bugdget

Keeping a budget is hard. We use a model-based approach to forecast what things you will be able to buy in the future.

Helpful resources

• https://en.wikipedia.org/wiki/Arithmetic_underflow

Solution

This challenge presents a user with a budget planning application. The application asks the user to enter their total income and expenses over a handful of months. We then compute the net savings (income minus expenses) for each month entered.

The application then switches into "forecast" mode -- we identify the trend in monthly savings by fitting a linear regression, then forecast the bank account's balance into the future.

The key is to manufacture a negative-sloped savings line, which will underflow the bank account balance to +MAX_INT. The underflow occurs on the following lines:

expected_savings = np.uint32(last_known_bank_account_balance) + np.uint32(slope) * count_months_to_predict

Note that your bank account balance must be positive before entering forecasting mode, or else the program quits.

See exploit.py