Shorthanded Poker: Flop Play Part II
by Jason Pohl

In Part I, we set the foundation for a flop strategy against a single preflop raiser. With top pair or better, we were able to definitively demonstrate the superiority of a check-raise on the flop against almost all opponents. To most educated players, this should come as little surprise. David Sklansky wrote in Hold'em Poker:

"It is frequently correct to check raise if:

1. You think you have the best hand (though not a slowplaying hand) and
2. You are quite sure someone will bet behind you if you check."

But is check-raising also a viable strategy when semi-bluffing or holding a marginal pair? Heads-up after the flop, the big blind will confront times when he has a draw, a medium pair, or suspects the button missed the flop and cannot call a bet. Against all but the most inattentive players, you cannot simply bet out with weak (but playable) hands and check-raise with strong hands. Such a pattern is too noticeable. Check-calling is also a weak, losing approach, as many authors have proven repeatedly. The alternative is to play strong and weak hands similarly (with some exceptions). We know that check-raising is advantageous with very strong hands. Now we will examine how disadvantageous check-raising might be with marginal holdings.

Scenario 1. Drawing Hands

Example 1. The Normal Draw.
Big blind holds Jc Th. Button holds Ad Qc.
Flop is Qh 9d 4s.

It is probable the button will not fold and will not check the turn, no matter what card comes. It is also probable the button will raise or 3-bet the flop.

• Betting out costs the big blind 2 small bets on the flop after the button's raise.
• Check-raising costs the big blind 3 bets on the flop after the button's 3-bet.

Either way, the big blind is going to check-raise the turn if a K or 8 hits while check-calling otherwise. Therefore, the amount won/loss is the same on the turn and river.

• Betting out adds 4 small bets, for a total of 8.5 bets in the pot.
• Betting out wins an average of (.342 * 8.5) = 2.907 small bets/hand for a profit of .907 small bets.
• Check-raising adds 6 small bets, for a total of 10.5 bets in the pot.
• Check-raising wins an average of (.342 * 10.5) = 3.591 small bets/hand for a profit of only .591 small bets.

In other words, check-raising loses about 1/3rd a small bet against a legitimate hand when the big blind has only 8 outs. However, the news is not nearly so grim with better draws.

Example 1b. Draw with Overcard
Big blind holds Ah 3h. Button holds Kd Qc.
Flop is Qh 9d 4h.

In this case, the button will again raise on the flop, but the big blind will hold 12 outs (nine hearts and three Aces.)

• Betting out wins an average of (.459 * 8.5) = 3.902 small bets/hand (1.902 profit.)
• Check-raising wins an average of (.459 * 10.5) = 4.82 small bets/hand (1.82 profit.)

There is virtually no difference between check-raising and betting out with 12 outs. With 13+ outs, check-raising will actually become profitable over betting out.

Example 1c. The Big Draw
Big blind holds Jc Tc. Button holds Ad Qs.
Flop is Qh 9c 4c.

• Betting out wins an average of (.563 * 8.5) = 4.786 small bets/hand (2.786 profit.)
• Check-raising wins an average of (.563 * 10.5) = 6.615 small bets/hand (3.615 profit.)

In this case, the big blind wants more money to go into the pot, and should strongly consider re-raising on the flop. It should be noted that even if the button held the Ace or Queen of clubs (but not both), the big blind would still be a favorite with two cards to come.

We need to look at one final example of a drawing hand. In this case, the button is ahead but will likely not 3-bet the flop and may fold the turn or river. Due to all the uncertainty of how the button will react, analyzing this final example can be a bit complicated.

Example 1d. Drawing Hand Against Drawing Hand
Big blind holds Jc Th. Button holds Ad Ks.
Flop is 9d 8h 4s.

The situation above is practically a coin flip (49.4% vs 50.6%). Therefore, the number of bets on the flop is effectively irrelevant. The only pertinent question is how the big blind's actions on the flop affect the turn and river play. Simply put, if a check-raise on the flop is more likely to cause the button to lay down its hand on the turn, then the big blind should unquestionably check-raise the flop and bet again on the turn.

If an Ace or King does fall on the turn, then the button might raise and cost the big blind an extra big bet. Figuring in the odds of sucking out on the river (Queen or Seven), the true price would be .818 big bets. Meanwhile, when the button fails to improve on the turn (on the 39 cards that are not an Ace or King), the big blind may well win the whole pot. Based on earlier assumptions, that pot would include 5.25 big bets.

It is important to realize that the big blind's flop play does not make a difference if an Ace or King falls on the turn. Either way (check-raising or betting out), the big blind will have control of the hand. The big blind will lose the same amount if an Ace or King hits or if the button refuses to lay down against two blanks. Again, the key is the likelihood the button will drop its hand when a blank hits. **Note: The big blind is a significant underdog on the turn if a blank hits, so the big blind wants the button to lay down the AKo.** If the chance of a fold is increased even 1% by a check-raise, then a check-raise should be employed.

But what if the button raises the flop with overcards and takes control of the hand? It should be obvious that allowing the button to take a free card or bet the turn to check the river would be very bad for our hero, the big blind. In both cases, the button has increased its chances of improving and/or seeing a showdown. At showdown, the button wins EVERY single time the big blind does not improve (and some times when both hands make a pair). Therefore, if a check-raise on the flop is more likely to take control of the hand, it is again a superior play since it vastly increases the odds the big blind will win the money already in the pot. This "control" factor strongly supports the case to check-raise with drawing hands.

I cannot emphasize enough that the only time the check-raise is disadvantageous is when the button holds a legitimate, strong hand. Even then, the check-raise only costs a fraction of a small bet. Meanwhile, the check-raise increases the likelihood of winning pots without making a hand.

