Why Abacus Doesn't Work

I have it in mind eventually to write a detailed article about the mechanisms of solitaire, describing the various elements which are combined in different ways to make different games: just as multiplayer card games can have elements such as tricktaking, various rules for following suit, trumps, melding, selecting from discard piles, etc., so does solitaire have elements such as foundations (built in various ways), different types of reserves (sequential/simultaneous, hidden/open, single/multiple, replenishing/non-replenishing, initially full or empty), discarding schemes, dealing and redealing protocols, etc. Looking at Abacus gave me some ideas on this and why certain combinations of mechanisms don't combine well or give the desired strategic effect.

Abacus is an original game by Randy Rasa, and is included in the Absolute Yukon package (available as free adware, or with no ads and additional games for $15). More information is available from Randy's website.   Despite my reservations about Abacus, I recommend Absolute Yukon. Yukon itself is one of my favorite non-open solitaires (along with Par Pyramid), and the program is very well done. I especially like the automatic column-squeezing feature, which works really well in Yukon.

When I say that Abacus doesn't work, I don't mean that the game is unplayable or unwinnable -- actually the contrary; it's far too easy to win and the game is too close for my taste to being self-working (perhaps you can read "doesn't work" as meaning "doesn't work for me" or "unsatisfying" -- I suspect that many people will like Abacus just fine). In Abacus, spaces in the tableau (i.e. empty columns) may only be filled with kings, as in Klondike. It's entirely right that this rule (KingOnly in Pretty Good Solitaire's terminology) be used, rather than allowing any card to fill an empty column (as in FreeCell or Thumb and Pouch). The KingOnly rule is a good one in my view in Yukon and Klondike (Thumb and Pouch being too easy for my taste), but not as good a rule in standard Scorpion, which is excessively difficult compared to Wasp. If Abacus used the Thumb and Pouch rule, the game would be virtually impossible to lose.

The only way to be blocked in Abacus is for every column to have a card at the bottom (exposed) which cannot go to a foundation, and where the next-lower card (the card which precedes it to the foundation and gets packed onto it in the tableau) is buried in the same column. The fundamental flaw in Abacus is that being blocked seems much too easy to avoid, and it would be almost trivial to avoid if you could move any card into an empty column.  The underlying cause of this "fundamental flaw" is that none of the cards in Abacus are hidden: in most Scorpion/Yukon games, what drives strategy is to try and uncover the cards which are dealt face down at the beginning. In Abacus there are no face down cards, and one can almost move cards around at random and probably win. What drove this point home to me was the realization after playing my first couple of games that one can deal out the entire deck at the start (clicking on the deck three times) and still win easily, provided you are careful, once the kings have all reached the tops of columns, to avoid the kind of blocked position I described earlier. One of the first few times I tried doing this, I was careless (deal 20172) and got blocked, but undoing a few moves made it easy to fix the mistake. Since then I have won almost a dozen games without difficulty. A quick calculation shows that the chance of being blocked after dealing the whole stock out at the beginning is 1 in 6.37 million. How can Abacus be fixed? I suspect that some of the cards need to be dealt face-down at the start, but probably not as many as the Klondike pattern -- Russian solitaire is much too unwinnable for my taste (I think, by the way, that the Wasp rule would make Russian solitaire better too). The game of Wasp in Absolute Yukon, on the other hand, seems too easy; the original version in Solitaire's Journey has an extra column of face-down cards which seems to make a *lot* of difference. I've always liked the game of Wasp; allowing any card to fill an empty column makes the game more strategic (far more interesting decisions to make) as well as easier to win, but the number of face down cards is a delicate balance and the original Scorpion tableau (as opposed to what is called Scorpion II in PGS) seems much better. For that reason, and because I don't like to see rules changed without a good reason, I'd be happier if Wasp in Absolute Yukon was the same as the original in Solitaire's Journey.

I'll close with a comment on the "Calculation" feature: I expect some people will like this a lot, but to me it is an irrelevancy -- it forces you to think about what card goes on which (good practice for Calculation itself, perhaps), but has no effect whatsoever on strategy. Bob Abbott (inventor of Eleusis, Ultima, and other games) says that clarity should be one of the guiding principles of game design, and Abacus in my view violates this principle here. An analogy would be a "variant" of bridge where the cards were ranked differently in each suit; confusing to players until/unless they got used to it, but having no essential effect on the game itself (since we assume deals are random). The ordering in Abacus is just a different mapping of the cards (if the second foundation suit is clubs, 4C-6C-8C are exactly like 2C-3C-4C would be if clubs was the first suit). [While this feature would be very bad in bridge, since the game would be undoubtedly be marred by mistakes in play in remembering which cards ranked higher in different suits, it's essentially irrelevant in Abacus since the program won't let you make an illegal play. Still, I wouldn't want to hand-deal Abacus.]  In Calculation, building up the foundations by different intervals works because the foundations are suitless, so each rank has a different strategic flavor and there are lots of interesting decisions to be made (should I use this queen on the third or fourth foundation?). In a game with suited foundations this doesn't happen, since every card can only go in one place.

This article is copyright © 2007 by Michael Keller.  All rights reserved.