Playing Lotto is just statistics... and luck

What exactly is the chance

The number of all possible sets of six numbers from a pool of 49 is the binomial coefficient C(49,6) = 13,983,816. This result is not an estimate or an observation — it follows directly from the rules of the game, in which we choose 6 numbers out of 49, and order doesn't matter. The table below shows how the chances of the various prize tiers break down for a single bet.

Probability of hits in a single Lotto bet (6 of 49)

Hits	Number of sets	Chance (1 in…)	In percent
6 of 6	1	13,983,816	0.0000072%
5 of 6	258	54,201	0.0018%
4 of 6	13,545	1,032	0.0969%
3 of 6	246,820	57	1.7650%
fewer than 3	13,723,192	1.02	98.1362%

This last row is the most important one in the table: in more than 98 cases out of 100 a single bet gives no winnings at all. The numbers in the „Number of sets” column add up to 13,983,816, that is, to every possible draw outcome — there is no room there for a „better” or „worse” set.

Why „hot” and „overdue” numbers give you nothing

The balls in the machine have no memory. If number 7 hasn't come up in a hundred draws, its chance in the next one is exactly the same as any other number's. The belief that „it has to turn around eventually” has a name in statistics — the gambler's fallacy — and it comes from transferring the law of large numbers, which describes the behaviour of an average over a very long series, onto a single event.

The opposite belief, that a number has „found its form”, is the same fallacy turned 180 degrees. Both stem from the same intuition: that randomness evens itself out over a short time. It doesn't. It only evens out on a scale no player will live to see — over tens of millions of draws.

The claim about a „defective ball”, which circulated in older texts about Lotto, doesn't hold up either. Balls and machines are subject to inspection, and even if one ball were a hair heavier, detecting this would require analysing thousands of draws and would give an advantage so tiny it wouldn't cover the cost of the bets needed to exploit it.

What statistics is actually good for

There is one thing a player does have influence over — and it isn't the chance of winning. It's the size of the prize once you do win. The prize at a given tier is split among everyone who matched the same set. If we pick a combination that a thousand other people play, we share it with a thousand other people.

People don't choose numbers randomly. They cluster them around birth dates (hence the overrepresentation of the 1–31 range), around diagonals and patterns on the ticket, around sequences like 1-2-3-4-5-6 and around numbers considered lucky, such as 7 and 13. Statistics lets us spot these clusters and avoid them. This increases the expected conditional payout, not the probability of winning.

• Sets with numbers above 31 are crossed out less often than birth dates.
• Regular patterns on the ticket (rows, diagonals, squares) are played by many people at once.
• Arithmetic sequences and „neat” sets like 10-20-30-40 tend to be heavily played.
• A set from a random generator is usually less similar to someone else's than one chosen by intuition.

Mode, median and mean — without the legends

Three concepts that show up most often in older texts about Lotto — and are most often used against their actual meaning.

Statistical measures in the context of number games

Measure	What it measures	Does it predict the draw
Mode	The value that occurs most often in a set of historical data.	No. It describes the past, it doesn't change the distribution of future draws.
Median	The middle value after sorting — for a pool of 1–49 that is always 25.	No. It's a constant that follows from the pool's range, not from the results.
Arithmetic mean	The sum divided by the number of elements; for a pool of 1–49 it tends toward 25.	No. Convergence toward 25 is a mathematical necessity, not a hint.
Standard deviation	The spread of numbers in a set relative to its mean.	No, but it shows how unusual a set is compared to other players' choices.

All of these measures are descriptive tools. They describe data that already exists. None of them contains a mechanism that would let you derive anything about the next draw from past results — because no such mechanism exists in a game of chance.

What this looks like in practice in 2026

Lotto draws take place several times a week, in the evening — around 22:00. The exact schedule, bet price and current guaranteed prize tiers are published by Totalizator Sportowy, and its website is the only binding source. Check there, because these details change independently of texts like this one.

On our site you'll find the historical results of every draw and the Check your ticket tool, which runs your set against the entire available history. It's worth doing this exercise, because it works both ways: it shows how much you theoretically could have won, but also — far more often — how many draws pass without a single hit. The second result is more instructive than the first.

What this means for a player

The honest conclusion is short: you cannot improve your chance of winning Lotto. You can choose less popular numbers so that any prize you win is split among fewer people. You can keep your own statistics for your own enjoyment, and because data analysis can be interesting in itself. What you cannot do is shift the 1 in 13,983,816 threshold by a single unit.