Do consecutive numbers avoid each other? The 49.5% surprise

Ask a lottery player whether 23 and 24 are likely to come up together and you'll usually get a wince. Consecutive numbers feel non-random — too tidy, too coincidental. So most players quietly avoid them. Here is the delightful truth: in a 6/49 game, about 49.5% of all draws contain at least one consecutive pair. It's nearly a coin flip. Let's derive it properly, because the derivation is one of the prettiest small arguments in lottery mathematics.

Setting up the count

Total number of possible 6/49 draws:

C(49,6) = 49! / (6! × 43!) = 13,983,816

We want P(at least one consecutive pair among the six numbers). "At least one" problems are almost always easier through the back door: count the draws with no consecutive numbers, and subtract from 1.

The bijection: squeezing out the gaps

Claim: the number of ways to choose 6 numbers from 1–49 with no two consecutive is exactly C(44,6).

The standard argument (a cousin of stars-and-bars): take any valid non-consecutive selection a₁ < a₂ < … < a₆ and define

b₁ = a₁, b₂ = a₂ − 1, b₃ = a₃ − 2, b₄ = a₄ − 3, b₅ = a₅ − 4, b₆ = a₆ − 5

— that is, bᵢ = aᵢ − (i − 1). Because each aᵢ₊₁ ≥ aᵢ + 2 (no two consecutive), each bᵢ₊₁ ≥ bᵢ + 1: the b's are strictly increasing, i.e. six distinct numbers. And since a₆ ≤ 49, we get b₆ ≤ 44. So every non-consecutive selection from 1–49 maps to an ordinary 6-number selection from 1–44.

The map reverses perfectly: given any six distinct numbers b₁ < … < b₆ from 1–44, set aᵢ = bᵢ + (i − 1) and you recover a non-consecutive selection from 1–49 (adding back the "spacers" pushes neighbours at least 2 apart). One-to-one both ways — a bijection — so the two collections are the same size.

Intuition for the stars-and-bars view: choosing 6 non-consecutive numbers from 49 is the same as arranging 6 chosen slots and 43 unchosen slots so that a mandatory unchosen "spacer" sits between each pair of chosen slots. Glue one spacer to each of the first five chosen numbers and you've used up 5 of the 49 positions in advance, leaving a free choice of 6 slots among 44.

The arithmetic

C(44,6) = (44 × 43 × 42 × 41 × 40 × 39) / 720

Numerator: 44 × 43 = 1,892; × 42 = 79,464; × 41 = 3,258,024; × 40 = 130,320,960; × 39 = 5,082,517,440. Divide by 720:

C(44,6) = 7,059,052

So:

P(no consecutive pair) = 7,059,052 / 13,983,816 ≈ 0.5048

P(at least one consecutive pair) = 1 − 0.5048 ≈ 0.4952 ≈ 49.5%

Nearly half of all draws. If you've watched a 6/49 lottery for a few weeks and noticed consecutive numbers "keep happening", congratulations: you observed mathematics operating normally.

Why the intuition fails

Our brains judge randomness by appearance. A draw like 8-19-24-31-38-45 looks random; 8-19-24-25-38-45 looks like a glitch. But the machine doesn't draw "a pattern" — it draws six balls, and with six numbers scattered across only 49 slots, the average gap between adjacent chosen numbers is about 43/7 ≈ 6.1 (43 unchosen slots split across 7 gaps), so gaps of size zero-between-neighbours are common. Spread-out, "random-looking" tickets are the aesthetic minority dressed up as the norm.

This is the same perceptual bug behind believing shuffled playlists aren't random when the same artist plays twice. Genuine randomness clumps. Smoothness is what design looks like.

The payoff: other people's wince is your margin

Here's where this stops being trivia. Every combination has probability exactly 1/13,983,816 — playing 23-24 together changes nothing about your chance of winning. But lotteries are parimutuel: jackpots are split among everyone holding the winning line. Your expected payout therefore depends on how many other people play your numbers, and other people demonstrably avoid consecutive picks.

This isn't folklore. John Haigh's analysis of UK National Lottery play (JRSS Series A, 1997) documented that player selections are far from uniform, with recognisably "random-looking" spread-out choices heavily over-represented, and concluded that deliberately choosing unpopular combinations increases the mean return. A study of manually picked tickets in the Israeli national lottery (Judgment and Decision Making, 2010) likewise found strong systematic patterns in what players select — human picks cluster on birthdays, lucky numbers and visually "balanced" lines, and shun the tidy-looking ones.

So a line containing consecutive numbers has the same 1-in-13,983,816 chance as any other, but — if the avoidance holds for the numbers you pick — fewer expected co-winners if it hits. Identical odds, better conditional payout. That's not a system; it's just declining to pay the herd tax. You can estimate the crowding on any line with our number-sharing risk tool.

Check it against reality

Pull up the draw history on our statistics pages — UK Lotto, Saturday Lotto, Mark Six — and count draws with consecutive pairs. Over any decent stretch you'll find close to half, right where C(44,6)/C(49,6) says it should be. (For games with other formats the exact figure shifts with the same formula: for choosing k from n, P(no consecutives) = C(n−k+1, k)/C(n,k).) The lottery keeps producing consecutive numbers at the advertised rate, players keep flinching at them, and the two facts together are the closest thing to an edge this game contains.