Unfortunately, I cannot describe here the subject of our patent, I can describe here only the way to our discovery. We took the ANCC catalogue and according methodology of Per Bak et al. (2002) we divided seismic events into two groups – one group of random – independent - events and the second group of non-random events, which depends each other. We found that random events depend not on external trigger forces like tides, LOD (Length Of the Day) anomalies, other seismic events, air pressure, precipitation, etc.. On the other hand, known external forces trigger mostly the second group of non-random events. For example the big Sumatra (2004) and Bam, Iran (2003) earthquakes were trigger by tides. The first events of earthquake swarms in Western Bohemia were triggered by tides too (Fischer et al. 2006). On the other hand Puget Sound earthquakes are triggered mostly by Chandler period anomaly of LOD.

We stated that seismicity itself carries the information not about big earthquake preparation stage, but about the stress state of rock mass. The second group of non-random events comes bigger in the time, when stress in the massif comes to the strength limit of rock mass. It was confirmation of LURR theory (Load Unload Response Ratio) by Yin et al. (2004). We stated that seismicity is predictable because seismic response of rock mass depends on the difference between stress state and strength limit of rock mass. If such difference comes to zero, the group of non-random events increases up to the time, when nucleation phase starts. At that time the nucleation process is unstoppable.
So we are able to predict the time of future big earthquake. The length of recognition phase of big earthquakes is couple of months, but the length of nucleation phase is only couple of days. The place of future earthquake is shown by area, where the stress was close to the strength limit. Magnitudo depends on the area and the period of anomalous (non-random) seismicity.