Sample mean vs. Theoretical mean

The app takes a sample from a discrete uniform distribution (numbers 1 through 9) and calculates the sample's mean. You can select different sample sizes and draw new samples to observe how the sample mean behaves. The expected value of this distribution is 5.

Here's a bet: you win $20 if the sample's mean is less than 4. The catch: you get to pick the sample size.

Try drawing samples at different sizes and observe how the sample mean moves relative to the expected value.

Which sample size is most likely to win you the $20?