RC2_PARAM_STEP_NOT_POSITIVE

Param <name> declares opt_step of zero or less.

What it means

opt_step is the DISTANCE between two values the optimizer tries. Zero never reaches the end of the range — it is an infinite grid — and a negative step walks away from the end it is supposed to reach.

Both used to be accepted and then defended against downstream: the frontend’s own walk clamps a non-positive step to 1, which means the sweep ran on different values than the author declared, silently.

Why it happens

A placeholder that was never filled in, or a sign that came along for the ride:

class ZeroStep(Strategy):
    period = Param(20, min=2, max=200, opt_step=0)
    trade = Symbol()
    sizing = FixedCapital()

    def execute(self):
        sma = SMA(self.period, on=self.trade)
        if sma.value > 0 and self.trade.is_flat:
            self.buy(self.trade)

class NegativeStep(Strategy):
    period = Param(20, min=2, max=200, opt_step=-5)
    trade = Symbol()
    sizing = FixedCapital()

    def execute(self):
        sma = SMA(self.period, on=self.trade)
        if sma.value > 0 and self.trade.is_flat:
            self.buy(self.trade)

How to fix it

Write a step greater than zero, or drop opt_step and take the default (1 for an int, 0.01 for a float). Direction is not the step’s business — the grid always walks from opt_min up to opt_max.

class RealStep(Strategy):
    period = Param(20, min=2, max=200, opt_step=5)
    trade = Symbol()
    sizing = FixedCapital()

    def execute(self):
        sma = SMA(self.period, on=self.trade)
        if sma.value > 0 and self.trade.is_flat:
            self.buy(self.trade)

Param()
RC2_PARAM_INT_STEP — the step is positive but fractional on an integer parameter.
RC2_PARAM_RANGE_INVERTED — the range the step is meant to walk is empty.