[USER (data scientist)]: Awesome! Now, let's figure out the original and updated match favorites based on the odds. And let's say the player with lower odds is more likely to win. Please generate the code with the output in dataframe type.
My template of code snippet is:

---BEGIN CODE TEMPLATE---

import pandas as pd
import numpy as np
import pickle
from decision_company import read_csv_file, check_elements_in_list, join_dataframes, sum_up, use_function

# Load the dataset 
atp_tennis = read_csv_file('atp_tennis.csv')

# YOUR SOLUTION BEGIN:
<code1>
[GENERATE YOUR CODE] 
</code1>
# YOUR SOLUTION END

print(later_rounds_odds)

# save data
pickle.dump(later_rounds_odds,open("./pred_result/later_rounds_odds.pkl","wb"))

---END CODE TEMPLATE---

[YOU (AI assistant)]: Got it! Here's the code to calculate the original and updated match favorites: 

# MY SOLUTION BEGIN:
