What value is, and how to tell whether you had any

§01A price, a chance, and the gap between them

Decimal odds are a probability wearing a different coat. A price of 2.00 implies a 50% chance, because a fair coin priced at 2.00 returns your stake twice and breaks even over time. A price of 1.91 implies 52.4%, and 4.00 implies 25%. The conversion is exact and reversible: implied chance = 1 / decimal odds. Nothing about the market's opinion is hidden in that number — it is printed on the price.

Value is the distance between that printed chance and the chance you believe is real. If you believe a selection wins 57% of the time and it is offered at 2.00, you are being paid as though it wins half the time. That gap is the value, and it is the only reason a bettor who bets at prices can expect to come out ahead. It is worth saying plainly what the gap is not: it is not a promise, it is not a fact, and it is not independent of you. Your 57% is an estimate you brought to the table.

So the whole practice reduces to three questions. What chance does this price imply? What chance do I think is real — and why? And can anything outside my own head check the second answer? The rest of this site takes them in order.

• A price longer than the chance justifies
• Your estimate of the true chance
• The closing line, as an outside reference

§02The margin is priced into both sides before you arrive

Both prices on a two-way market do not add up to 100%. A typical football market might be 1.91 and 1.91; converted, those are 52.4% each, and together 104.7%. The 4.7 points above 100 are the operator's margin, sometimes called the overround or the vig. It is not a fee charged at the window, it is a haircut taken out of the price itself, on every selection, whether the bet wins or loses.

That changes the size of the task. To break even you do not need to be right as often as the price suggests; you need to be right more often than the price suggests by enough to clear the margin. At 1.91 the break-even win rate is 52.4%, not 50%. A bettor who is genuinely a fraction better than the market at 52% would still lose money at those prices, which is uncomfortable and true.

§03The only outside test is the closing line

An estimate of the true chance is unfalsifiable on its own. You can believe anything about a match, and a win or a loss tells you almost nothing about whether you were right. What does move is the price: as money and information arrive, the market drifts toward a number that reflects everything known at the moment betting stops. That final number is the closing line.

The closing line is not truth. It is the sharpest publicly available consensus, because it is where the largest and best-informed money has finished arriving, and it is the closest thing to an independent judge of your estimate that you will get. If you consistently take prices longer than the line closes at, you were consistently reading the market early — which is what a real edge looks like from the outside. If you consistently take prices that then shorten further against you, the more likely explanation is that the market knew something you did not.

This is the whole argument for measuring closing line value instead of counting winners: it produces a signal from every bet, including the ones you lost, and it can be evaluated in a few dozen bets rather than a few thousand.

§04Why a real edge still loses for months

Three things make a genuine edge feel like a losing habit. First, variance: a bettor winning 55% of even-money bets still loses five in a row roughly once every fifty sequences of five, and will spend weeks below the starting bankroll. Second, sample: to distinguish 52% from 55% at a reasonable confidence takes on the order of thousands of bets, not dozens. Third, the estimate itself moves — a model that was right about a league last season may be stale this season.

None of that is an argument for ignoring the arithmetic. It is an argument for not reading a short run of results as evidence either way, and for keeping a record that is honest about which numbers are measured and which are guessed.