Why a builder never pays what the legs suggest
Combine four selections from four different matches and the arithmetic is straightforward: multiply the prices. Combine four selections from the same match and the arithmetic breaks, because the outcomes are not independent. A team winning comfortably is more likely to have kept a clean sheet, more likely to have had its main striker score, and more likely to have led at half time. Those events travel together, and a price that ignored the fact would be an invitation to be picked off.
Correlation, positive and negative
Positive correlation means two outcomes tend to happen together, and a builder combining them is worth less than the multiplied prices imply. Negative correlation means they tend to exclude each other — a low-scoring match and a hat-trick, for instance — and such a combination is worth more, if the platform allows it at all. Pricing engines estimate these relationships from models rather than from a lookup table, which is why two operators can quote noticeably different numbers for what looks like the same builder.
Why multiplying is wrong
Multiplication assumes independence. Within a single fixture almost nothing is independent, so the multiplied figure overstates the true price of a positively correlated bundle. The builder price you are offered is the model's estimate with margin applied, and it will usually be lower than the naive product.
Some combinations are refused
Platforms block combinations that are effectively guaranteed by each other, because those would be a certainty dressed as a multiple. A refusal is not a glitch; it is the pricing engine declining to sell a bet whose legs are too tightly bound together.
More legs, more model risk
Every extra leg adds another relationship the model has to estimate, and estimates carry error. Operators respond by widening the margin on longer builders. The result is that the most creative-looking slips are usually the most expensive ones on the coupon.
Compare builders, not legs
Because engines differ, the only meaningful comparison is between the final quoted prices for the identical builder at different operators. Comparing the individual legs tells you very little about which platform will price the bundle better.
Questions readers send us
Why is my builder priced lower than the legs multiplied?
Because the legs are positively correlated and the engine has accounted for it, and because margin has been applied to the bundle. Both effects push the quote below the naive product. It is the expected behaviour of the product rather than a deduction being taken from you.
Can a builder be voided if one leg is affected?
Treatment varies more here than in standard multiples, because the price was calculated on the bundle as a whole rather than leg by leg. Some operators void the entire bet, others recalculate. The specific rules for builders are worth reading in advance, as they are not always the same as the rules for ordinary accumulators.
