The same thing happened to me when I first got started - you didn't mention which limits you were playing, i.e. .25/.50, .50/1.00 or higher?
I agree with Hank that there is no way you can expect to repeat play money results at real money tables. In my opinion, if you don't start winning right away at hold-em, you should read a book or two, and as Hank says, the Sklansky books are widely recognized as the best out there. I read Lee Jones' 'Winning Low Limit Hold-Em', and although I had to adapt some of the techniques he describes to the micro-limit tables, reading that book was the difference between losing and winning for me. Another thing you should know - in my experience, the variance (the amount by which your bankroll can swing up and down) is just crazy at the micro-limits (i.e. .02/.04 to .50/1.00). I have heard some poeple say that things settle down for good players once you hit the 3.00/6.00 level. Anyway, you can expect to have wild winning and losing streaks at the micro-limits.
If you are really keen on learning, spend some time at the 2+2 website - the chat boards there are full of people like us who are just starting out and the veterans will always answer your questions.
Re: the debate about whether a computer can emulate good poker play...I am not sure that the chess analogy holds up, because there is no element of bluff in chess. How could the computer tell if the betting patterns of the human player represented a bluff or not? Modelling based on past revealed bluffs? WOuldn't that be almost impossible to program? And what about the play of the computer - would bluff be programmed into it? How would it go about making the decision to bluff or not?
All you AI programmers out there can now step up and explain why I don't know what I'm talking about
