Ichess Strategy added #231
Conversation
user_data/strategies/Ichess.py
Outdated
| df['kumoBreakoutScore'] + | ||
| df['senkouCrossScore'] + | ||
| df['chikouCrossScore'] | ||
| ).cumsum() |
There was a problem hiding this comment.
this will not work identically during backtesting and live runs.
backtesting has all data available - so you'll calculate this on a range of potentially 100ds of days (the end value can be pretty high).
You should instead use .rolling().cumsum() (didn't check for other such problems in this strategy ...).
There was a problem hiding this comment.
thanks but I got two errors:
TypeError: rolling() missing 1 required positional argument: 'window'
which window is best default for rolling?
after add a number .rolling(9).cumsum()
AttributeError: 'Rolling' object has no attribute 'cumsum'
Whats the best change for fix?
user_data/strategies/Ichess.py
Outdated
| else: | ||
| return 0 | ||
|
|
||
| df['tkCrossScore'] = df.apply(calcTkCross, axis=1) |
There was a problem hiding this comment.
This is a very slow way of doing things.
i don't see any reason not to do this in a vectorized manner - which will be a lot quicker.
It may seem like a more complicated approach (which it might be if you're not used to thinking that way) - but the performance you'll gain from that will be substancial - especially for backtesting/hyperopt.
There was a problem hiding this comment.
oh than what can we do?!! you said befour i can use dt.hour! can I use some thing like dt.hour instade of that, here?
There was a problem hiding this comment.
please do your own research - pandas is a wide enough topic so you can find probably 100ds of resources there - i'd suggest to properly learn pandas first.
There was a problem hiding this comment.
np.vectorize is NOT what i meant, and will have no performance benefit over using apply(). it's the equivalent to pd.apply - which both execute the function n times (n = number of rows).
A vectorized approach is for example the following: dataframe['whatever'] = dataframe['low'] / dataframe['high'] - which is operating on series, not on single values.
Okay thanks👌👌👌 I think I will work more on it than ;) |
|
I reached good things about this strategy! |
|
I see no updates to this PR that address the review so would recommend this is closed and reopened if any further work has been done. |
Ichess Strategy ichimoku_score
Associate various ichimoku signals with a score.
For example, bullish signal => positive score,
and bearish signal => negative score. If the total score is above 0, it may indicate a bullish trend .
Otherwise, if it is below 0, it may indicate a bearish trend .
We used two smoothed moving averages to find the trend.
More info:
#97
https://www.tradingview.com/script/P1bybHZA-Ichimoku-Cloud-Signal-Score-v2-0-0/
Author: @mablue (Masoud Azizi)
github: https://github.com/mablue/