A full box on six horses in a part box bet horse racing tricast produces 120 combinations. A banker tricast with one horse fixed first and the same five others floating for second and third produces 20. The cost drops by more than 80 percent. The trade-off is that you’re now betting on your banker finishing first — not just in the top three — but if your opinion supports it, the saving is enormous. Fix what you know, box what you don’t.
Full boxes are the blunt instrument of combination betting. They cover every permutation equally, which is useful when you genuinely cannot rank your selections. But opinions are rarely that flat. More often, you have one horse you’re confident about, two more you rate highly, and a couple of outsiders you want to include for coverage. The full box charges you the same for every permutation involving the outsiders as it does for the most likely outcome — and that’s where the money leaks.
Banker and part-box structures exist to match your bet’s shape to your opinion’s shape. This guide covers the mechanics of both, with formulas, examples, and the specific scenarios where each structure outperforms a full box.
Banker Forecasts and Tricasts: Fixing One Horse, Boxing the Rest
A banker bet fixes one horse in a designated finishing position and allows the remaining selections to fill the other positions in any order. The most common version is a banker first tricast: Horse A is fixed in first place, and Horses B, C, D, and E are boxed for second and third. The number of permutations drops dramatically because only one horse can occupy the first-place slot.
The formula for a banker first tricast with one banker and n floaters is n times (n minus 1) — the same as a box forecast formula, because you’re effectively boxing the floaters for two remaining positions. With four floaters: 4 times 3 equals 12 permutations. With five floaters: 5 times 4 equals 20. Compare that to a full five-horse box tricast — which is 5 times 4 times 3, or 60 permutations — and the saving is clear.
At £1 per line, a banker first tricast with one banker and four floaters costs £12 instead of the £60 you’d pay for a full five-horse box. That’s an 80 percent reduction. The tricast formula for a full box — n times (n minus 1) times (n minus 2) — grows cubically, while the banker formula grows only quadratically. The gap widens with every additional floater.
| Structure | Floaters | Permutations | Cost at £1 | Full Box Equivalent | Saving |
|---|---|---|---|---|---|
| Banker 1st + floaters | 3 | 6 | £6 | 24 (4-horse box) | 75% |
| Banker 1st + floaters | 4 | 12 | £12 | 60 (5-horse box) | 80% |
| Banker 1st + floaters | 5 | 20 | £20 | 120 (6-horse box) | 83% |
| Banker 1st + floaters | 6 | 30 | £30 | 210 (7-horse box) | 86% |
Banker forecasts work the same way but are simpler: one horse fixed first, the floaters competing for second place only. A banker first forecast with four floaters produces 4 permutations at £4. The full box on five horses would be 20 permutations at £20. The savings are substantial, and the bet is clean: you believe Horse A wins, and you want to cover four possibilities for second place.
The risk is concentrated and explicit. If your banker finishes second instead of first — or worse, finishes third while two of your floaters take the first two spots — the entire bet loses. A full box would have covered those outcomes. The banker deliberately excludes them to reduce cost. That’s not a flaw; it’s the design. You pay less because you’re betting a stronger opinion.
Part-Box Structures: Assigning Different Roles to Different Groups
A part-box takes the banker concept further by dividing your selections into groups with different positional assignments. Instead of fixing one horse and floating the rest, you create two or three groups and specify which positions each group can occupy.
The most common structure is a two-group part-box tricast. Group A contains horses you believe will finish first or second. Group B contains horses you believe could finish second or third but are unlikely to win. Every permutation where a Group A horse finishes first, another Group A horse finishes second, and a Group B horse finishes third is included. So are permutations where a Group A horse is first, a Group B horse is second, and another Group A horse is third. But permutations where a Group B horse finishes first are excluded.
The practical application is clearer with numbers. Suppose you have two Group A horses and three Group B horses for a tricast. The part-box covers: A1-first with any combination of the remaining four for second and third, plus A2-first with any combination of the remaining four for second and third. That’s 2 bankers times 4 times 3 equals 24 permutations — identical to a four-horse full box tricast, but with five horses involved. The coverage is shaped differently: you’ve excluded permutations where a Group B horse wins, and included permutations where Group B horses fill the minor places alongside your two principals.
Research into optimal selection counts for box tricasts suggests that for fields of 14 to 20 runners, four to five horses represent the productive range — enough to capture the likely top three, but not so many that the cost overwhelms the dividend. Part-boxing lets you reach five-horse coverage at four-horse prices by being explicit about which horses you expect to win and which you expect to fill the places. It’s a more honest representation of a typical racing opinion than a flat box, and the cost savings fund additional bets elsewhere on the card.
When Banker Beats Full Box: The Decision Framework
Use a banker when you have one horse that stands out clearly above the others in your selection group. The classic scenario: a well-handicapped horse trained by a top yard, returning to a course where it has won before, with a jockey booking that signals intent. You’re confident it wins. You’re much less sure about the places. A banker first tricast with this horse fixed and four or five floaters captures your opinion precisely and costs a fraction of the full box.
Use a part-box when you have a clear top tier and a clear second tier but the tiers overlap in the places. Two horses you think will be first or second, three more that could fill third — that’s a classic part-box tricast setup. The cost is lower than a full five-horse box, and the coverage reflects your actual view rather than treating all five horses as equally likely for every position.
Use a full box when you genuinely cannot rank your selections. Four horses that you consider equally capable of finishing first, second, or third — no standout, no hierarchy. The full box is the honest bet for a flat opinion, and the cost is the price of that honesty. Don’t use a banker just to save money if your opinion doesn’t support it. A banker on a horse that finishes second is a lost bet that a full box would have won.
The decision is always the same: does your opinion have a shape? If yes — one horse above the rest, or two tiers — use a banker or part-box. If no — every horse equally likely in every position — use a full box. The structure should mirror the opinion. The cost follows from there.