You are correct that no one is doing good short-term prediction at this time. Like a drug,
it is necessary to do a long-term study of each forecast system before any hope can
be raised that such and such a system can accurately forecast/predict earthquakes.
There are many forecast systems which accurately forecast earthquakes but under
analysis are no better than random, just as there are many drugs which look like they
cure some disease, but which under closer analysis are no better than a placebo or
do significant harm to other parts of the body while curing the disease. The reason for
archived forecasts is so that such long-term analysis can be done on apparently
successful techniques. Careful and long-term research is often needed to prove or
disprove theories and claims.

Now I put it to you and other board members. Which is more useful - a short-term prediction
for a Magnitude 5 earthquake in an area where such events occur on average about
once a week, or a 2-month forecast for a magnitude 7 where events occur on average
every 25 years. In considering this, you should realize that both forecasts are successful.
The question is which is more likely to be able to prepare people for a disaster?