Heikin-Ashi: Synthetic Candles and Real Prices

Keep original and transformed fields separate

A common Heikin-Ashi close is the average of the current real open, high, low and close. Its open is the average of the previous Heikin-Ashi open and close. The synthetic high is the maximum of the real high and the two synthetic body values; the low is the corresponding minimum. TradingView describes the Heikin-Ashi chart type.

The recursive open requires a seed convention at the start of history. Different starting data can affect early candles. A transformed candle should therefore not overwrite the original OHLC columns in a dataset used for accounting, fill analysis or order validation.

Calculate an original example

Assume the current real bar has open 100, high 106, low 99 and close 104. Its Heikin-Ashi close is (100 + 106 + 99 + 104) / 4 = 102.25. If the previous synthetic open and close were 98 and 100, the next synthetic open is 99.

The synthetic high is 106 and low is 99, so the transformed body runs from 99 to 102.25. Neither body endpoint should be assumed to be the actual opening or closing transaction. The real close remains 104 for a calculation that requires the observed close.

For the next real bar, suppose open 105, high 110, low 104 and close 109. Its synthetic close is 107, while its synthetic open is (99 + 102.25) / 2 = 100.625. The resulting synthetic low is 100.625, below that real bar's low of 104. A stop or fill at that synthetic low would not be justified by the bar's actual traded range.

Understand apparent continuity

Because the synthetic open carries earlier information, successive candles can show smoother bodies and repeated colors. That is the intended transformation, not proof that gaps disappeared from the real market. A rapid change can take time to alter the smoothed appearance.

Measures calculated from transformed candles also become measures of the transformed series. An RSI on synthetic closes is not the same experiment as RSI on real closes. The chart style is therefore part of the input specification, not merely a cosmetic display setting.

Separate signal generation from execution

TradingView’s non-standard-data documentation distinguishes synthetic chart prices from market prices relevant to execution. A test may use a transformed series to define a signal, but it still needs a realistic decision timestamp and actual executable prices for entries and exits. An ordinary-price fill setting does not fix every possible look-ahead or bar-path assumption.

Compare Renko's price-driven construction, inspect what ordinary OHLC can establish, and follow the testing protocol. Preserve both real and transformed columns so every attractive historical result can be traced to prices that could actually have been observed and traded.