Submissions are evaluated on accuracy. Once you submit your predictions, Kaggle will run the accuracy calculation and display your score. You can view how you rank in the class in the Leaderboard tab. 

## Submission Guidelines

For this prediction task, you have to create a submission file for the Kaggle competition. It should follow the format of `submission.csv` which has two columns: ID and SHOT_MADE.

### Submission File
For each ID in the test set, you must predict whether or not the shot went in (binary SHOT_MADE variable). The file should contain a header and have the following format:

```
ID,SHOT_MADE
3036018,0
1132843,0
567965,0
etc.
```

Refer to the following code snippet: 

```{python}
submission_df = pd.read_csv("./submission.csv", index_col=0)
submission_df['SHOT_MADE'] = # your predictions
submission_df.to_csv("./submission.csv", index=True, index_label="ID")
```

IMPORTANT: Make sure that the IDs of your predictions and the IDs on the submissions match!


## Grading

The leaderboard will rank participants based on their model performance (accuracy) on a held-out test set. Part of grade depends on your accuracy score and ranking - to get full credit, beat the baseline accuracy score. You will receive bonus points accordingly if you rank in the top 10. The other part of your grade will depend on the accompanying discussion you put in the notebook - your model selection process, hyper-parameter tuning and feature engineering (if any). 

Good luck, and may the best analysis win!