I read an interesting interview not long ago with Emma Donoghue and two of her editors – one from the US and one from Canada. The considerations one has with trying to reconcile the comments of various people who don’t necessarily agree with one another is an intriguing thing. But the thing that struck me was when Emma discussed her third (UK) editor, and how she’ll assign less weight to a problem only one of the three mentions. In particular, she mentioned that one of her three editors had figured out who the murderer was early in Frog Music, and this was too high of a percentage.
Now, look. You’ve only got three editors, which sort of gives you a whopping two degrees of freedom in this statistical analysis you’re performing. Not something you really want to make decisions based on. But worse than that, you can’t go around shouting “Thirty-three percent!” when you’ve got exactly one sample falling into that category. One slight tip in the other direction and you’d think zero percent of your editors had guessed the murder, and where would that put you? What if you decided the ideal number of people who figured it out was ten percent? How would you ever know how close you were? This is why all writers need a solid education in mathematics and statistics.
Also, I had a similar problem myself not long ago, and I thought 33% was about right.