Schell's mailbag, January 2003
|January is here again and so is the mailbag.
As usual it features the most interesting reader questions and comments from
the past year. When you
write to me,
please remember to include your name and where you're writing from. If
you're asking how to play a specific hand, be sure to indicate the score and
who is dealing. I'll publish as many messages as I can, reserving the right
to edit for length and clarity.
Becoming an expert
Unlike some games, cribbage lacks a formal definition of expertise — there is no "expert" rating class per se, and ACC-awarded titles such as Bronze Award and Master recognize accomplishment rather than pure playing strength. As an arbitrary baseline, I'll consider someone an expert if they can beat a minimally-competent novice of the time and bat .540 or better in serious tournament competition (by which I mean bone fide events like the ACC Open, not some freebie online tourney). I'd guess there are about 500 players in the world that would qualify as experts under this standard. If you prefer a standard that is less dependent on the caliber of the competition, consider yourself an expert if you can average two errors or less in a nine-game session. It took me about two years of intensive study and practice to reach this level.
I spend an average of one hour per day working on my game. This includes sparing with HALSCRIB and Royal Cribbage, drilling myself on discard averages, analyzing problems, practicing endgames with Cribbage for Windows 97, reviewing my cribbage database, and inputting notes from the past week's Grass Roots session. It doesn't include time spent writing about cribbage, nor the time spent in actual competition, which over the course of a year amounts to roughly 36 Grass Roots tournaments, eight to ten sanctioned weekend tournaments and the occasional Internet match. That's a lot of time devoted to cribbage, but there are quite a few diehards — mostly retired people — who travel to sanctioned tournaments every weekend, play in Grass Roots every week, and spend their spare weekday evenings playing on the Internet.
If you're looking for one of the cribbage books I recommend, try clicking on the title when you see an underlined reference to it at Cribbage Forum. You'll be taken to a Web page with ordering information for the particular book. You might also look at the playing tips section of the ACC Web page, where there is a listing of mailing addresses and Web pages for books published in North America. At present, only Dan Barlow's Play Cribbage to Win enjoys widespread distribution in bookstores here, but most of the others can be ordered directly from their author.
At 117-118* or 118-118* you must accept that you are an underdog to win, since dealer pegs an average of 3¼ points under normal conditions. Occasionally you'll win a game by pegging out, but you should still generally plan to play desperation defense and hope to limit dealer to one or two pegs. With that in mind, your discarding priorities are roughly as follows:
In the first example you made the right discard. You must keep the run to ensure that you'll have enough points to win if you get to count your hand. That rules out 2-3-9-K, which looks attractive from a defensive standpoint, but leaves you vulnerable to a 7 or 8 cut. 2-3-4-6 ordinarily pegs better than 2-3-4-9, and 2-3-4-K is a tad safer if the opening lead isn't paired (since the K is less likely to get paired than the 9), but the 9 covers your 3 lead, and that's a higher priority in this position (your chances of winning go up dramatically if you survive dealer's first play). Once the 5 is cut, leading the 3 is clear. If dealer offers a 15-2 you should probably take it, since your remaining 4-9 gives you a good shot at scoring a winning go. Otherwise it's basically a matter of minimizing your losers. In the event, you were unlucky that dealer scored a 15-3 with a two-on-none at the end, but you still played the hand correctly.
In the second example you have very little chance of pegging out with the cards you were dealt, so I'd keep 3-3-K-K, which spaces your cards out a little better than 2-3-3-K. Additionally, once you've led the 3, you nothing but "safe" cards remaining in your hand, whereas with 2-3-3-K you'll always have that vulnerable 2 to worry about. Note that 3-3-K-K would have won against 4-5-Q-Q.
You have come a long way in five months! I'm glad you're playing cribbage and finding partners in Russia. There is also Internet cribbage, a nice way to meet people from other places. The better connected we are in the civilized world, the better our chances of staying one step ahead of the Neanderthals who are currently trying to destroy it.
Actually the overall average value of the crib is about 4.8 points. What's important to remember is that the 91 possible combos are not discarded with equal frequency. As pone you'll be tossing low-scoring dyads like 8-K, 9-Q and 10-K far more often than high-scoring ones like 2-3, 5-5 and 7-8. As dealer it's the opposite. (It seems I throw my opponents 10-K two or three times a night in Grass Roots play, but I doubt I make that same toss to my own crib more than a half-dozen times a season.) Thus pone's discards will cluster around the bottom end of the scale, while dealer's discards will favor the higher end. The crossing point between equally-matched players comes at 4.8 points. An expert player, though, will typically have an advantage of about ¼ point over an average player.
Once you understand this, it's easy to see why a dyad tossed to your opponent's crib scores more on average (typically by about a point) than the exact same dyad tossed to your own crib. If you throw your opponent 10-K, you take the risk that she'll toss one or two 5s in with it (dealer is equally likely to toss or retain a dealt 5). If you discard 10-K as dealer, you're quite unlikely to receive this kind of gift from your opponent, though she might still cut you a 5. Accordingly, 10-K, like any other dyad, will improve more in your opponent's crib, where is it more likely to combine with valuable cards.
