Here's is a practical example for Roger's program to deal with.
Back on March 13, Don in Hollister made a forecast for Hastings, New Zealand
The forecaster had expected the following earthquake:

F: 18MAR2002 +/- 2 days 40.0S 176.8E Ml 3.5-4.3 Hastings, North Island, N.Z. within 65 km

The following event was reported today by Red Puma - now don't complain, I realize this
is probably an incorrect location - but that is a problem when multiple locations are
given. Any forecaster could choose the location closest to his/her forecast and claim
a correct hit. The purpose of this is to see how Roger's program deals with a forecast
for an unusual event where the data on the event itself is inconsistent.

O: 17MAR2002 21:43:2 39.1S 176.0E mb=6.3 MAD NORTH ISLAND, NEW ZEALAND

If this were the correct epicenter - it certainly is near the location Don had
given (about 60 km to the east and 90 km north), it is within the time window,
and depending on the network you choose could even be close in magnitude.
Not only that, but it is a rare event.

Another epicentral locations given by Red Puma is:

17Mar2002 21:43:30.4 36.5S 178.3W MS=5.1 GSR EAST OF NORTH ISLAND, N.Z

Both are manual determinations. Of course, I would be tempted to use the GSR
location as it is probably pretty close (experience pays in matters like these).

How would Roger's evaluation program work on this event and forecast?

The forecast had actually stated:

Hastings, New Zealand Earthquake Forecast
Posted by Don In Hollister on March 13, 2002 at 01:27:37:

Magnitude: 3.0-4.3Ml
From: 03/16/2002UTC - To: 03/20/2002UTC
Location: Hasting, New Zealand
Lat: 40.0S - Long: 176.8E - Range: 65Km

Hi All. Looking for a quake at 40.0S/176.8E. This is about 35Km south of Hastings, New Zealand. Should occur on 03/18/2002. The window will open on 03/16/2002 and close on 03/20/2002.

The radius for this is 65Km. Take Care…Don in creepy town

• RED PUMA earthquake listing