I play low-limit Hold 'Em pretty much exclusively. I have a lot of thoughts about it. For one thing, KoreyKruse is right and Arimathea is critically wrong above (though her writeup is good). First of all, at a typical low-limit table, you can never see the flop using Arimathea's philosophy since even a pair of A's has much less than a 50% chance of winning (I've written computer simulations to analyze it). Second of all, if you use such a conservative philosophy after the flop, you're not giving yourself an opportunity to make up for your pre-flop bids. In poker, you lose money by being excessively conservative. As KoreyKruse points out, "expected pot odds" vs. odds of winning determine the correctness of every Hold 'Em decision (with the exception that possibly you might want to fool other players by pseudorandomly playing poorer-than-acceptable hands, making them more likely to stay for your better hands--my feeling is that generally in low-limit games the players aren't clever enough to worry about that).

But even though KoreyKruse is right, correct play is still extraordinary difficult. For one thing one must accurately predict the "expected pot" (the amount of money that will be handed to you in the event you win). Odds of winning are extremely hard to analyze as well. Suppose the flop just occurred and there are 5 players left. You can write a computer program to determine your odds of winning in the event everyone stays in to the showdown. But what if, for example, you have an inside straight draw on the turn and you fold it (often correctly--only 1 in 11.5 chance of getting the straight, high likelihood of split pot, chances for raises on the turn to make it not worth your while, etc.), but you would have made the straight on the river? An accurate computer model would have to take into account your play on the turn to determine your odds of winning after the flop! And what about the reality that other people also may fold their winning hands? Not the easiest program to write!

And even more importantly, unless you have good position (Arimathea is right--position is huge) you cannot know whether your call will be followed by a raise. While it is often worth it for you to see the river card for the minimum bid on the turn (say $20 in $10-$20), raises to $40 or (gasp) $80 could very well ruin the math for you. You say to yourself--I am going to lose this $20 X% of the time, but (100-X)% of the time I'll get it back and the expected pot--so I'll come out ahead over the long run by Y amount. Then raises make it so you will lose major money in the long run. No computer analysis or probability calculations are going to save your ass here. So...

You have to read the players! That's what makes poker so fascinating. Sure, knowledge of your odds of winning is critical. But if you can watch people's behavior and determine, for instance, whether they intend to raise on the turn, you have a massive advantage.

The neatest thing about Texas Hold 'Em is that it is nearly impossible to "learn to play" simply from experience, in my opinion. Since every hand is based on luck, you'd often have to play the exact same hand 1,000 times (and be able to remember that fact and the results of each hand) to be able to analyze your proper move based on experience. So there is TONS of room for innovation and ingenuity...