Try to imagine a "typical" discard you'd make to your opponent's crib. Usually you'll endeavor to toss garbage like A-K or 9-K, but sometimes you have to make a dangerous toss like 6-8 or 5-K in order to keep your scoring cards together. A median toss is probably something along the lines of 2-J or J-K — not ideal, but not too dangerous either. Now consider the same question as dealer. You aim to avoid throwing yourself dross like A-K or 4-9, but you're not always in a position to discard juicy tidbits like 2-3 or 5-J. Probably the median toss is something like J-Q or Q-Q — not great, but still promising. Care to guess the average value of these four "typical" discards? Yep, 4.8 points!
Leading from magic fives
This play of leading the higher card of a magic five has got to be the oldest in the book — it always seems to be one of the first things novices learn when they start to improve their game. The main reason it's usually best to lead the 3 from 2-3 (or the 4 from A-4) is that it prevents dealer from safely playing an uncovered 5 or 6, thus making it more likely that he'll play a ten-card instead, which is exactly what you want. Yes, it also makes it easier to get the count over 15 if your lead is paired, but even if you're leading from something like 2-3-4-8, where you can't get the count above 15 if either your 2 or 3 lead is paired, it's still best to lead the 3.
Of course, nothing in cribbage is simple, and there are plenty of times where the lower card is the right lead. If you're playing defensively, and the lower card is paired in your hand or matched by the starter or one of your discards, then the lower card is probably safer to lead. And starting with a hand like 2-3-4-5, I generally prefer to lead the 2, since if it's paired I can safely dump the 5 next, whereas leading the 3 I'm screwed if it's paired. Dan Barlow even suggests you occasionally lead the 2 from a 2-3 combo just to keep your regular opponents guessing about what's in your hand!
Positional holes and opponent's style
You should definitely adjust the positional/par holes against opponents who habitually play aggressively or conservatively. Against an aggressive opponent you can expect to peg a bit more and score more in your crib than you otherwise would. It will also be less effective than usual to play defensively. Move the positional holes backwards one or two points (say to 16, 42, 69 and 95) to compensate for this. Conversely if your opponent habitually plays conservatively — never pairing leads if he can't retaliate, for example — then move the positional holes forward one or two points. You'll need better position early on to make up for the extra difficulty you have scoring against such a player down the stretch, and conversely he'll need better position if he's the sort who's unwilling to take chances as the game goes along. By the way, the 11-12% edge you have over your dad is exactly what I would expect between two otherwise equally-skilled players of which only one plays board strategy.
Recording live games
The annotated games published at Cribbage Forum have prompted a number of questions like the following:
Transcribing cribbage games between two humans is a challenge given the fast pace of the game and the lack of a tradition of published play. I'll occasionally bring a video camera to major tournaments to tape matches for later transcription. If shoot over the shoulder of one player, I can catch that player's discards, then reconstruct the other player's initial six cards from the crib composition (assuming both players follow ACC rules and expose the crib after each deal). Alternatively I'll just stand behind one of the players with a notepad and write down the cards as they are played. I don't bother recording individual scores, but I do note the score between deals from the board so that when I reconstruct the scoring later I'll have a checkpoint. It takes some practice to do this fast enough to keep up with tournament players. It helps to ignore suits except for flush hands and His Nobs. If I tell the players beforehand that I'm interested in recording the game for Cribbage Forum, then they're usually helpful about not burying cards too quickly after the deal.
The vast majority of the games I record are between two other players, but occasionally I'll write down one of my own if it seems interesting. Usually this isn't apparent until the game is a few deals old, so in those cases I have to rely on my memory to reconstruct what happened before I started notating. In general, though, when I take notes during play it's to record a specific problem for later analysis, perhaps a a tough positional decision, an unusual discarding problem or a pegging play that went against me. Basically I'm looking for things that might help me improve my game or make good instructional material at Cribbage Forum. Generally I only take a few such notes during any particular cribbage session, so this doesn't consume much time (I do most of the actual writing when my opponent is shuffling the cards). Occasionally someone does grumble about this (it seems to happen every year at Reno), but most of my regular opponents are accustomed to seeing my ubiquitous notepad and don't mind if I scribble in it from time to time.
Handling A-4-x-x as pone
Holding A-4-x-x, if dealer pairs your 4 lead you have a dilemma between dropping the A, which gives up a 15-2 to a 6, or playing a ten-card, which leads to disaster against hands like A-A-4-x, 2-3-4-4 and 2-3-4-5. Hessel's statistics on hand frequency (which may be slightly biased toward having dealer toss himself mid-cards) say that dealer will hold both a 4 and a 6 about 6.3% of the time, whereas he'll have A-A-4-x, 2-3-4-4 or 2-3-4-5 only 2.0% of the time. These numbers are static, and don't factor in how your own holdings affect the percentages, though I suspect the edge would go against A-A-4-x if you're already holding an A. It looks like dropping a ten-card has a slight mathematical edge, since dealer is much more likely to have a 6 than a hand that traps your lone A. Because of this, HALSCRIB 4.9 figures playing a ten-card is about ¼ point better in the long run. Nevertheless, playing the 4 now only gives up two extra points if dealer can exploit it, whereas having your A trapped high can give up as many as eight extra points. Ouch!
