In a message dated 2/25/2004 1:40:34 PM Central Standard Time, jim_donohue@computer.org writes: Depends upon what you mean by anticipate. Servoes of the autopilot sort certainly can anticipate in tailoring their responses to the actual conditions. The capabillity of the auto pilot to handle such things as cyclic events i.e. waves...is limited only by the amount of money you want To spend. Servoes simply work in a different domain than humans...and in most of that domain humans are not even capable of playing. Oh oh, here we go. Sooo, if we are all ready to play. Jim, there is no marine AP on the market that uses anything like what can be called artificial intelligence. They use simple adaptive algorithms and to say that they learn anything beyond how far to turn the rudder to accomplish a given turn rate is to overstate the case. AND, if there was such an AP it still would have insufficient inputs to reach their "learning" potential. (flux gates and flux gate rate compasses simply won't do. Once a second GPS inputs won't do. Gyro and ring laser gyro's will do. accelerometers will do.) AND if they did have sufficient inputs they still could not "learn" a "chaotic state." Sadly waves, in water or otherwise,can not properly be modeled as "cyclic events." AND as if that were not enough, if they could do all the forgoing the boat's steering system, be it hydraulic or linear actuator,can not move the rudder rapidly and smoothly enough. We have an initial state problem here. In heavy weather, e.g., surfing conditions, you can see the out of phase wave, coming, you can estimate it's effect and you can start the rudder and throttl e inputs before it arrives. Otto can't. Having used AP10, AP20, WH, ComNav and Raymarine a lot in past few years IMHO they are all okay autopilots but nothing special. They'll do a good job most of the time. None of them steer a particularly straight course in heavy weather. Observe the GPS track, look back and see the snake. None of them can steer better than an adequate helms person. All of them can steer longer than a human. Not sure what to make of the "servoes" comment. Is this a reference to AP servo motors? Most don't use them. Is it a software reference. If so it's not used in any of the neuro or baysian networks with which I am familiar. Always willing to learn chuck TOAD HALL