CorrelationTrend: Correlation Trend (Ehlers)

The Pearson correlation between price and time over the window, from −1 to +1. It measures how steadily price has been moving: +1 is a perfectly straight rise, −1 a perfectly straight fall, and 0 means price has no straight-line relationship with time at all.

Signature

CorrelationTrend(period, on=<source>)

Argument	Type	Required	Notes
period	int Param or literal	Yes	Window. Default 20, bounds 3–400.
on=	symbol, OHLCV field, or indicator	No	Source. Defaults to the trade symbol’s close.

Formula

$$ \text{CT} = \frac{\sum_i (i - \bar{i})(C_i - \bar{C})}{\sqrt{\sum_i (i - \bar{i})^2 \sum_i (C_i - \bar{C})^2}} $$

with i the bar’s position in the window (oldest first) and C the source value.

Warmup

period bars; 0 until then, and 0 on a flat window.

Reading the value

ct.value
ct.value > 0.8
ct.crossed_above(0)

Example

class StraightLineTrend(Strategy):
    """Enter only when price and time are strongly correlated."""
    trade = Symbol()
    sizing = FixedCapital()

    def execute(self):
        ct = CorrelationTrend(20, on=self.trade)
        if ct.value > 0.8 and self.trade.is_flat:
            self.buy(self.trade, reason=f"Correlation {ct.value:.2f}")
        if ct.value < 0 and self.trade.is_long:
            self.sell(self.trade, self.trade.position, reason="Trend correlation lost")

Notes

• John F. Ehlers, “Correlation as a Trend Indicator”, Technical Analysis of Stocks & Commodities, May 2020.
• The square of this value is LinearRegression.r2: the same fit, signed here so it carries the direction.
• Proven by an assert-via-trade run on SPY daily (matched a strategy-side Pearson computation to 1e-4 on every bar; zero alarm trades).