Posterior Probability

If statistical models can track down spam, why not murders? Just input all the details of a crime, and ask a bayesian model for the guilty part. That had been your idea, hopefully a way to save lives. And it worked.

For a while.

Then came the random killings, spurious data to keep the model confused. Pretty soon more people were dying than ever before.

You don't go out much nowadays. You just stay in your office, recalculating endlessly the rising odds of your own violent demise.

.finis.