Personally I think the alternatives are close enough that the right choice comes down to which kind of hand you prefer to defend against in your particular situation. For example, at 92-86* with a 5 cut showing, I'd play a ten-card, since with dealer so far out of position I'm more worried about him having a 4-5-6 run combination than a ten-card and a couple of aces. If the score was 92-96* however, I'd probably figure there was little reason to defend against a 6, since this probably gives dealer excellent position regardless. I'd drop the A to guard against A-A-4-x, which doesn't combine so favorably with the 5 cut.
Welcome to the game Ron! There is no such thing as a double (or triple or quadruple) run in the pegging. The cards in runs must be played contiguously to be valid, so clearly a pair would break the sequence. If there is a local variant of the game where this is interpreted differently, then I have yet to encounter it.
As for cutting the cards before dealing, this is always done in tournament play, and it should also be done in casual play. The notion that cribbage is a "gentlemen's game" and that therefore the deck should not be cut before dealing is a bit of latter-day nonsense kept alive through its repetition in sophomoric reference books and instruction sheets accompanying cribbage boards. In fact, the deck has always been cut in cribbage prior to dealing, just as in any other card game, gentlemen's or otherwise.
I've never encountered a cribbage variant where the right J doesn't count in the crib, but given that I mainly play at clubs, tournaments and Internet sites, all of which use the standard international rules, there's a good chance that someone out there plays the way you describe. Ordinarily though, the only difference in scoring between hand and crib is that flushes count in the crib only if they match the suit of the starter.
To play triples (two teams of three players), deal four cards to yourself and your right-hand opponent, and five cards to the remaining four players, who each discard one card to the crib. Players from each team sit in alternating seats, so that each player is between two opponents. The player to dealer's right cuts the deck prior to the deal, while the player to dealer's left cuts for starter and makes the opening lead. Triple doubles (three teams of two players) plays the same way, except that players one and four (counting around the table) are on the same team, as are players two and five, and players three and six. For more information, have a look here.
Ask the experts again?
In January 2002 we featured the following Ask the experts problem (click here to see how the experts responded):
Pat Hayslett writes:
If dealer has A-A-9, the most he can have with a 10 cut showing is eight points (if his last card is 4 or 5). I don't think that's worth worrying about at this score, even with a four-point peg. Dropping the 6 saves you one point against A-A-9, but leaves you vulnerable to the scenario you should be most concerned with: that dealer has a high multiple run and can trap your second J on the next play series.
At 62 you are within ten holes of a positional hole that dealer has not
yet reached. That means you have a positional surplus, and dealer, who is
back at 52*, has a positional deficit. Reckoning to the next positional hole
(70) you are +2, while dealer is -18. Clearly dealer is not likely to make
up that large a deficit in one deal, even if you make a dangerous toss. On
the other hand, with only a two point positional surplus yourself, you don't
have much margin for error on offense, since a single bad hand could wreck
your standing from here on out. Therefore I would play this hand
aggressively and toss Q-Q, retaining 5-7-8-8.
If you run this hand through Cribbage Hand
Evaluator or DS.EXE,
you'll see that 5-7-8-8 returns the best average hand and
expected average by far. The alternative toss of 8-Q makes
sense if you need defense, since it gives up about 1½
points less in opponent's crib than Q-Q, and keeps a safer
defensive pegging hand (since the cards are better spaced). But it costs you
2 points of scoring potential
— not worth it at this score.
Supposedly this was a commercial game published in 1977 by a company in Sumner, WA. I've never seen it, and it appears to be long out-of-print. Does anyone out there know where a copy might be available? [Update: reader Joe Savage from Bridgewater, NJ reports "the board this person is looking for has surfaced a few times recently on eBay. I think it generally goes for about $17-20 or so." Good luck.]
There is a similar variant called cribbage football, which Dan Barlow describes thusly: "Turn on the [football] game, play as usual, but if your team scores, you get just as many points. If your team turns it over, you must jump backwards." (Fun with Cribbage, p. 55).
I'd never heard of this before, but I did an Internet search and found a description of it here. I don't know what is "Chinese" about it, and I haven't met too many Chinese cribbage players, so I would put it down to the Western proclivity to attribute anything exotic to the Far East.
See you all next time. Until then, keep on pegging.
- January 2003
prior article | Cribbage Forum home | next article
Cribbage Forum features articles on cribbage strategy and tactics by Michael Schell.
Original Material and HTML Coding Copyright © 2003 by Michael Schell. All Rights Reserved.