hide answers

Dan Barlow:

I toss 6-7 in all three cases. I realize there are players who base this decision on the expectation that both players will get average hands for the rest of the game, but unless this game is to 1000 instead of 121, that's unlikely. Throwing away four or more points just because dealer might have thrown a pair of 8s makes no sense. He probably tossed a pair of Qs.

John Chambers:

In the first situation, you are already in average position (70) and you don't really need the points. However, in order for your opponent to get into position on this hand, your opponent will need to get to hole 86. Your opponent will need 34 points between the hand and crib. I would give the 6-7 to your opponent. If you get a cut, you can skunk your opponent. If your opponent gets a cut, it's not likely he'll get that close.

In the second position, I would do the same thing and try to go for the skunk, or you can chose to play defensively.

In the third situation, I would play it more defensively. Your opponent is only four holes from average position (70*). I would discard 6-J (of different suits). First, you need to break up the crib. Second, remember that odd cards make the crib. Third, I don't like to discard 7s into the crib unless I have no choice.

George Rasmussen:

With dealer at hole 54*, I would toss 6-7. Same with dealer at hole 60*. Hold the dozen (5-5-J-J) as dealer is well short of Third Street critical position zone (69-73). With dealer at hole 66*, I would play ultra- defense and discard the safest combination in this hand. Goodbye 6-J! Keeping the crib small in score is more important than the size of my hand with dealer at hole 66*.

Michael Schell:

Bear with me, since I'm going to take you through how I'd tackle this over the board. I'm +10 at the start of the deal in each of these scenarios. There's little chance of a skunk, even in the most favorable case: with dealer at 54*, he's +14 for foiling a skunk while I'm -16, so I'd pretty much need to cut a 5 cut to 5-5-J-J. Therefore, unless nothing less than a skunk helps me (say, in the last game of a qualifying round), I'm just going to eliminate them entirely from consideration.

Thus, at 70-54* dealer is starting -16, which compared to my +10 tells me that I should just make a normal balanced discard. 6-7 goes into the crib. At 60*, dealer is -10, the same as my +10, so again it's 6-7 to the crib.

At 66* we're closer to a classic yellow zone decision, albeit on on Third Street, rather than Fourth Street. In such a situation, the four-to-one rule is a useful guideline. As you may recall, this rule suggests that with me and dealer near a positional hole, if an aggressive keep improves my average score by four times as many points than it costs me in the crib and pegging, then it's worth doing, even in an ostensibly defensive position. Using the addition method, which adds a point for a double run, but subtract ½ point for breaking up all runs, I'll give 5-5-J-J a value of 11½ and 5-5-6-7 a value of 9. 5-5-6-J would normally get penalized down to 5½ for run breakup, but I'll make an arbitrary exception due to the two 5s and leave it at 6 (sorry, but I'm allowed to break my rules when experience weighs in favor of it). This gives me a quick estimate that keeping 5-5-6-7 will cost me 2½ points, and 5-5-6-J will cost me 5½ points. In reality the average hands are about 14.30, 12.22 and 9.46 respectively (depending on Nobs likelihood), which isn't off by too much give that what I'm gauging is relative rather than absolute value.

The next step is to factor in the points given up in my opponent's crib. With my discard averages, 6-7 and J-J both give up 6.4 points, while 7-J gives up 4.7 points. The two "blocked" 5s will mitigate the risk of any of these discards, so I'll consider that a wash. Clearly, there's little to be gained by keeping 5-5-6-7 instead of 5-5-J-J. But 5-5-6-J is a more compelling option, sacrificing an estimated 5½ points to save 1.7 points in the crib. 5-5-6-J is a slightly more dangerous pegging hand due to its 3/1 distribution, and that might be worth ¼ point or so, but even 1.4 or 1.5 is more than ¼ of 5½, so I would still lean toward the defensive 5-5-6-J keep, even with dealer four holes short. I suspect, based on these estimates, that the winning percentages are closer than you might expect either way.

To many it would be blasphemy to consider keeping 5-5-J-J in this board position, but I think that traditional concepts of board position have underestimated the risk of getting stuck in the yellow zone on Fourth Street, and facing dealer's pegging on the game's final hand.

