Which soroban abacus should you buy? (a cheap one is fine)

You have decided to try the soroban, you have opened a shop page, and there are forty frames on it — plastic and wood, thirteen rods and twenty-three, colour-coded beads, one that takes batteries. The good news is that this is the cheapest decision in the whole project and one of the hardest to get badly wrong. Three things matter. Everything else on that page is decoration.

The only three decisions that matter

• Frame type. For mental calculation you want the soroban layout: one bead above the beam, four below it.
• Rod count. Thirteen, fifteen and seventeen are the usual beginner sizes, and any of them is enough for years of work.
• Bead action. A bead should move with a light push, then stay exactly where you left it.

Frame type: buy a 1:4 soroban

The modern soroban uses a 1:4 bead configuration: one bead above the reckoning beam — the go-dama, or heaven bead — worth five, and four beads below it worth one each, so a single rod encodes every digit from 0 to 9 in exactly one way. That one-and-only-one property is the reason the soroban became the mental-calculation instrument. A number has a single bead picture, so a child memorises the picture instead of re-deriving it every time.

The heaven bead is not a decimal-point marker, which is a common confusion — the decimal position comes from choosing a unit rod and reading places outward from it. The other bead frames you will meet are not wrong, they are different tools. A suanpan carries extra beads, so one rod can hold a value above nine, which historically suited Chinese weight arithmetic where one jin equalled sixteen liang: a use case, not a number base. A rekenrek has twenty beads in two rows of ten, each row split five red and five white, and every bead is worth one wherever it sits, so it teaches counting and number bonds rather than place value. A Russian schoty has ten beads on most wires and no dividing beam at all, which makes it a counting board rather than a place-value trainer.

How many rods should a soroban have?

Each rod holds one decimal digit, so rod count is just how wide a number you can lay out. A beginner course spends its first months on one- and two-digit addition and subtraction, which needs two or three rods; the rest of the frame is headroom for later multiplication and division, where you keep one operand on one side and build the product on the other. Frames sold for beginners commonly come with thirteen, fifteen or seventeen rods, and instructional guidance treats any of those as ample — fifteen or seventeen are the usual recommendation, as a balance between capacity and a frame small hands can still reach across. Larger twenty-three-rod frames are typically taken up later, if at all.

There is a research reason not to over-buy. The standard reference here is Frank and Barner, writing in the Journal of Experimental Psychology: General in 2012: trained mental-abacus users appear limited to representing about three or four abacus columns at a time, which suggests each column is held as a separate object in visuospatial working memory. That is a description of how the representation seems to work, not a measured ceiling on what practice can reach — but it does mean extra rods buy room on the desk. They do not buy room in the head, and the head is where this training is going.

Bead action — and why a sticky cheap frame is still fine

Put a finger on one bead and push. It should move with a light flick, stop where you left it, and not drag its neighbours along. Then tilt the frame: nothing should slide. Loose beads that drift make a child doubt a correct answer, and beads so tight they need two fingers slow the technique down at exactly the age when speed is the reward. Wooden frames with a slightly stiff rod tend to loosen with use; plastic frames with wobbly rods tend not to tighten. That is the whole quality test, and it takes five seconds in a shop.

And if the only frame you can get is a cheap one, buy the cheap one. In a June 2019 thread on the Australian parenting forum Whirlpool, a parent weighing commercial lessons at “$30 a lesson” wrote that they had “ordered a cheap $5 soroban online and will teach him myself at his own pace” — one anonymous anecdote, in what was almost certainly Australian dollars, and not a market rate. But the ratio it describes is real: the frame is the smallest line in this budget. For contrast, ALOHA Ireland is one of the few programmes publishing a headline figure on its own site, listing 420€ per term per level for two hours of class a week in groups of at most ten students, with seventeen-week terms and two terms a year (as listed on 30 July 2026). That fee buys live in-person teaching, which no piece of software provides. Either way, the instruction is the purchase and the beads are a rounding error.

What not to buy

• Decorative frames. Anything sold as an ornament, a desk piece or an antique reproduction — the beads are usually either glued tight or completely free, and neither is usable.
• Colour-coded place-value toys. A frame with one colour per column is built to show hundreds-tens-ones to a five-year-old, not to run soroban technique, and the colours become a crutch the mental picture later has to unlearn.
• Program-specific frames, unless you are enrolled in that program. Some providers supply their own redesigned tool rather than a standard soroban: Indian Abacus, for instance, describes its own as a newly invented and patented digital and non-digital abacus, where the verifiable public record is published patent filings. If you are enrolled with that provider, use the tool they supply — it is built for their curriculum. If you are not, a standard soroban is the frame that fits whichever curriculum you choose afterwards.
• Electronic gadgets sold as a “digital abacus”. A bead frame does not need a battery, and once the calculation moves into the head a screen is in the way.
• Tablet-only apps, if you do not already own a tablet. SoroTouch, for example, states that it is only available on tablets and cannot be used on smartphones or computers, and lists Android 10.0 or later, a 64-bit CPU and a four-point multi-touch display among its requirements — a hardware bill that has nothing to do with beads. If a tablet is already in the house, that requirement costs you nothing; if it is not, price it in before you compare.

Where the frame stops being the point — and what to spend on instead

The physical soroban has a job with an end date. It teaches the finger technique and builds the bead picture, and then the goal is to take it away, so the child calculates on a frame they can see with their eyes closed. Chunjie Wang’s 2020 review in Frontiers in Neuroscience concludes that abacus-based mental calculation training “has the potential to enhance various cognitive skills including mathematics, working memory and numerical magnitude processing” — while noting that most studies detecting a benefit did so after around three years, that the gains look like near transfer to tasks closely related to the trained ones, that on fluid intelligence the effect “appears to not go beyond a placebo effect”, and that small samples, cross-sectional designs and passive control groups in the underlying studies “may artificially inflate the overall effect”. Read that as a schedule rather than a warning. What decides whether your child gets three years of practice is not the frame.

So the honest answer to “which abacus should I buy” is: a plain fifteen- or seventeen-rod soroban, in whatever material reaches your door cheapest, and then put the rest of the budget into structure — an ordered level ladder, a daily drill that grades itself, and someone or something that notices in week two when a finger habit goes wrong. Kani publishes about USD 7 a month with a free tier and self-serve checkout, which is well below per-hour centre tuition and more than several abacus apps charge: Abacus4Everyone lists INR 399 a month and Vedaavi lists USD 2.49 a month, both as published in July 2026. Compare against the model you would actually use, and never against the frame.

Brand names mentioned in this article are trademarks of their respective owners. Kani is not affiliated with, endorsed by, or licensed by any of them.