hide answers

Dan Barlow:

I'll feel pretty silly if dealer pegs only one hole, and I didn't save six points. I think I'll save 2-5-Q-Q and lead the deuce. Dealer plays an 8, and I score 15-2. He plays a 4 for 19, hoping to bait me into forming a run, but all I have are Qs. I play one, and he says go. Now he lays down a 3, I play my other Q, and he has a 7. I win!

²  ₈  ^5 (15-2)  ₄  ^Q (29-1)    ₃  ^Q  _7 (20-1)

I point out to him that if he had played his 7 on my 2, I probably would have played a Q, and he'd have won:

²  _7!  ^Q  ₄  ⁵  _3 (31-5)

He rudely fails to thank me for pointing this out.

John Chambers:

In this situation, with your opponent only two points from going out, you would keep the 2-5-5-Q. Why not the 6? Because you want a spread of cards so you don't get caught into letting your opponent peg. This is a pegging situation, and your opponent will try to keep four different cards with low cards. With this in mind — and the fact that you have two 5s (which means you opponent is not as likely to have a 5) — you lead the Q. And since he probably has low cards you would only play the 2 if absolutely necessary.

DeLynn Colvert:

I'll buck the trend and hold 5-5-Q-Q, planning to lead a Q. That way, if I survive the lead I'll then have three cards available that can't be paired.

George Rasmussen:

In this case, with dealer within two points of winning, I would choose the safest lead and let that determine the cards retained. Of course, as pone I'll keep six points and plan on counting out. Let's retain 2-5-5-Q and lead the deuce, forcing dealer to pair to score two points. On the second play, I recommend dropping a 5 if you can get the count at 15 or more. You've got to have some luck to keep the dealer to a single peg.

The other option is to play the dozen (5-5-Q-Q), lead a Q and use the 5 where possible to break up scoring prospects, hoping to give up only a single go. This is not as bad as it might seem as dealer is unlikely to hold ten-cards needing two points to win. And if he does have a ten-card, the odds favor him preferring a J. My holding two 5s cuts into the probabilities that dealer will have a 5. Still, I favor the first option. The downside is that you have two 5s — the easiest card in cribbage to trap. And such a play is often sought by the dealer. With two 5s you're providing dealer two opportunities to execute a trap.

Michael Schell:

At 115-119* my strategy will be to hold six points — or something extremely likely to produce six points — and try to limit dealer to a single peg. Hoping to peg a 6/1 in this position is asking a bit much. I'll have six sure points with either 5-5-Q or 5-Q-Q. The 2 is attractive since that gives me an ostensibly safer lead than a Q (three losers compared to four losers, a measurement I call objective risk). So the choice is between 2-5-5-Q and 2-5-Q-Q. I'm generally loathe to hold 5s in defensive pegging situations, because it's the one card dealer is most likely to be gunning for, and also because it's easy to get stuck leading one on the second play series. So my preference is 2-5-Q-Q, with only one 5. If I survive the 2 lead from that hand, I'll have just one "problem" card to worry about (the lone 5), and if I can dump it second, my remaining Q-Q has an excellent change of stealing last card, since dealer could be forced to play an extra card on an early go.

The only twist comes if I cut a 5 or Q, in which case a Q lead will have the same objective risk as the 2. There are two tiebreakers that favor a Q in that scenario. First, it would leave me with a variety of ranks among my remaining cards, perhaps even allowing me to snatch a crucial "go" in the first play series with my 2. Second, at this score most competent dealers will favor holding cards of rank A through 5, and thus my opponent is unlikely to be able to pair a Q lead. As I write this in early 2003, it's not known just how much this discarding preference affects the real-life odds of leading a Q as opposed to a 2. For now, my policy is to lead a ten-card over a low card if the two plays have equal objective risk, but to otherwise make the lead with the fewest absolute losers.

Phyllis Schmidt:

Keep 2-5-5-Q. Lead the 2 and try to avoid getting stuck with a 5 as your last card. Otherwise stay away from everything else.

Peter Setian:

I would keep 2-5-Q-Q and lead a Q. You have to keep six points (as opposed to two points), and keeping the two Qs seems to guard better against possible pegging trouble with a run (13 or 14 count). Leading the 2 from the same hand would be my second choice. But I prefer to lead a Q since I have seen two 5s, and since the dealer is not likely to have kept high cards in this position.

HALSCRIB:

I don't like my chances here, probably about 13% at the most. I start by eliminating the candidate hands that don't guarantee me at least six points after the cut. Then I rank the survivors by how few pegging points they give up on average:

Keep	Lead		Average dealer pegging
2-5-5-Q	Q		2.22
2-5-Q-Q	Q		1.76
5-5-6-Q	Q		2.02
5-6-Q-Q	Q		1.72

5-6-Q-Q wins by a hair. Next I'll calculate the pegging sequences from this hand against all the possible opposing dealer hands. I'll weigh the results based on the relative likelihood of the opposing hands, mindful that dealer will favor "pegging cards" at this score (unlike other bots, I have special hand frequency tables that I can call upon in endgame situations). The result of this brute force approach is that the obvious Q is the best choice.

The humans may reason that 2-5-5-Q or 2-5-Q-Q would be better keeps because either one affords the lead of the 2, with only three immediate losers. But this overlooks opponent's bias to keep a 2 rather than a 6 or Q in this position. Essentially the lead of the Q has only two likely losers: the remaining 5s. This makes it a tad safer than the 2. Once you accept this, the next question is whether the better rank distribution you'll have retaining a 2 outweighs its greater chance of eventually being paired. By a tiny margin I conclude that it does not. Thus I'm keeping the 6 with my 5-Q-Q and leading a Q.

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.

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.