The latest model of the Antikythera Mechanism was published in Nature in November 2006. The evolution of models of the Mechanism has followed a long historical struggle to reconcile the evidence from the fragments with a coherent design based on a rational structure and function.
Price's classic model is the basis for all recent models of the Mechanism, though many of the details of this model have now been rejected. Wright introduced a new model in 2005. This incorporated some major changes to Price's model as well as a mechanical realization of Price's idea that there might have been a Planetarium at the front of the Mechanism. Though highly ingenious with its eight coaxial pointers, there is much debate about the plausibility of this system and the absence of evidence from the fragments themselves may mean that no consensus can ever be reached. Most of the back of Wright's model has now been superseded by the model of the Antikythera Mechanism Research Project (AMRP), though we have retained Price's idea of a 235-month Metonic calendar on the Upper Back Dial. In particular, the AMRP model rejected Wright's older Draconitic Month Dial, which was driven by a modification of Price's Differential gear.
The major changes in the AMRP model are a 223-month Saros eclipse prediction dial with an Exeligmos (Triple Saros) dial as a subsidiary for the Lower Back Dials and a complete reassessment of Price's Differential. We believe that it is an epicyclic system that works in an entirely different way and realizes Hipparchos' theory of the Moon's irregular motion with a wonderfully ingenious design. Taken together, these changes represent a major reappraisal of the gearing structure and suggest that there was a fundamental confusion in the models of Price and Wright between inputs and outputs in the epicyclic system at the back of the Mechanism. In addition the output of this system in our model is to the front not the back dials, as in the models of Price and Wright. One great benefit of our model is that the Mechanism no longer emerges as a case of "Hellenisitc bodging" but as an economical design with a beautiful harmony and coherence. In addition, all the extant gearing (except a single unknown gear) can now be understood to be entirely based on two great cycles of the solar system from Babylonian astronomy: the Metonic Cycle and the Saros Cycle.
We acknowledge our great debt to our predecessors - in particular the pioneering work of Price, Bromley and Wright, which included many brilliant ideas. But in making our model, we have also rejected much of the previous research. We believe that our model represents an altogether more persuasive account of the evidence than any that has gone before. We very much appreciate Wright's immediate acceptance of our ideas and their direct incorporation into the model that he has made. We cannot accept that these changes only involve a slight alteration - maybe it was easy to make the changes mechanically, but in conceptual terms our model is very different. It represents a fundamental re-think of the whole structure and function of the Mechanism.