Peter Setian:

In all three of the dealer's positions, I would break up the hand and toss 7-J. Even without help from the cut card, I will already be in a decent position when dealing the next hand. And if I play aggressive and get a cut for a 16‒20 point hand, I will still be short of the next "ideal" dealing position. So there's no reason to take a chance tossing 6-7 (or J-J). Exception: I'll toss 6-7 if I must get a skunk — such as having nine game points in Grass Roots play with this being the only game left.

REX:

Tsk tsk, except for Schell, none of the humans have considered defensive pegging. And even Schell thinks that 5-5-J-J is only a little safer in the pegging than 5-5-6-J, whereas I think that it's way safer. Enough so, in fact, to recover more than half of added crib exposure tossing 6-7 instead of 7-J:

Keep	Toss		Average  hand:		Opponent's   crib:		Pegging: Net (pone/dealer)
5-5-6-7	J-J		12.22		6.16		-2.70	(1.20/3.89)
5-5-6-J	7-J		9.46		4.21		-2.28	(1.30/3.59)
5-5-J-J	6-7		14.07		5.75		-1.52	(1.15/2.67)

As a result, 5-5-J-J is the right keep in all three scenarios. Note that three deals hence, my winning odds remain unchanged, but as dealer gets closer to position in four deals hence, dealer's odds increase and mine decrease:

Keep	Toss		% in three deals:			% in four deals:			Projected score after four deals:
			Loss	Win		Loss	Win		Me	Opponent
Score: 70-54*:									127	105
5-5-6-7	J-J		0.5	7.9		6.2	64.9
5-5-6-J	7-J		0.2	4.5		4.0	55.0
5-5-J-J	6-7		0.6	11.1		5.5	71.4
Score: 70-60*:									127	111
5-5-6-7	J-J		2.2	7.9		15.9	57.1
5-5-6-J	7-J		1.1	4.5		11.5	50.1
5-5-J-J	6-7		2.2	11.1		13.8	64.0
Score: 70-66*:									127	117
5-5-6-7	J-J		7.7	7.9		33.2	42.7
5-5-6-J	7-J		4.4	4.5		26.9	39.8
5-5-J-J	6-7		7.2	11.1		28.9	50.2

Click here for a guide to cribbage notation and symbols.

Panelists

Dan Barlow won the 1980 National Open Cribbage Tournament, and made the 1985 All American Cribbage Team. His cribbage strategy articles appeared in Cribbage World for many years, and can be seen on the ACC Web site. He also provides strategy tips at MSN Gaming Zone. He has written seven books on cribbage, two of which have been glowingly reviewed in Games Magazine. All, including his latest book Winning Cribbage Tips, are available at The Cribbage Bookstore.

John Chambers was one of the original founding members of the ACC. He is a Grand Master, winner of seven major tournaments, and author of Cribbage: A New Concept, He also directs three annual tournaments: the Ocean State Cribbage Classic, New England Peer Championship and Charity Cribbage Challenge.

George "Ras" Rasmussen is a Life Master - Two Stars, a four-time All-American, the national Grass Roots Division 1 champion in 2009, a former state champion in Virginia, Montana and Washington, and holds a Gold Award and a President's Award. He also directs the Washington State Championship, held each year in Centralia, WA. His articles on cribbage are available on the ACC Web site.

Michael Schell is a pioneer of modern cribbage theory, which synthesizes traditional concepts of expert play with new computer-informed insights and analysis. He has published Cribbage Forum since 2000. Schell holds a Bronze Award, is a Washington State Champion (2001), and was one of the principal architects of ACC Internet Cribbage.

Peter Setian has played cribbage for over 20 years, and has been a member of the ACC for about 14 years. During that time, he has won seven major tournaments and earned his Life Master rating. He plays in about eight tournaments per year, including the ACC Tournament of Champions and the annual Grand National. He enjoys participation in Grass Roots Club #72.

HALSCRIB is widely regarded as the world's strongest computer cribbage player. Its opinion was solicited using a special analysis version of the program. Since HALSCRIB only speaks binary, its thoughts have been translated into English by Michael Schell and its creator, Hal Mueller, a retired mathematics professor and eight-time ACC tournament winner. For more information, see the HALSCRIB home page.