Betting — From a View to a Ticket
Buy only the outcomes you have assessed, within a fixed risk budget.
Theory: the selection is only the beginning
You can have a sensible opinion about a horse and express it through a poor bet. The pool may charge more, the combination may be overbet, or the ticket may expose too much of your money to one outcome.
Decide the maximum amount you are willing to lose before building the ticket. For these exercises, use imaginary units. Chapter 15 deals with a real bankroll; learning the mechanics requires no deposit.
Start with the simplest bet that expresses the view. Extra combinations need their own justification, not just the hope of a larger return.
Know what the ticket buys
A win bet pays if the selection wins under the applicable settlement rules. A place bet covers a specified finishing position, which varies by product and field. The place probability is not simply the win probability multiplied by the number of places.
An exacta or forecast requires the first two in the correct order. A quinella covers a selected pair in either order. A trifecta or tierce requires the first three in order; other products cover unordered combinations.
A multi-race bet requires qualifying selections across specified races. Product names and settlement rules differ between operators. Check the nominated legs, unit stake, scratches, consolation rules and any limits before considering value.
Backing a horse buys the winning outcome. Laying on an exchange takes the other side. For an illustrative lay stake of 10 units at 2.50, the potential profit is 10 and the liability if the horse wins is 15, before commission. Do not confuse the displayed lay stake with the maximum loss.
Joint outcomes need joint probabilities
The chance of A winning and B finishing second is:
P(A wins) × P(B finishes second given A wins).
It is not the product of A's and B's win probabilities. B's chance of finishing second depends on how the race unfolds and on A having won.
Across different races, independence can sometimes be a useful approximation, but shared weather, track bias and systematic forecasting errors can connect your exposures. More importantly, a tote multi-race dividend is not generally the product of individual win odds. The money placed on combinations and the product's rules determine the payout.
A small apparent edge in each leg does not establish a particular ticket's profitability. You need a view of its combined probability and the return available.
One structural fact from the Scandinavian V-pools we measured: upsets paid roughly nine times what favourites paid per winning ticket. That is a property of how the pool distributes money, not evidence that outsiders are underbet; the favourite–longshot bias in Chapter 12 points the other way. It does explain why the multi-race returns worth having come from a well-founded selection the crowd is against, and not from buying one leg wider.
A/B/C/X is a budgeting shorthand
You can organise a ticket with groups:
- A: the main selections supporting the ticket.
- B: credible alternatives you may cover selectively.
- C: less likely or less attractive alternatives considered only in limited scenarios.
- X: omitted from this ticket.
The letters are choices about coverage, not objective properties of the horses. X means you did not buy that outcome, not that it is impossible.
A very likely winner can be unattractive at the price. A less likely horse can be a useful selection if its chance is sufficiently underrepresented in the pool. Probability and value should both influence the groups.
An invented four-leg example
Suppose a hypothetical four-leg product has two candidates per leg:
| Leg | A selection | B selection |
|---|---|---|
| 1 | 1 | 2 |
| 2 | 3 | 4 |
| 3 | 5 | 6 |
| 4 | 7 | 8 |
Buying every combination at one unit costs 2 × 2 × 2 × 2 = 16 units. It covers the all-A combination, four combinations with exactly one B selection and eleven with at least two B selections.
A different 16-unit allocation is:
| Ticket | Units on this combination |
|---|---|
| 1–3–5–7 | 8 |
| 2–3–5–7 | 2 |
| 1–4–5–7 | 2 |
| 1–3–6–7 | 2 |
| 1–3–5–8 | 2 |
| Total | 16 |
This weighted ticket buys only five distinct combinations. It concentrates on the view that the A selections are strong and at most one needs replacing. If two B horses win their legs, it loses completely.
The wider ticket is not automatically wasteful, and the weighted one is not automatically clever. Their value depends on the probabilities and payouts. The example shows what concentration changes: coverage and exposure, not the underlying chance of the horses winning.
A box can be deliberate
A box covers permutations or combinations of selected horses under that bet's rules. It can be a reasonable expression of an approximately equal view. It can also buy many weak combinations merely because the interface makes that convenient.
Before submitting, inspect the actual tickets. Count the combinations, account for repeated lines and check the stake attached to each. “Small unit stakes” can still produce a large total.
If you cannot explain why a combination belongs, remove it or return to the analysis. If you cannot estimate the relevant joint outcomes, the simpler win market may be a better learning environment.
Carryovers change the pool, not the standard of evidence
Money carried from an earlier pool can improve the relationship between new stakes and money available for prizes. The benefit depends on the rules, deductions, eligible winning outcomes and new money attracted.
It does not make every combination profitable. More competition may arrive, and an apparently generous pool can still reward the combinations you bought poorly. Treat a carryover as a change in the economics to investigate, not a free lunch.
Count the exposure more than once, but not the money
A win bet, an exacta and a multi-race ticket can all depend on the same horse. They are separate tickets but a shared risk.
Record the total amount you can lose if the horse disappoints. Do the same for a shared assumption, such as a strong front-running bias. Diversifying ticket names does not diversify the underlying opinion.
Practical: compare two tickets
Quiz: what did you buy?
Up next: how to review a decision when the result is trying to rewrite your memory.