Well, I'll try to add my $0.02 to these comments. First I'll tackle the limit vs. no-limit and tournament vs. cash game discussion.

I play tournaments when I judge they are positive EV for me compared to the amount I can win in cash games in the same time investment. I also play "no-limit" in live games wherever possible, if the money is deep compared to the blinds (I'll explain this in detail).

Harrington's book excellent, and it is the best thing out there for no-limit. The whole first volume and first chapter of the second volume apply just as well to cash games as tournaments. It's a great book, and I have no doubt that you are beating the small NL games after studying that book. It is an EV focused book and everything I said in my first post applies to this as well.

These days, NL is almost always played with a "capped buy-in", which technically means it isn't NL at all. It's a kind of spread-limit where you can bet anywhere from the big blind amount up to the maximum buy-in. Notice that tournaments are *identical* in this respect, you never have the option of buying in for ten times as many chips as everyone else if you wanted to, which you *can* in a true NL game. The only true NL Holdem game I am aware of is spread at the Wynn in Las Vegas -- everything else is this crippled capped buy-in version.

There is an important reason for this, as well as some important implications for an expert player. The reason why capped buy-in games predominate is because "true" no-limit (with no buy-in caps) allows an expert player to have too much edge over average or even good competition. Prior to the recent surge in popularity of NL Holdem caused by the TV coverage of the WSOP, no-limit holdem was not played, except for in that tournament main event. All the amateur players had been busted and no-limit games were no longer spread. It is widely believed among experts that no-limit is bad for poker, because it throws the balance of luck vs. skill too much in favor of skill; the recreational players don't stand a chance when the "money is very deep".

True no-limit poker is a game of implied odds. To use your example of overbetting the pot to force out flush draws, in a real NL game with deep stacks, I can call this overbet on the flop, provided that you still have lots of money in front of you. Now if I catch my draw, you'll make another pot-sized bet and I can set you all-in, or I can wait until the river to do this. The odds on the flop don't mean nearly as much when the money is deep. How do you like your set when you get check-raised all-in on the river when the 3rd flush card comes? A tough NL player can do this as a bluff, too.

With capped buy-in games, you NEVER have to deal with these difficult and highly skillful decisions. The money simply isn't deep enough. They way the game is played today, the so-called NL games are usually played with two betting rounds. Often all the money is in preflop. What doesn't go in preflop goes in on the flop. You should probably be playing that way in a capped buy-in game, too, in most cases. Even if there is some money left by the turn, usually it isn't much and the pot-size makes a call pretty much automatic with any kind of hand. What this does is diminish the advantage of the expert for three reasons: he can't use his superior hand reading ability for profit on later streets, he only has two betting rounds (in most big hands) to profit from his better judgement, and it diminishes the importance of position (once you're all-in position is irrelevent). Basically, capping the buy-in puts more "gamble" in the game. When the money is really deep, though, and the expert can make a huge bet on the river because there is still a lot of money left, then it's a whole different ball game, to say the least.

As a matter of fact, there is a strategy for NL games called the "Short-Stack Strategy" that relies on getting all the money in preflop or on the flop. It is extremely effective and profitable, and can be learned in about 15 minutes. Ed Miller gives the outline of this strategy in his second book, Getting Started in Holdem, and there is an excellent detailed analysis of it on the 2+2 forums. Basically, it relies on a guranteed preflop +EV situation, and totally eliminating implied odds of your opponents.

Online tournaments are always short-money affairs, except for a few major large buy-in events which may start out a bit like a deep money game. The Sit-N-Go's are particularly short stacked. Because of this, it is extremely easy to protect your hand with a pot overbet (because there are no implied odds, nobody can profitably call short), and in my opinion, it becomes a very simple game to play. Ferguson and Harrington have both said that the #1 tournament poker skill is knowing the end-game strategy for very short-money preflop all-in moves. Harrington explains how to develop that strategy in his section on what he calls "Structured Hand Analysis" in Volume 2. The shorter the money, the more quickly it becomes necessary to make these "low m" decisions. I was working on a multi-dimensional parametric model of that, but I haven't spent time on it recently since I don't play many tournaments anymore. In my opinion, there is way more money to be made in the cash games, and with much less variance, meaning a smaller bankroll is required.

