Optimization returns the same set of parameters, does not cycle through the parameter ranges

azzam

Trying to run a simple strategy, but the optimizer doesn't seem to cycle through all parameter options. It returns the same set of parameters.

This is my strategy

class Momentum(BasicStrategy):
    params = (
        ('maperiod', 25),
        ('rsiperiod', 3),
    )

    def __init__(self):
        # Add a MovingAverageSimple indicator

        self.wma = bt.indicators.WMA(self.datas[1], period=self.p.maperiod)
        self.rsi = bt.indicators.RSI(self.datas[1], period=self.p.rsiperiod)
        self.slope = bt.indicators.RateOfChange100(self.wma, period=1)
        self.buy_signal = bt.And(self.rsi() < 35.0,  self.slope(-1) > 0.0)
        self.sell_signal = bt.And(self.rsi() > 65.0, self.slope(-1) < 0.0)
        self.order = None
    
    def next(self):
        if self.order:
            return
        if self.buy_signal and self.position.size <= 0:
            self.buy()
        if self.sell_signal and self.position.size >= 0:
            self.sell()

The optimization part

    cerebro.optstrategy(
        Momentum,
        maperiod=range(30, 50, 10),
        rsiperiod=range(3, 5, 2),
    )

    
    cerebro.addanalyzer(bt.analyzers.DrawDown, _name="drawdown")
    cerebro.addanalyzer(bt.analyzers.Returns, _name="returns")
 
    # clock the start of the process
    tic = time.perf_counter()

    # Run over everything
    results = cerebro.run()

    # clock the end of the process
    toc = time.perf_counter()
    # import pdb; pdb.set_trace()
    for run in results:
        for strat in run:
            par_list = [[strat.params.maperiod,
                         strat.params.rsiperiod,
                         strat.analyzers.returns.get_analysis()['rnorm100'],
                         strat.analyzers.drawdown.get_analysis()['max']['drawdown'],
                         0
                         #x[0].analyzers.sharpe.get_analysis()['sharperatio']
            ] for x in results]
    
    par_df = pd.DataFrame(par_list, columns=['maperiod',
                                             'rsiperiod',
                                             'return',
                                             'dd',
                                             'sharpe'])
    print(par_df)

here is the output

  maperiod  rsiperiod    return        dd  sharpe
0        40          3  0.000152  8.268218       0
1        40          3  0.000152  8.268218       0
Time used: -217.93248052499985

First, it seems that I arbitrary get a single set of parameters, here for example it is (40, 3). If I change the parameter range, I get another pair that keeps repeating. Notice it is not the first of the range (which should be (30,3)) and note top of the range.

Any help into this?