Profiling Opponents in the Absence of Statistical Convergence Pt. IV

In my previous three columns, we noted how we must constantly make playing decisions in poker without the benefit of statistical convergence, either because we simply do not have enough trials with a given opponent for any statistic to approach convergence (for example, we are facing players new to us) or because a given situation occurs so infrequently that a given statistic will essentially never approach convergence (for example, flush board check-raise stats).

At this point, we have three reasonable choices:

Default to basic strategy. With little or no information on an opponent, you should default to basic strategy, which involves making the best play – on average – against the universe of opponents.

Play the percentage directly. Take the statistic at face value and play the percentage as if it were statistically significant, or – and more to the point…

Profile the player. Use the statistic in the context of other statistics, the way the player has played other hands, your history with the opponent, and other profiling factors in order to improve the accuracy of your decision on a case-by-case basis.

We discussed the first two options – defaulting to basic strategy and playing the percentage directly – in the last two columns. Now let’s discuss the third option.

Profiling the Player

When faced with a decision on a player, our best and final option is to profile the player. What we want to know is, is the player capable of making the play? Is he the type of player who would make such a play? Moreover, how likely is he to make the play? Does he do it often enough to materially affect our playing decision (meaning, deviate from basic strategy)?

In live play, poker players rarely get the benefit of anything resembling statistical convergence. But for as long as poker has been played, poker players have been able to answer these questions and make accurate judgments about their opponents with far less statistical information than is available to online poker players today, and based on far fewer trials.

How is this so? The answer is context.

Generally speaking, you don’t want to look at any one statistic in a vacuum, because there is always more information available. Rather, you need to look at a statistic or action within the context of a human being. That is, you want to look at a player’s action in the context of:

The player’s actions in previous hands.

The player’s VP$IP/PFR (voluntarily put $ in pot/preflop raise) profile.

Other statistics related to the specific action.

Inter-player history, your image, and game dynamics.

The Player’s Actions in Previous Hands

The first clue that a player is capable of making a play is if you’ve seen him do it before. If in your first ten hands at the table, a player three-bets preflop and shows down 7-6 suited at the river, then you know he is capable of three-betting light – and you can know this long before you get a 1,000-hand sample and an accompanying three-bet percentage figure. Alternatively, if a player check-raises on a flush board flop, gets called and shuts down, then you know he is capable of check-raise bluffing on flush boards.

Certain plays may also convey other information about the player’s profile.
The free card play in limit hold’em is a good example. I remember when I started playing poker back in 2003, when no-limit hold’em cash games were virtually non-existent, and the vast majority of the games being played were limit hold’em. Every once in a while, the flop would come something like 9-8-2 rainbow. There would be a bet and raise; the turn would be a blank, and the turn would get checked around.

Any time you saw that play, you could tell a couple of things. One is that the player who put the raise in on the flop and checked back the turn probably held J-10 for an open-ended straight draw; and the other was that he probably read at least one of a number of books on limit hold’em, most likely Hold’em Poker for Advanced Players by David Sklansky and Mason Malmuth, or Winning Low Limit Hold’em by Lee Jones, if not Hold’em Excellence by Lou Krieger (which was the first book I read).

Consequently, you could also figure that the player was probably tight-aggressive, and that you should respect his under-the-gun raises preflop.

It doesn’t necessarily mean that the player is good, or that he even understood the concepts presented in these books, or even that every detail in these books is unequivocally correct. As it is, a lot of these players would raise the flop and check back the turn at times when they should continue betting. But this action reveals information about a player’s influences, which in turn yields information about how the player plays the game.

Light three-betting preflop in no-limit or PLO is another excellent example of a tell-tale play. If you see a player three-bet light preflop – or three-bet more than once in a few rounds – there’s a good chance that the player has watched online poker training videos, participates in or at least browses online poker discussion forums like Two Plus Two, and/or has read some of the more recent poker strategy books on the respective games.

Again, this doesn’t necessarily mean the player is good, or that he knows what he is doing. But it does mean that he is probably a regular, and that he is at least a moderately educated poker player. It also means there is a good possibility that he is capable of some of the plays that are presented in online poker training videos, forums, and books.

If the player is the type to three-bet light, he also might be the type to check-raise bluff, or float, or value bet thin, or donk bet (betting into a player who has the betting initiative) light, or all of the above.

