Reading a Payout Table Without Fooling Yourself

Sportsbook prices announce themselves. A number sits next to the market, and anyone who has spent an afternoon learning to read it can convert it into an implied probability and see roughly what they are being charged. Pick style platforms present the same information in a shape that hides it: a grid of entry sizes and multipliers, sometimes with variants that pay something for a near miss. The fee is fully present in that grid. It is just spread across cells instead of printed on the line.

Reading one properly is a learnable skill, and it takes about ten minutes to learn. This article teaches the method rather than the numbers, deliberately. Platforms revise their tables, run promotions, and offer different structures in different places, so any figure quoted here would be wrong for someone by the time they read it. Learn the method and you can price whatever table is in front of you today.

What a payout table actually is

A payout table is a complete list of the outcomes a structure pays for and what each one returns per dollar of entry. That is all. On a simple entry there is one paying outcome: every selection correct. On a protected or insured structure there are several: everything correct pays the top rung, and missing one selection pays a smaller rung instead of nothing. Some structures pay across more than one near miss level.

Two things are conspicuously not in the table, and both are supplied by you. The table never states how likely each outcome is, and it never states what your selections are worth. Those are the inputs that turn a list of multipliers into a price, which is why two people can look at the identical grid and correctly reach opposite conclusions about it.

The arithmetic, in one paragraph

For each outcome the table pays, multiply the payout by the probability of that outcome. Sum across all paying outcomes. That sum is the expected value of the structure per dollar staked. Compare it to one dollar. If the sum is below one, the structure returns less than it costs at your probabilities, and the shortfall is the fee you are paying for the product. If it is above one, your probability estimates say the structure is favorable, which is a statement about your estimates as much as about the platform. How a platform settles a tied leg also changes which outcomes pay, and the push article works through that on integer lines.

The comparison unit matters. Always reduce a structure to expected value per dollar, never to expected profit on a particular entry size, because entry sizes differ across structures and comparing dollars of profit conflates the price with how much you happened to put in. Per dollar is the only basis on which two different tables are the same kind of thing.

A worked table, entirely invented

Everything in the two examples below is made up. No platform pays these numbers and none is being described. The arithmetic is the point.

That last sentence is the entire lesson about protected entries, and it generalizes. Now change one input and watch the ordering flip.

That is the uncomfortable and useful conclusion. Which structure is better is not a property of the structure. It is a joint property of the structure and your probabilities, and a comparison that does not state the probabilities it assumed has not compared anything.

Protected and insurance style entries, fairly described

Structures that pay something for a near miss are neither a gift nor a trap by nature. They are a redistribution. Paying a rung for missing one selection has to be funded, and it is funded by lowering the top rung. Money moves from the rare outcome to the common one. Your average result barely moves; your distribution of results narrows sharply.

Whether that is worth it depends on what you want from the structure and on the specific numbers.

• Lower variance is real value to some people and none to others. A narrower distribution means fewer long empty stretches, which matters if those stretches drive decisions you would rather not make.
• The near miss rung is usually the most probable paying outcome, which means it carries a lot of the structure's expected value even though it looks like the consolation prize. Never price a protected table from the top rung alone.
• Protection is worth relatively more when your selections are weaker, because missing one is more likely. It is worth relatively less when your selections are strong, which is exactly what the second fictional example showed.
• Some protected structures are priced worse than their plain counterparts and some better. There is no general rule. Compute both at your own probabilities every time the table changes.

Why a bigger headline multiplier tells you nothing

Marketing for these products leads with the top number, because a large multiplier is easy to display and feels like generosity. It is not information. A multiplier is only half of a price; the other half is how often the outcome it pays for actually occurs, and larger multipliers exist precisely because their outcomes are rarer. Doubling the top payout while requiring two more correct selections can easily make the structure worse.

The same trap appears one level down, at the individual selection. Discounted and boosted versions of a prop change the payout in exchange for changing the line, and the boost is only good if the probability it costs you is smaller than the payout it adds. The modified payout article works through that trade, and the piece on alternate lines separates it from moving to a different number at standard pricing. In both cases the discipline is identical to the one in this article: never evaluate a payout without the probability attached to it. One important consequence is that many of these offers are sold over only, so when the research says the boosted side is bad, the conclusion is to pass, not to look for the other side. There is no other side to take.

Correlation breaks the multiplication

Every calculation above multiplied probabilities together, which assumes the selections are independent. Real selections frequently are not. Two players in the same game share a pace, a game script, and a weather condition. A quarterback and his receiver share the same completed passes. Two players on the same side of an outcome rise and fall together, which makes the all correct outcome more likely than the naive product suggests and also makes the total miss more likely, at the cost of the middle.

This matters differently for different structures. Positive correlation inside an entry helps the top rung and hurts the near miss rung, so it can make a plain structure look better than the independent arithmetic said and a protected one look worse. Negative correlation does the reverse. Either way, the product of individual probabilities is the wrong number, and the size of the error grows with the number of selections. The article on entry legs and edge covers how correlation and leg count interact, and it is the necessary companion to this one.

Making the comparison honest

A payout table comparison done properly has three parts, in order. Write down the probability you believe for each selection, from research done before you looked at the payouts. Compute expected value per dollar for each structure at those probabilities, including every paying rung. Then, and only then, decide whether the variance shape of the better priced structure is one you want. Doing it in the other order, picking a structure and reverse engineering probabilities that justify it, is a well documented way to talk yourself into anything, and it is the same after the fact reasoning the comparison article warns about when evaluating tools.

The reason we spend so much effort on probability estimates is that they are the input every one of these calculations depends on and the one nobody can hand you. Our engines simulate games and produce a full distribution for each market rather than a single number, so a probability exists for every line and every outcome a table might pay on, and each of those probabilities is graded afterward against what happened, at posted lines, with the record public in the Model Room. The projections themselves are on the Signal Board, and the multi book view lives in the Prop Grid. Bring your own numbers or bring ours, but bring numbers. A payout table without probabilities attached is a price tag with the price torn off.