A live price is a model with a margin, updated continuously

§1Three layers: model, trader, spread

The model. An automated pricing model takes the current state of the event — scoreline, clock, and a set of statistics about the teams or players — and produces a probability for each outcome. It reprices whenever the state changes, which in a fast sport can be many times a minute. This is where the fair price lives, and it is the layer that responds to an event within a fraction of a second.

The trader. Human traders sit above the model. They set the parameters, adjust for information the model does not know, widen the market when the feed is uncertain, and suspend it when quoting would be indefensible. The trader is the reason a market can behave in a way that looks slightly judgemental — because it is.

The spread. The model and the trader produce a fair estimate. The price you are offered is that estimate with a margin taken out, expressed as a back price and a lay price that do not sum to a round hundred per cent. The width of that gap is a commercial decision about the market, the sport, the time of day and the size they expect to take.

§2The two-sided spread, converted into a percentage

The cleanest way to see a margin is to convert both sides of a market into implied percentages and add them. Two outcomes priced at 1.95 and 1.95 imply roughly 51.3 per cent and 51.3 per cent, which is 102.6 per cent — a margin of about 2.6 per cent, or about 1.3 per cent per side. That gap is the operator's compensation for making the market, and it is paid by every bettor who trades into it regardless of whether their bet wins.

For a worked description of the same arithmetic from the pre-match side, this site's earlier material covers implied probability directly; the point here is different and simpler. The spread you are quoted is not the same on both sides of the book, and it is not the same at all times of the event. It is a live commercial variable, and reading it is part of reading a live price.

§3Why the margin usually widens in play

Compare a pre-match market with the same market in play and the offered price is normally shorter, which means the margin is wider. There are several unglamorous reasons and they all point the same way.

The practical consequence is worth stating plainly: the same selection, bought in play, is usually bought at a worse price than the pre-match market offered. If your reason for buying it in play is that the price looks generous, check it against what the pre-match market said, because the apparent generosity is often just the margin being wide in a thin market.

§4Liquidity, stake limits and the size of a live market

Live markets are usually thinner than pre-match markets, and operators cap the stake they will accept on a live selection. The cap is not a judgement about you; it is the operator's statement of how much risk it is willing to hold at that price. Once the cap is reached the operator either shortens the price or stops accepting bets on that selection, and either way the market has moved somewhere you cannot follow.

Thinness has a second effect that is easy to miss. In a thin market the price can move a long way on a small amount of money, including your own. A stake that is unremarkable in a pre-match market can be the largest trade in a live one, which means part of the move you see after your bet may be caused by your bet. If you have ever watched a price shorten immediately after you took it, that is not coincidence and it is not flattery.

§5Requested odds, and the bet that comes back different

Many live products let you request a price rather than accept the one shown. You type odds and stake, the request goes to the pricing engine, and the engine decides whether to take it. Three outcomes are typical: the odds are accepted and the bet stands; the odds are unavailable and you are offered the current price instead; or the request is rejected outright because the market moved while it was pending.

Requested odds are often described as a way of avoiding the moving market, and that is not what they do. They give you a way to name your price and let the market decline. The value of the feature is that it makes rejection visible — you find out that your view of the fair price is outside the market's, rather than having it quietly taken. That is useful information, and it is not an advantage.

§6Reading a live price properly

• Convert both sides before you judge. The margin is in the gap between the implied percentages, not in either price on its own. A price that looks short on one side looks generous on the other.
• Compare with the pre-match market. If the in-play price is longer than the pre-match price was, ask what has changed in the event to justify it, and whether the change is in your favour.
• Notice the width, not just the level. A market whose spread has widened sharply is telling you the operator’s own uncertainty has risen. That is a warning about volatility rather than an opportunity.
• Respect the stake cap. If your intended stake is near the maximum the market will take, your own bet is part of the price move.
• Treat a rejection as information. A consistently rejected price is a measurement of your model, not an obstacle put in your way.

§7Notes on numbers and sourcing

The conversion between a price and an implied percentage is arithmetic and needs no source. The statement that live margins are typically wider than pre-match margins is a general observation about how these markets are built rather than a measurement of any particular operator, and no operator is named. The illustrative margins on this page are examples chosen to show the arithmetic¹.

1. 1An example is not an observation. The figures on this page are chosen so the arithmetic is easy to follow, and they describe no operator, market or event. Where a real number would require data that this site does not have, it is not given.