Betting versus check-raising

I like Ed Miller's columns in Card Player magazine. He is the writer who is most likely to have a specific idea that offers a relevant way to tweak my game, something concrete that I can implement the next time I play.

In the April 4, 2012, issue, he discusses a hand played by one of his students. In a $1-2 NLHE game, the student raised under the gun with Js-Jc and picked up three callers. On a flop of 9d-7d-4s the student checked, had an opponent bet, which folded the other two still in the hand. The student then check-raised. When asked about his play, the student said that his goal with the check-raise was to get his opponent to fold.

Miller spends most of the column discussing (appropriately) how wrong-headed this is. Not that he needs my two cents, but here's how I would have phrased the general idea:

I don't make money when my opponents make correct decisions. I make money only when they make mistakes. The most fundamental categories of mistake they will make are bad folds (when they have the best hand) and bad calls (when they don't). Of these two basic types of error, bad calls are far more lucrative overall. This is because of two factors. First, bad calls result in bigger pots than bad folds. Second, bad calls are a much more frequent error than bad folds, especially among low-stakes players such as those that populate Vegas's $1-2 games.

The consequence of these observations is that opponents' bad calls are the single most important key to making money in poker. Well, I'll qualify that. One could argue that avoiding making bad calls oneself--or, more generally, making fewer and less costly mistakes than one's opponents--is even more important. But in terms of what you might be able to manipulate opponents to do, inducing bad calls is the vein of gold.

One of the most frequent post-hand commentaries that bad players offer shows how poorly they grasp this concept. Hardly an hour of play goes by without somebody making a prohibitively large bet on the flop, folding the field, then showing an overpair with a comment like, "I didn't want anybody drawing at that flush/straight." If you ever want to show the poker world that you're unclear on the concept, be one of the guys who does that. What you should really want, of course, is for players to call too much money with inferior or speculative holdings, while simultaneously being able to deny them proper implied odds for the call by being able to get away from your hand on the occasions when they hit their draw, miracle two pair or trips, etc.

All of which is a long way of introducing what I thought was the most mysterious lacuna in Miller's column: the near-total lack of comment on his student's check on the flop. At the end, he does add this: "The irony of all this [i.e., his long criticism of his student's stated goal] is that I like his raise just fine. (I would have simply bet the flop, but given his check, I like the raise.) It's just his reasoning that was bunk. I like the raise specifically because lots of worse hands--tens, nines, diamonds, straight draws, and so forth--will call. My student raised thinking he would get folds, but in fact his bet is going to get action from all the right hands."

I don't disagree, but I would have put much more emphasis on what is passed over lightly with just the parenthetical remark. I would insist that betting out should be the default plan, with a check-raise an acceptable alternative only if the student can articulate very specific facts about his opponents that can justify it. Perhaps there's an aggressive player on the button who always bets when everybody checks to him. Maybe the student has been playing in a perfectly straightforward manner--check/folding flops when he has less than top pair, betting them only when he has better than that--and he has reason to think observant opponents have noticed this tendency and will try to steal from him. He may know one of the other players always bets draws rather than trying to take a free card. But there has to be something very specific about this particular confluence of opponents and circumstances to make going for a check-raise a viable alternative to betting.

I would give two reasons for this argument. First, as Mike Caro repeats ad infinitum, the obvious, straightforward action is usually the right one, and you need special circumstances to make deviating from it profitable. A continuation bet while holding an overpair to the board is clearly the most obvious, straightforward choice, the one least likely to trigger the fatal consequences of Fancy Play Syndrome.

Second, on a relatively draw-heavy board, giving a free card to three opponents would be a mortal sin, not forgivable with any subsequent act of atonement. One would need to be highly confident that a bet from one of them was going to be forthcoming before choosing to go for the check-raise. In this particular hand, without running any numbers, offhand I'd put the requisite level of confidence in an opponent betting at something like 80-90%. In order to have that high degree of confidence, one would need to be able to articulate specific facts about the situation based on previous observations and patterns.

So my only criticism of Miller's column is that I think he passes too lightly over the bet versus check-raise decision, as if it's kind of just a matter of taste or style. That may, in fact, not be his view of it. Maybe there just wasn't room to go into it while still airing out his main point about the flawed rationale his student gave. But I think that the check on the flop (assuming there were not adequate, specific justifications for it) it as big a problem as the warped reasoning about inducing folds.