One major advantage to using dynamical, physics based climate models (run on large supercomputers) is that they are initialized with the observed conditions. Meaning if it is all over warm everywhere then the models should start from that point and then run forward using equations that describe the movement of our atmosphere/ocean.  Many statistical based models based on past history (e.g. analogs) are probably at a distinct disadvantage if the climate changes rapidly then they will not find past cases that are comparable.  With that said, scientists will be busy studying the current circumstances of this developing El Nino-- it does seem quite unique when compared to similar developing events.