Identify a cross-over

nerdy

What is the best way to identify that price has crossed over an indicator. Do I need to use self.candles[:,2][-2] in combination with self.vars[‘previous_candle_indicator_value’] to determine if price has crossed?
Or do we have access to something like this? https://tulipindicators.org/crossover

markus

There is actually a function already implemented:
https://github.com/jesse-ai/jesse/blob/136ae029ea8a76d05eb1700c2448474c253e49ca/jesse/utils.py#L130
You could use that, but for me thats easier:
self.ema20[-1] > self.ema100[-1] and self.ema20[-2] <= self.ema100[-2]
This checks if the current [-1] value of the ema 20 is greater than the ema 100 and before that the ema 20 was either smaller or the same. Thats a cross from the ema 20 above the ema 100.
The cross below would be: self.ema20[-1] < self.ema100[-1] and self.ema20[-2] >= self.ema100[-2]
You can use a fixed value to compare too. Like:
self.ema20[-1] > 150 and self.ema20[-2] <= 150]
The indicators would need sequential=True. Using self.vars is way more complicated in my eyes, but you could the last/previous values you stored there for this too.

nerdy

markus Cheers. I'll give it a go.

nerdy

Hmm, appears I can’t use sequential with the ichimoku indicator.
ichimoku_cloud(candles, conversion_line_period=9, base_line_period=26, lagging_line_period=52, displacement=26)
I added sequential=True and received.

============================== EXCEPTION TRACEBACK:
  File "/usr/local/bin/jesse", line 8, in <module>
    sys.exit(cli())
  File "/usr/local/lib/python3.8/dist-packages/click/core.py", line 764, in __call__
    return self.main(*args, **kwargs)
  File "/usr/local/lib/python3.8/dist-packages/click/core.py", line 717, in main
    rv = self.invoke(ctx)
  File "/usr/local/lib/python3.8/dist-packages/click/core.py", line 1137, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/local/lib/python3.8/dist-packages/click/core.py", line 956, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/local/lib/python3.8/dist-packages/click/core.py", line 555, in invoke
    return callback(*args, **kwargs)
  File "/usr/local/lib/python3.8/dist-packages/jesse/__init__.py", line 286, in backtest
    backtest_mode.run(start_date, finish_date, chart=chart, tradingview=tradingview)
  File "/usr/local/lib/python3.8/dist-packages/jesse/modes/backtest_mode/__init__.py", line 58, in run
    simulator(candles)
  File "/usr/local/lib/python3.8/dist-packages/jesse/modes/backtest_mode/__init__.py", line 225, in simulator
    r.strategy._execute()
  File "/usr/local/lib/python3.8/dist-packages/jesse/strategies/Strategy.py", line 827, in _execute
    self._check()
  File "/usr/local/lib/python3.8/dist-packages/jesse/strategies/Strategy.py", line 619, in _check
    if self.should_long():
  File "/home/btc-1/strategies/E2E/__init__.py", line 28, in should_long
    if self.ichi.conversion_line > self.ichi.base_line:
  File "/home/btc-1/strategies/E2E/__init__.py", line 71, in ichi
    return ta.ichimoku_cloud(self.candles, conversion_line_period=20, base_line_period=60, lagging_line_period=120, displacement=30, sequential=True)
=========================================================================

 Uncaught Exception: TypeError: ichimoku_cloud() got an unexpected keyword argument 'sequential'

nerdy

Tried anyway and failed. You don’t have a quick example of how to use the crossed function?

Failed ichi crossover attempt.

if self.ichi.conversion_line[-2] < self.ichi.base_line[-2]:
                        if self.ichi.conversion_line > self.ichi.base_line:
  File "/home/btc-1/strategies/E2E/__init__.py", line 28, in should_long
    if self.ichi.conversion_line[-2] < self.ichi.base_line[-2]:
=========================================================================

 Uncaught Exception: IndexError: invalid index to scalar variable.

markus

Thats because Ichimoku Cloud doesn't offer sequential yet.
Compare the docs: https://docs.jesse-ai.com/docs/indicators/reference.html#ichimoku-cloud
In that special case you have to use the self.vars approach for Ichimoku.