Array of questions / clarifications to get started

[julian-dotcom]

Hi Oleg, thanks for the cool library! I think it has a lot of potential, and I am getting started learning how it works right now.
There are a few questions that I would like some clarification with, that the documentation doesn't quite answer. Hopefully, I don't ask any redundant questions. Thanks in advance :)

1. What does freq="1D" do in
   pf = vbt.Portfolio.from_signals(mult_comb_price, entries, exits, freq='1D')

2. It would be really cool, if you could have a few examples of more complicated strategies than just RSI or SMA. That way, we can kind of see how to develop more intricate strategies than just comparing 2 values.

3. How do you plot a custom strategy? I am testing pandas_ta "supertrend" and want to see the results plotted, i.e. OHLCV chart with the trend lines.

Thanks in advance, I hope to post additional questions as the pop up. :)

[polakowo]

Hi @julian-dotcom,

1. It tells portfolio that the frequency of your data is daily and is used to compute annualization factor for various return metrics. Sometimes frequency cannot be automatically determined, that's why it must be set explicitly.
2. You can find explanation of all building blocks in https://vectorbt.dev/docs/indicators/factory.html. If you're looking for an example, see how default indicators were created: https://github.com/polakowo/vectorbt/blob/master/vectorbt/indicators/basic.py.
3. vectorbt uses plotly, and in plotly you can append traces to the same figure: fig = df.vbt.ohlcv.plot(); your_indicator.trending_line.vbt.plot(fig=fig). Again, you can look and take inspiration of how basic indicators are implemented.

[julian-dotcom]

Thanks! What are the other possible parameters for freq? "1M" or "1m" for minutes? How can add a freq of 5 minuts?

Regardless, I think it would be super helpful, if you could have a more complex strategy as an example on the web page. I am not a novice anymore (not super experienced either) but I really struggled getting started.

[polakowo]

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Timedelta.html#pandas.Timedelta

"5 minutes"

Give me an example of a complex strategy, I will implement and include it in the docs.

[julian-dotcom]

Thank you so much for the offer.

I am currently trying to test a supertrend strategy using pandas_ta supertrend() function:
SUPERTREND = vbt.IndicatorFactory.from_pandas_ta('supertrend')

My problem is that I want to test a wide array of parameters, like you did in the example on GitHub for crossovers:
windows = np.arange(2, 101)
fast_ma, slow_ma = vbt.MA.run_combs(price, window=windows, r=2, short_names=['fast', 'slow'])

In this example all potential slow and fast combinations are tested.

How can I do the same, for a strategy that takes 2 parameters, i.e. length and multiplier:

supertrend = SUPERTREND.run(high=df['high'], low=df['low'], close=df['close'], length=windows, multiplier=windows)
Let's say I generate a window of np.arange(2, 31), how do I test all potential combinations of this window for the length and multiplier? E.g.: When length = 1, test all values from 2-31 for multiplier.

Thanks and have a great weekend

[polakowo]

param_product=True

[CharlieAlphaFox]

Really cool librabry, on the original question, can I expand on the 2d point ?

A simple combination of indicators from the example of RSI on the BTCBRL example, is it possible to:

A - Get a signal from MACD?
B - How would you combine it with the RSI/other signals?
C - Add more indicators/signals like ADX or others from pandas.ta into one strategy ?

#!/usr/bin/python3
from time import sleep
import vectorbt as vbt

data = vbt.BinanceData.download('BTCBUSD', start='2021-12-06', interval='5m').get()
precio  = data.get('Close')

indicador_rsi = vbt.RSI.run(precio, window=24)
ind_macd = vbt.MACD.run(precio, fast_window=6, slow_window=34, signal_window=9)
# output_names=['macd', 'signal'],

signals = ind_macd.signal

#buy_signals = ind_macd.macd(signals, crossover=True)
#sell_signals = myind.macd(signals, crossover=True)

entrada = indicador_rsi.rsi_below(20, crossover=True)
salida= indicador_rsi.rsi_above(80, crossover=True)

print('MACD: signals')
sleep(1)

#print(buy_signals)
#print(sell_signals)

[polakowo]

@CharlieAlphaFox

1. https://vectorbt.dev/docs/indicators/basic.html#vectorbt.indicators.basic.MACD.signal
2. You can combine arrays using pandas, or use comparison methods of the indicator: https://vectorbt.dev/docs/indicators/basic.html#vectorbt.indicators.basic.MACD.signal_above
3. Define your own indicator function that takes your inputs, all parameters, calls the respective indicators by distributing parameters among them, combines their outputs, and returns them. An example for pandas-ta: https://vectorbt.dev/docs/indicators/factory.html#without-numba