for the rapid, shallow breaths he was taking, the man in the gun sights didn’t move. His hands were held out stiffly in front of him as if they could stop bullets. I remember wondering for a moment if it would hurt to be shot, but another part of my mind jerked me back to the job I’d taken on. I could not miss a detail.
The woman appeared to relax and then she became silent again. Though some observers might have viewed this as a favorable indicator, I had to assess if her quiet pauses were used for a rallying of reason or a contemplation of murder. I noticed that she was not wearing shoes, but discarded the observation as irrelevant to my task. Details are snapshots, not portraits, and I had to quickly determine which bore on my prediction and which did not. The mess of papers on the floor near an overturned table, the phone knocked off the hook, a broken glass likely thrown when the argument was more innocent—all assessed and quickly discarded.
I then saw a detail of great significance, though it was just a quarter-inch movement. (In these predictions, the gross movements may get our attention, but they are rarely the ones that matter most.) The fraction of an inch her thumb traveled to rest on the hammer of the gun carried the woman further along the path to homicide than anything she had said or could have said. From this new place, she began an angry tirade. A moment later, she pulled the hammer of the pistol back, a not-so-subtle underscoring that earned her new credibility. Her words were chopped and spit across the room, and as her rage escalated, it might have seemed I had to hurry and complete the prediction. In fact, I had plenty of time. That’s because the best predictions use all the time available. When effective, the process is completed just behind the line that separates foresight and hindsight, the line between what might happen and what has just happened.
It’s like your high-stakes prediction about whether the driver of an advancing car will slow down enough to allow safe passage—a fantastically complex process, but it happens just in time. Though I didn’t know it that day, I was automatically applying and re-applying the single most important tool of any prediction: pre-incident indicators.
Pre-incident indicators are those detectable factors that occur before the outcome being predicted. Stepping on the first rung of a ladder is a significant pre-incident indicator to reaching the top; stepping on the sixth even more so. Since everything a person does is created twice—once in the mind and once in its execution—ideas and impulses are pre-incident indicators for action. The woman’s threats to kill revealed an idea that was one step toward the outcome; her introduction of the gun into the argument with her husband was another, as was its purchase some months earlier.
The woman was now backing away from her husband. To someone else, this may have looked like a retreat, but I intuitively knew it was the final pre-incident indicator before the pulling of the trigger. Because guns are not intimate weapons, her desire for some distance from the person she was about to shoot was the element that completed my prediction, and I quickly acted.
I backed quietly down the hall through the kitchen, by the burning and forgotten dinner, into the small bedroom where a young girl was napping. As I crossed the room to wake the child, I heard the gunshot that I had predicted just a moment before. I was startled, but not surprised. The silence that followed, however, did concern me.
My plan had been to take the child out of the house, but I abandoned that and told her to stay in bed. At two years old, she probably didn’t understand the seriousness of the situation, but I was ten, and knew all about these things.
▪ ▪ ▪
It wasn’t the first time I’d heard that gun go off in the house; my mother had accidentally fired it toward me a few months earlier, the