The idea has some merit.

Back in my government days, my program had built-in ranges for all parameters. I used a 6 point system; year, month, day, time, location and magnitude. Each point had an allowable range (arbitrary) assigned and the prediction received a score from 1 to 6. The "value" of the prediction came from comparing it to a graph based on 1000 random predictions similarly scored.

You need a series of such predictions to establish anything. I used a statistical test called the Kolmogoroff Smirnoff test to tell if a person's record was significantly better than the random predictions.

No one was.....

Roger