Probability & statistics, explained simply.

Random experiment

Any process with an outcome you can't predict in advance, but that you could repeat under the same conditions. It's the starting point of all probability. Before you can talk about chances, you need a defined thing that could turn out different ways.

In lottery terms: every draw is a random experiment. Same barrel, same rules, same number range, and yet nobody, including the operator, knows which balls will come out until they do.

Sample space

The complete list of every possible outcome a random experiment could produce. Nothing outside this list can happen; everything inside it, in principle, can.

In lottery terms: for Oz Lotto, the sample space is all 62,891,499 possible 7-number combinations from 1–47. One of those, and only one, gets drawn each time. Every combination in that space is exactly as valid as any other.

Random variable

A way of turning the outcome of a random experiment into a number you can actually work with, instead of a messy description.

In lottery terms: "how many of your numbers matched" is a random variable. The draw itself is a jumble of balls, but that one number is what decides which prize division, if any, you land in.

Probability

A number between 0 and 1 (or 0% and 100%) describing how likely a particular outcome is. Zero means impossible, one means certain, and everything real sits somewhere in between.

In lottery terms: matching all 7 Oz Lotto numbers has a probability of 1 ÷ 62,891,499, about 0.0000016%. See the full spread across every game on our Odds page.

Independence

Two events are independent if knowing one happened tells you absolutely nothing about whether the other will. This is arguably the single most important idea on this entire site.

In lottery terms: every draw is independent of every other draw. The barrel has no memory. A number that's run "hot" for months carries zero extra chance next draw. This is the actual mathematical reason "past frequency does not predict future results" is true everywhere on this site, not just a slogan we like. We put it to the test properly in our review of the hot-and-cold-numbers strategy.

Conditional probability

The probability of something, given that you already know something else happened. Written P(A given B), it answers "does this extra information change the odds?"

In lottery terms: "given it's Saturday, what's the chance Saturday Lotto draws tonight" is a fair conditional question. But "given number 7 has come up three times this month, what's the chance it comes up tonight" isn't, because draws are independent (see above), conditioning on the past changes nothing. The conditional probability and the plain probability are exactly the same number. That's the actual maths behind why "overdue" numbers aren't a thing.

Joint probability

The probability that two or more things happen together, at once. When the events are independent, you find it by multiplying their individual probabilities.

In lottery terms: Powerball's Division 1 needs a joint event: all 7 main numbers correct and the Powerball itself correct, together. Since the main barrel and the Powerball are drawn independently, their probabilities multiply, which is exactly how that famous 1-in-134,490,400 figure is built.

Expected value

The long-run average result you'd get if you repeated something many, many times, with each outcome weighted by how likely it is. It's not what happens any single time. It's what the average settles into eventually.

In lottery terms: the expected value of any lottery ticket, anywhere in the world, is mathematically always less than what you paid for it: a fixed share of every dollar funds the prize pool, the retailer, and the state, before a single ball drops. That's not pessimism, it's arithmetic. It's exactly why our house line is that hope is worth what you paid for the ticket, and nothing more. Our reviews of syndicates and system entries both come back to this same number.

Law of large numbers

As a random experiment gets repeated more and more times, the observed results drift closer and closer to the true underlying probabilities. It only kicks in at scale. Small samples stay lumpy and uneven.

In lottery terms: this is why frequency charts look roughly balanced after thousands of draws, not because any number is "due," but because that's simply what randomness looks like once you've collected enough of it. Zoom in on just the last 10 draws and you'll see clumpy, uneven patterns instead. That's also completely expected, and also telling you nothing about tonight. We put this exact idea to work in our look at the most and least drawn numbers.

Shape bands (Scorcher, Straight Down the Line, Brass Monkeys)

Our house classification for how a draw's numbers are arranged. A draw's shape is two counts taken together: how its numbers split odd against even, and low against high, where low means the bottom half of the game's number range. Each axis carries a measurable share of probability, so shapes can be banded by how common they are: a Scorcher axis carries more than 20% of the probability mass, Straight Down the Line sits between 15% and 20%, and Brass Monkeys is anything rarer.

A draw only belongs to a band when both axes land in it. When the two axes disagree, say a dead-common odd/even split alongside a rare low/high one, the draw is a mixed shape that belongs to no single band, and everywhere on this site it gets called exactly that rather than being squeezed into the nearest label.

One rule that applies everywhere: shapes are counted on the winning main numbers only. Supplementaries and the Powerball never enter the odd/even or low/high counts, on this page or any other, because they are drawn as a separate event and belong to a different part of the prize table. When you see "4 odd, 2 even" it always means four of the six main numbers. And as with every stat on this site: a shape describes what happened, it never predicts what comes next.