Scenario 2. Middle Pair
Holding middle pair in a heads-up confrontation is certainly tricky. It is beyond the scope of this article to suggest all the proper ways to handle this delicate situation, but we can begin to consider the impact of our two main options.

Example 2a. 5-out Middle Pair
Big blind holds Tc 9c. Button holds As 4s.
Flop is Ah 9s 6c.

This is a good example of how a middle pair can be dangerous. The big blind is behind, and the button will likely raise a bet or (maybe) even a check-raise on the flop. An argument could be made that this is one of the exceptional cases where a bet out is superior since most opponents would not raise the flop without an Ace.

Let's assume the button will raise or re-raise with top pair on the flop. Let's also assume that the big blind will call a raise on the flop to try for a suck-out on the turn.

• Betting out wins an average of (.232 * 8.5) = 1.972 small bets/hand (-.028 loss.)
• Check-raising wins an average of (.232 * 10.5) = 2.436 small bets/hand (-.564 loss.)

This is a situation where the check-raise is clearly an inferior play. The big blind is a significant underdog, with 5 outs for two pair or runner-runner clubs for the flush. It should also be noted that the calculations above assume all the outs are clean. If the button held As Ts or As 6s, then the big blind would be drawing to only two outs. In those cases, the loss is even worse.

Example 2b. Middle Pair with Counterfeit Outs
Big blind holds Tc9c. Button holds AsTs.
Flop is Ah 9s 3d.

• Betting out wins an average of (.136 * 8.5) = 1.156 small bets/hand (-.844 loss.)
• Check-raising wins an average of (.136 * 10.5) = 1.428 small bets/hand (-1.572 loss.)

Knowing that sometimes we will face a losing proposition with second pair or worse, should we still regularly check-raise with middle pair rather than bet out? The answer continues to lie in our assumption of the opponent's holdings. In each example above, the button has a legitimate hand, often with a significant piece of the flop. The real profit of a check-raise occurs when the button did not connect with the flop and will lay down against action from the big blind.

Example 2c. Middle Pair vs. No Pair
Big blind holds Tc9c. Button holds Ad Jd.
Flop is Qd 9s 3h.

The button holds a legitimate raising hand preflop, but failed to connect with the flop. The button has six outs, along with a backdoor flush and/or straight draw (but only the Ace appears to be clean from the button's perspective). If we assumed the button would fold to either a flop bet or flop check-raise, a check-raise is preferred because it earns an immediate additional small bet. Even if the button calls on the flop, they will probably not call past the flop unless a good turn card falls. In that case, the check-raise only earns an extra bet when a blank falls (any card except a diamond, Ace, Jack, or Ten--24 out of 45 cards.) Even if the button might also call with a King or Eight, there are still 18 blank cards for a profit of .4 small bets resulting from a check-raise.

In truth, the big blind would prefer if the button kept calling. With only six outs, every extra bet makes money for the big blind. Even if a non-diamond King or Eight fell, the button would not add any outs since a Jack would now give the big blind a straight. While the middle pair might be anxious to see a fold, a button calling station is giving away money.

Scenario 3: Button Misses Flop
The ultimate difference-maker in our bet out/check-raise debate is this scenario. When the button will not give further action after the flop, it is best to get money into the pot as quickly as possible, so that the payload is larger when the button does fold. This critical concept makes check-raising the superior play almost all the time when heads-up against a preflop raiser. Not only is there an enormous profit to be made from taking down the pot uncontested, but the reward is even higher when taking down the pot after a check-raise.

Example 3. Big blind holds Ac 5h. Button holds Kh 7h.
Flop is 8d 5d 2c.

Let's give the button some credit and assume they are prepared to fold, considering their apparent lack of outs.

• Betting out wins 4.5 small bets.
• Check-raising wins 5.5 small bets.

It is clear that this is the kind of scenario where the big blind is very concerned to make sure no free cards are given, so my advice is different if the button will check the flop. But our basic assumption is that the button will not check the flop, but will come out attacking every time. Unless the button will call a check-raise significantly more often than a bet out, the choice is clear. A check-raise earns an extra bet.

So, how often does the button miss the flop? Let's pick a hand with a high likelihood of hitting the flop (all percentages are inexact): JTs. JTs will flop:
2%: Straight or Flush
20%: Flush Draw or 8-out Straight Draw
33%: Jack or Ten
20%: 4-out Straight Draw

Calculating for redundant flops (e.g. pair with a draw, simultaneous flush and straight draws), JTs, the most prolific preflop holding, will still miss the flop completely well over 1/3rd of the time.

In other words, over one-third of the time, the JTs is likely to fold to either a bet out or a check-raise. During that one-third of the time, a check-raise earns an extra small bet. That's an average extra profit of over .333 small bets/hand, if our base strategy utilizes the check-raise.

Wrapping It All Up
I am not suggesting that check-raising on the flop is an infallible solution to our heads-up dilemma. After all, there are ways for the button to use its positional advantage to assign a penalty on every strategy, and that is no different against the player who check-raises instead of betting out. Many things could go wrong. The button might check the flop, slowplay for an extra bet on the turn, or use aggression to wrest back "control." For those reasons and many more, I cannot emphasize enough the need to mix up one's plays. The big blind must bet out sometimes, slowplay other times, and even check-call on occasion, all in the name of variation. But all things being equal, I believe it is apparent that check-raising is most often the 'correct' strategy against the average preflop raiser who automatically bets the flop.

Next article, we continue looking at flop play by considering bluffing and semi-bluffing with only a few outs. You can email me at Jason@PokerPages.com with any questions, comments, or ideas for future articles.

Until next month, good luck!