What is the disadvantage of using a discrete probability distribution for bid amounts?

• A. Loss of information due to narrowing it down to certain numbers.
• B. It increases computation time.
• C. It requires more data to calibrate.
• D. It produces a biased estimate of the mean.

Answer: (C)

Introducing a discrete distribution for bid amounts means you must assign a probability to each specific bid value you consider. The main downside is the data requirement: the more bid levels you include, the more probabilities you need to estimate. With a wide or finely gridded bid space, you quickly need a lot of observed bids to calibrate those probabilities reliably; otherwise the estimated distribution becomes noisy or biased and the Monte Carlo results lose credibility. A continuous or simpler parametric form often achieves a good fit with far fewer parameters, meaning less data is needed to calibrate. The other points aren’t inherent drawbacks in the same way: discretization isn’t guaranteed to raise computation time, and it doesn’t automatically bias the mean if estimated properly. The key issue is the data needed to reliably pin down the probabilities across many possible bid values.