Because of these issues with NL poker, capped buy-in or not, the long-term future of Holdem is limit. Limit is a vastly more complex game (compared to capped buy-in NL), for reasons you suggest -- you can't easily protect your hand with an overbet. You've got to do some complicated things to protect your hand in limit poker. "Real" NL Holdem probably is the "Cadillac of Poker", and is definitely a very, very complex game. But that is NOT the "NL Holdem" that people play today, and the difference is enormous.

All poker is about playing the players by the time you get to an intermediate level of play. That is definitely not only the case with the NL games. In fact, because of the way the NL is played today, and how easy it is to get an edge, it could be argued that adjusting for opponents is more important in limit holdem, because you are sure to have 4 betting rounds to employ this information.

Regarding the legality of online poker, I think it is very much a grey area. It doesn't really matter for taxes, though. You are still required to pay taxes on illegal income in the US. Most pros I know declare the winnings as "other income". There are accountants who specialize in gambling taxation.

Many people are concerned about poker-playing computer programs before they become serious online players. There are a few poker playing "bots" out there, and they can win in small-stakes games for a small amount. I'm can't be certain, but I believe I encountered one once. My experience was similar to what others have reported on 2+2: the bot is pitifully easy to play against, and relies on an extremely weak-tight "nut peddler" strategy. It's really easy to stay out of the way of this kind of thing. I've read game theory papers about certain poker situations and the general consensus is that a really tough bot that can play well in more than just heads-up (2 player) situations is beyond the existing mathematical understanding of the game. The game really IS that complex.

However, let's consider: what would happen if someone programmed an awesome, world-class, unbeatable bot? It would be no big deal. To understand why you need to understand how you make money in a poker game. You make money from other players making -EV decisions. The sum of everyone's expectation must be zero. If we have a +EV situation, it's because someone else has a -EV situation. This is referred to as the "Fundamental Theorem of Poker", a term coined by David Sklansky. All it means is this: we profit from the mistakes of others.

If all 9 opponents in a 10-handed game were these "killer bots", then sure, we couldn't win since they would play perfectly and make no -EV mistakes against us. If there were 1 killer bot in the game and 8 fish, then the game would remain hugely profitable. We wouldn't make money from the bot, but who cares? We're there to make money from the fish! Having a killer bot in the game would be NO DIFFERENT than having another expert player in the game. This is no big deal -- it's just close to zero EV for both of us.

It is often said that the most important poker skill is game selection. You want to play in games with poor players who are making lots of technical mistakes, and avoid games filled with players who don't make many mistakes. The bots don't change that at all.

Even though I believe bots are irrelevant to the game (they are weak, and wouldn't matter even if they weren't), the online sites see this as a PR concern, and take measures to detect automated playing. Most big cardrooms have something similar to those "bot proof" squiggly letter systems you see on some web sites. If your play matches a bot profile (by playing 24 hours nonstop, or other metrics) then you need to type in this machine unreadable letter sequence. There are actually lots of anti-cheat software precautions they use. They are much better able to detect and stop multiplayer collusion than a live cardroom is, for example.

As far as what sites I play on, most of my play is on Party Poker, but I also play on Bodog, Pacific, Eurobet, Interpoker, and Paradise Poker. They are all reputable. Party Poker is definitely the most popular and probably has the best software. Eurobet or Pokerroom have a client that works on Macs. The rest are Windoze only, unfortunately. I think all the big sites are reputable. Party Poker is a publicly traded company and has bigger fish to fry that stealing the money on deposit. Like Jeff said, you use an online bank account similar to Paypal. All the poker sites accept Neteller, which is kind of the standard. I use Neteller, Firepay, and Moneybookers, depending on the situation.

Jim