps across several brick thresholds. The chart may draw a neat staircase, but that does not establish that a trader could transact at each step after receiving the signal. A completed brick can identify a threshold crossing without proving that its plotted price was still available when an order arrived.

Renko backtest accuracy depends on both signal accuracy and execution accuracy. A strategy can get one right and the other wrong.

Ask three questions about every simulated entry:

• What underlying market event confirmed the signal?
• When could the strategy first observe that event?
• What executable price was available after that observation?

Apply the same questions to exits. A target touched on a reconstructed chart is not necessarily a filled limit order. A stop crossed during a gap is not necessarily an exit at the stop price.

For general terminology around chart types and backtesting, Investopedia provides background. The decisive evidence, however, is your platform’s fill model and the underlying data used in your test.

3. Investigate the missing intrabar price path

A time-based candle provides an open, high, low, and close. It does not fully describe the order of all trades inside that interval.

This creates the Renko intrabar price path problem. If both an upward threshold and a downward threshold fall inside one candle’s range, different sequences of movement can generate different signals and outcomes.

The market might move upward first, reverse, and then recover. It might move downward first and rally later. Both paths can share the same candle summary, yet produce different entries, stops, or reversal bricks.

A historical engine may impose an assumed sequence. Another may use only closes and ignore intermediate movement. Neither approach should be mistaken for a complete reconstruction of what happened.

Lower-timeframe data can reduce uncertainty, but it does not eliminate movement hidden inside those smaller candles. Tick data offers more sequencing detail, although trade ticks alone may still omit the bid-ask quotes needed to model execution accurately.

When the order of events cannot be established, classify the trade as ambiguous or test alternative plausible paths. Choosing whichever sequence produces the best result quietly turns missing information into an advantage.

4. Worked example: the same candle, different outcomes

Hypothetical example only: every number below is invented for illustration, not taken from a market or study.

Assume an instrument has a completed Renko brick ending at 100, with a fixed brick size of 10. Under the selected construction rules, an upward brick confirms when the underlying price reaches 110.

The strategy buys after that confirmation. For simplicity, assume an idealized entry at 110, a stop at 100, and a target at 120, with no costs. These assumptions isolate the path problem; they do not establish achievable fills.

One underlying candle has these hypothetical values:

• Open: 100.
• High: 120.
• Low: 90.
• Close: 110.

Now compare two possible paths:

1. Path A: 100 → 90 → 110 → 120 → 110. The decline occurs before the entry signal. After the assumed entry at 110, price reaches the target at 120.
2. Path B: 100 → 110 → 90 → 120 → 110. The entry signal arrives before the decline. After the assumed entry at 110, price reaches the stop at 100 before later reaching 120.

The candle summaries are identical. The hypothetical trade outcomes are not. For this comparison, consider only the first trade and allow no re-entry.

A backtest using only that candle cannot establish which outcome occurred. Even after resolving the path, spread, latency, and slippage still need attention. If the first executable offer after confirmation is above 110, the actual entry assumption must change as well.

5. Distinguish repainting from ordinary brick development

An unfinished brick changing as price moves is not necessarily a software fault. It may be a normal preview. The problem arises when a backtest treats that preview as a confirmed historical signal.

Renko repainting risk concerns whether the information you saw in real time remains the same after additional data, recalculation, or a chart reload.

Potential sources include:

• Projected bricks appearing and disappearing before confirmation.
• Historical reconstruction using different data granularity from live updates.
• Volatility-based brick sizing that changes how history is rebuilt, depending on implementation.
• Revised source data or a different starting point for construction.

A volatility-based brick size does not automatically make a test invalid. The key question is whether each historical decision used only the volatility information available then, and whether past bricks remain stable under the engine’s rules.

Keep a live or replay log of signal timestamps, brick sizes, and underlying prices. Compare that log with the chart after reloading it. A visually tidy historical chart is weaker evidence than a timestamped record of what the strategy actually knew.

6. Common mistakes across Forex, indices, and crypto

The first mistake is assuming one cost model fits every instrument.

In Forex, a single plotted price can hide the bid-ask spread. Check whether the feed represents bid, ask, or another price, and model the correct side for entries and exits. Include applicable commissions and overnight financing.

For indices, distinguish an index calculation from a tradable futures contract, ETF, or CFD. They have different sessions, costs, and pricing characteristics. Continuous futures data also needs careful treatment around contract rolls.

In crypto, identify the exchange and product. Spot and perpetual futures are not interchangeable; fees, funding, liquidity, and venue-specific price paths can differ.

Other common mistakes include:

• Optimizing brick size repeatedly on the same sample.
• Ignoring gaps because the chart draws continuous-looking steps.
• Counting several bricks as several independent execution opportunities.
• Assessing drawdown only at brick closes, missing adverse movement between them.
• Assuming costs remain unchanged around major scheduled releases.

For event timing, check the official Federal Reserve meeting calendar and BLS CPI page. These sources establish schedules and releases; they do not validate a Renko strategy or predict market direction.

7. A step-by-step validation checklist

1. Freeze the specification. Document the instrument, venue, session, source feed, brick rules, and strategy logic before evaluating results.
2. Check data resolution. Identify which signals depend on an unknown intrabar sequence. Obtain finer data where possible and flag unresolved cases.
3. Separate signals from fills. Generate signals from confirmed Renko information, but simulate execution on underlying prices available afterward.
4. Audit order timing. Ensure an order cannot fill before the signal exists. Pay special attention when multiple bricks form from one update.
5. Model costs and constraints. Include spread, commissions, slippage, financing or funding, and relevant contract specifications.
6. Inspect individual trades. Review examples around reversals, gaps, session boundaries, and releases. Reconcile each signal and fill with the source data.
7. Test stability. Compare nearby brick settings, separate market periods, and an untouched out-of-sample segment. Avoid selecting only the best-looking configuration.
8. Forward-test and reconcile. Log signals and hypothetical executions as they occur. Compare them with the later historical reconstruction before drawing conclusions.

The bottom line

Renko is useful for organizing price movement, but a cleaner chart does not create cleaner execution. Credible testing preserves the distinction between synthetic bricks, observable signals, and executable prices.

Treat unusually smooth results as a reason to inspect assumptions, not as proof of an edge. If profitability disappears when ambiguous paths and realistic costs are included, the original result was not robust evidence.

Continue learning free on Trade Feeld, and follow @tradefeeld on X for trading education. The goal is better testing discipline, not confidence borrowed from an attractive backtest.