ZigZag Indicator, Sometime Showing Incorrect Waves for Deviation: 0.1% or more

[divyankm]

Which version are you running? The lastest version is on Github. Pip is for major releases.

import pandas_ta as ta
print(ta.version)
0.4.19b0

Have you tried the development version? Did it resolve the issue?

$ pip install -U git+https://github.com/twopirllc/pandas-ta.git@development
Yes

Describe the bug
Sometime the ZigZag Waves are not formed correctly for Deviation: 0.1% or greater

To Reproduce

!pip install -U git+https://github.com/twopirllc/pandas-ta.git@development
from pandas_ta import zigzag
import pandas_ta as ta
import pandas as pd
import numpy as np
import plotly.io as pio
from datetime import date
from datetime import datetime, timedelta
from pytz import timezone
import plotly.graph_objects as go
import plotly.figure_factory as ff
import plotly.express as px
import plotly.io as pio
import yfinance as yf
# Define the stock ticker symbol (replace 'ES=F' with your desired stock symbol)
ticker = 'ES=F'
# Fetch the last 1 day of data with 1-minute intervals
data = yf.download(ticker, interval='1m', period='5d')
# Convert the datetime index from UTC to Eastern Time (US/Eastern)
data.index = data.index.tz_convert('US/Eastern')
df2 = data.copy()
# Flatten the MultiIndex and remove the ticker part (symbol name)
df2.columns = df2.columns.get_level_values('Price')

# Reset index to get a default integer-based index
df2.reset_index(drop=False, inplace=True)

# Using the Pandas-TA zigzag indicator (or other equivalent library like ta-lib or ta)
zigzag_result = ta.zigzag(
    high=df2['High'],      # High prices of the asset
    low=df2['Low'],        # Low prices of the asset
    close=df2['Close'],    # Closing prices of the asset
    legs=3,                # Minimum number of price points required for a significant trend change
    deviation=0.01,        # Minimum 1% price movement to form a new zigzag segment (1% deviation)
    retrace=True,          # Consider retraces before identifying new points
    last_extreme=True,     # Consider the last extreme point in the calculation
    offset=0               # No time offset, meaning Zigzag points will be calculated from the current time
)

df2 = df2.join(zigzag_result)

zigzag_result2 = ta.zigzag(
    high=df2['High'],
    low=df2['Low'],
    close=df2['Close'],
    legs=3,
    deviation=0.1, #0.07
    retrace=True,
    last_extreme=True,
    offset=0
)

df2 = df2.join(zigzag_result2,rsuffix='_zigzag')
df2 = df2.iloc[:-1]

# Filter non-NaN values in ZIGZAGv_0.01%_3 for plotting
valid_zigzag_data = df2[df2['ZIGZAGv_0.01%_3'].notna()]

# Create the figure
fig = go.Figure()

# Adding Candlestick trace
fig.add_trace(go.Candlestick(
    x=df2['Datetime'],  # Assuming `Datetime` is set as index
    open=df2['Open'],
    high=df2['High'],
    low=df2['Low'],
    close=df2['Close'],
    name='Candlesticks'
))

# Adding Zigzag line trace with dots
fig.add_trace(go.Scatter(
    x=valid_zigzag_data['Datetime'],  # Non-NaN indices
    y=valid_zigzag_data['ZIGZAGv_0.01%_3'],  # Non-NaN Zigzag values
    mode='markers+lines',  # Both markers (dots) and lines
    line=dict(color='rgba(0, 0, 255, 0.9)', width=2.5),  # Customize the line color and width
    marker=dict(color='red', size=4),  # Customize the dots color and size
    name='Wave Formation 0.01% DEV'
))

# Filter non-NaN values in ZIGZAGv_0.01%_3 for plotting
valid_zigzag_data2 = df2[df2['ZIGZAGv_0.1%_3'].notna()]
# Adding Zigzag line trace with dots
fig.add_trace(go.Scatter(
    x=valid_zigzag_data2['Datetime'],  # Non-NaN indices
    y=valid_zigzag_data2['ZIGZAGv_0.1%_3'],  # Non-NaN Zigzag values
    mode='markers+lines',  # Both markers (dots) and lines
    line=dict(color='yellow', width=2.5),  # Customize the line color and width
    marker=dict(color='red', size=4),  # Customize the dots color and size
    name='Wave Formation 0.1% DEV'
))


# Set up layout
fig.update_layout(
    title="Price with Higher High and Lower Low Patterns",
    xaxis_title="Date",
    xaxis=dict(
        rangeslider=dict(
            visible=True  # Set to True to make the range slider visible
        )),
    yaxis=dict(
        autorange=True,  # Enable dynamic Y-axis range adjustment
        fixedrange=False  # Allow zooming on the y-axis
    ),
    yaxis_title="Price",
    template="plotly_dark",
    showlegend=True,
    legend=dict(
        orientation='h',  # Horizontal legend
        yanchor='top',
        y=1.5,  # Position legend slightly above the plot
        xanchor='left',
        x=0.5,  # Center the legend horizontally
        font=dict(
            color="grey"  # Change the legend text color here (e.g., 'white', 'black', or any valid CSS color)
        )
    )
)

# Show plot
fig.show()

# Optionally, save the figure to an HTML file
pio.write_html(fig, file='Wave Pattern Formation 5d.html', auto_open=True)

Expected behavior
As Data is coming on 1min Basis, sometimes zigzag are plotted correctly as below.

Screenshots
While Rest of Time Zigzag is not plotted correctly, like below.

[Ish2K]

Could you share the dataset which results in inconsistent results? Perhaps you can make a temporary csv file to store results. My intuition is that maybe there are null values reported in dataset by yfinance. Did you try using other tickers? If yes, did you face occasional inconsistent results there too?

[divyankm]

DataSet ES=F:
df2_output.csv

Output ScreenShot (Yellow Wave: Deviation: 0.1%):

My intuition is that maybe there are null values reported in dataset by yfinance.
No Null Values in OHLC Data, on df2_output.csv file

Datetime	0
Adj Close	0
Close	0
High	0
Low	0
Open	0
Volume	0
ZIGZAGs_0.01%_3	3460
ZIGZAGv_0.01%_3	3460
ZIGZAGd_0.01%_3	3460
ZIGZAGs_0.1%_3	5007
ZIGZAGv_0.1%_3	5007
ZIGZAGd_0.1%_3	5007

Current ScreenShot of ZigZag Waves on 0.1% Deviation for ES=F:

Did you try using other tickers? If yes, did you face occasional inconsistent results there too?

Yes, I check on Other Ticker: YM=F (Dow Futures) on 1 min interval, Same issue coming there.
Dow Futures DataSet Attached as well.

ds=f_output.csv

[twopirllc]

Hopefully @aligheshlaghi97 has some insight on this or how to fix perhaps? 🤷🏼‍♂️

[aligheshlaghi97]

Hi @twopirllc
Thanks for mentioning me! I’ll take a look and try to address it during the next week.