The Player’s VP$IP/PFR Profile

Is the player loose or tight, passive or aggressive? More specifically, is he a TAG (tight-aggressive), a nit, or a LAG (loose-aggressive) player? And if we can classify the player as a nit, is a nit capable of three-betting light preflop, or bluff check-raising?

In Small Stakes No-Limit Hold’em, Miller, Mehta, and Flynn classify players in no-limit six-max play in the following categories (plus a few more categories) based on the player’s VP$IP/PFR:

Category VP$IP/PFR
Nit 13/9
Nitty TAG 17/13
Solid TAG 21/17
Loose TAG 24/20
Thinking LAG 29/24
Crazy LAG 53/39*

Source: Small Stakes No-Limit Hold’em by Ed Miller, Sunny Mehta, and Matt Flynn

Let’s say that over 50 hands, a player has a VP$IP/PFR of 16/12 and appears to be nitty TAGish, but has three-bet four times in 20 opportunities for a 20 percent three-bet percentage. How likely is it that the 20 percent three-bet figure is an aberration? Is it more likely that the three-bet percentage figure is higher than we can expect this opponent to actually three-bet over the long run, or is it more likely that the VP$IP/PFR figures are lower than the opponent actually plays?

In this case, I would take the VP$IP/PFR figures to suggest that the three-bet percentage in our sample is probably higher than the player’s actual three-bet percentage, and more likely to be under 18 percent than over. A big part of the reason I trust the VP$IP/PFR figures more is because they converge faster than the three-bet percentage figure, and thus are more likely to be a more accurate reflection of the player. Consequently, if I were holding 6-5 suited in the cutoff facing a three-bet from the button sitting 100 big blinds deep, I would more likely fold rather than four-bet.

Other Statistics Related to the Action

When deciding whether to three-bet bluff on a flush-board flop against a given opponent’s check-raise, the most directly applicable stat is his flush-board check-raise percentage. But instead of focusing solely on flush-board check-raise percentage, you might also – or instead – look at his check-raise percentage in general. If a player has a check-raise percentage in the high single digits, there is a strong possibility that he check-raises even more often on flush boards.

On the flip side, if you are the other player and are thinking about check-raising on a flush-board flop, you might consider the player’s general continuation bet (c-bet) and fold to check-raise stats, as well as the player’s flush-board c-bet percentage and fold to flush-board check-raise stats.

Similarly, if you are considering three-betting preflop against a given opponent, you might also look to group statistics. You would start with the player’s PFR percentage (is he loose and aggressive?), but also look at how the player responds to three-bets. Thus, you might take into account the player’s fold to three-bet percentage and fold to c-bet percentages, as well as his four-bet preflop percentage.

Inter-Player History, Your Image, and Game Dynamics

Existing history between you and your opponent could be another factor in evaluating your opponent’s play.

Let’s go back to the 6-5 suited hand. Again, it is $1-$2 no-limit hold’em six-max with 100BB stacks, you open with a raise to $7 in the cutoff with 6-5 suited, and the button three-bets to $22. Over 50 hands, he has a three-bet percentage of 20 percent. If in the current session, you’ve already four-bet him pre-flop twice, you might hold off on four-betting him with 6-5 suited this time, because he might be more apt to call you or five-bet shove you lighter.1

There are two key points here. One is that the profitability of your four-bet is dependent not only on his three-bet frequency, but also on the range of hands that he will continue with when you four-bet. If your opponent widens his four-bet calling/five-bet shoving range beyond tens plus-A-K, this will hurt the value of you four-betting with 6-5 suited, because there is essentially nothing he will add to his range that you can beat.

The second point is that your opponent is likely adjusting to your play the same way you are adjusting to his. Indeed, this game is not played in a vacuum.

Your own image, as well as general game dynamics, could also play a role in your opponent’s play. Your general table image is a factor whether or not you’ve yet to play a hand against this particular opponent, as the hands your opponent has seen you play could affect his judgment of your actions in the same way that your previous history does. General game dynamics might also be a factor in that your opponent might respond to differently if the game is wild than if the game is generally tight. ♠

1 Footnote:

Though on another level, some opponents might know that you know you look over-aggressive, and give you credit for a stronger hand when you four-bet again because they know if you are going to four-bet for a third time in 50 hands, you are probably not going to do it without a hand.

Card Player columnist Jeff Hwang is author of Pot-Limit Omaha Poker: The Big Play Strategy and the three-volume Advanced Pot-Limit Omaha series. Jeff is a consultant for the recently released PokerTracker 4.