Code for Not feeling the buzz: Correction study of mispricing and inefficiency in online sportsbooks
FixedEffectModel==0.0.5
jupyter==1.0.0
linearmodels==6.0
numpy==2.0.0
pandas==2.2.2
requests==2.32.3
scikit-learn==1.5.1
seaborn==0.13.2
tqdm==4.66.4
openpyxl==3.1.5
To install a new environment with the required packages, run the following commands:
- Create a new environment using conda
(assuming miniconda installed on machine: https://docs.anaconda.com/miniconda/ )
conda create --name conda-python-buzz-env python
- Activate the environment
conda activate conda-python-buzz-env
- Install the required packages
Using the requirements.txt file:
pip install -r requirements.txt
Or using the install.sh script:
install.sh
Files are ordered as the replication study reads.
-
dataprocess_ramirez.ipynb ⟶ ramirez_matches_cleaned.csv:
Process dataset for use in replication and correction
-
dataprocess_clegg.ipynb ⟶ clegg_matches_cleaned.csv:
Process dataset for use in extended dataset.
-
replication_and_correction.ipynb:
Replication: Obtain model parameters and results to verify RRS.
Correction: Identifies, removes, and explores the effects of the problematic Hercog bet.
Results in Tables 1-3 and Figure 1.
-
extension.ipynb:
Conducts new mispricing and inefficiency test using extended dataset.
Results in Table 4 and Figure 2.
-
p_bs.ipynb:
Conducts a
$p_{bs}$ simulation as outlined in Wunderlich and Memmert 2020.Results in Appendix C.
-
ramirez_matches_cleaned.csv:
As processed by dataprocess_ramirez.ipynb.
-
clegg_matches_cleaned.csv:
As processed by dataprocess_clegg.ipynb.