Hot and cold lottery numbers: what a frequency chart can and cannot tell you

Some numbers really have been drawn more often than others. Here is why that is expected, how to tell ordinary variation from a real fault, and why it does not help you pick.

Lottery · Updated 5 October 2026 · 4 minute read

Open the draw history of any lottery and count how many times each number has appeared. The counts will not be equal. One number will lead, another will trail, and one will not have been seen for months. People call these hot, cold and overdue numbers, and a great deal of advice is built on them. The counts are real. The conclusions usually drawn from them are not.

Uneven counts are what fairness looks like

Flip a fair coin 100 times and you will rarely get exactly 50 heads. Getting 46 or 55 is ordinary. A lottery drum behaves the same way with more outcomes. In a game that draws 5 balls from 69, each ball has a 5 in 69 chance of appearing in a given draw, a little over 7 percent. Across 1,000 draws the expected count for each ball is about 72. Chance alone will scatter the actual counts around that figure, typically by 8 or so in either direction, and with 69 balls in play a few will land further out than that. A table where every ball sat at exactly 72 would be the suspicious one.

So the first question to ask of a hot number is not whether it is ahead. It is whether the lead is larger than chance would produce among that many balls over that many draws. Almost always it is not.

The mistake with a name

The belief that a number is due because it has been absent has been studied for long enough to have a name, the gambler's fallacy. It rests on an intuition that is half right. Over a very long run, each number's share of the draws does settle toward its expected share. What the intuition gets wrong is how that happens. The drum does not compensate. An early shortfall is never made up; it is diluted, as thousands of later draws make it a smaller and smaller fraction of the total.

Each draw is independent. The balls carry no record of last week, and the machine is loaded the same way every time. A number that has been missing for 40 draws has exactly the same chance in the next one as a number drawn last night.

The opposite belief, that a hot number will stay hot, fails for the same reason. Both treat the history as if it were a force acting on the next draw.

Long gaps are common, not strange

If a ball has a 7.2 percent chance of appearing in each draw, the chance that it misses 20 draws in a row is about 22 percent. With 69 balls in the drum, at any moment you should expect a handful to be in the middle of a drought that long. A gap only looks remarkable because attention goes to the one ball that is having it and not to the 68 that are not.

When a frequency chart does matter

There is one legitimate use for these counts, and it belongs to the lottery, not to players. A physical fault, such as a ball that is slightly lighter or a machine that mixes badly, would show up as a number whose count drifts further from the expected value than chance allows and keeps drifting. Lotteries test for exactly this. Ball sets are weighed and measured, rotated between draws, and the results are audited. A statistical test such as chi-squared compares the whole table of counts with what a fair draw would produce.

That is also the honest way to read a frequency chart yourself. Look at how far each count sits from its expected value, and at whether the spread across all the numbers is the size that chance predicts. In the long published histories of the major games, it is.

What choosing numbers can change

Your choice of numbers cannot change your chance of winning. It can change how much you would win, because a jackpot is split among everyone holding the winning combination. Many players choose birthdays, which crowds the numbers from 1 to 31, and many copy patterns from the play slip or last week's results. A combination that few other people hold is no more likely to be drawn and is less likely to be shared if it is. A random pick achieves that about as well as anything.

A fair summary

• Frequency tables describe the past accurately and predict nothing.
• Uneven counts and long gaps are what independent draws produce.
• The useful comparison is each count against its expected value, not one number against another.
• No selection method changes the odds. The only thing a player controls is how many tickets to buy, and whether to buy any.