Surprise is a variable you can model, but the variance is always higher than the mean. If you want to hit the sweet spot, calculate the probability of detection and offset it with a buffer. I’ll walk you through the math if you’re serious, but first let me know if you’re planning for a birthday or a tactical advantage.

Birthday or battlefield, the math is the same. First you estimate the enemy’s (or guest’s) baseline expectation E and the standard deviation σ of their reaction. Then you choose a target surprise level S. The optimal surprise is when S lies within one standard deviation above E so it’s noticeable but not predictable. I’ll lay out the equations if you want, or we can just talk tactics. Your move.