"We are on this board with the goal of coming up with a useful prediction methodology." - Some people, maybe.

Odds of 7.5% aren't very good, agreed, but if you fire off twenty predictions at a time on a regular basis you will soon have enough "hits" to claim that you have a "record".

Knowing the odds certainly isn't useful for making accurate predictions, but it IS useful for evaluating pseudo-predictions based on the simple rule "when an earthquake occurs, predict another one" as suggested by Mulargia some years ago. It's analagous to the weather forecasting rule, "observe the weather today and predict the same for tomorrow" - which gives quite a good success rate, but is nevertheless trivial and not really very useful.