Consecutive Pairs in 30 LOTTO 6/49 Draws: Counting Pairs and Draws

How can 30 draws contain 18 consecutive pairs, yet only 14 draws with a pair? Some draws contribute more than one pair. Even then, two pairs need not form a sequence of three numbers: only one draw in this window contained a run of at least three. The dated results below show how these counts fit together.

Analysis revised 2026-09-15: general examples were replaced with this dated 30-draw analysis. The original publication date, March 8, 2026, is retained; the June–September results were analyzed in this later revision.

The reading estimate covers the core text and ordinary table and example reading, excluding a full audit of the dated appendix and source links. It is an editorial estimate, not a measured reading time.

Which draws are included?

This analysis uses 30 LOTTO 6/49 Classic draws from June 3 through September 12, 2026, inclusive, based on saved official Loto-Québec results. The dataset is version 2 for that fixed window. Every supplied scheduled date is retained, with no filtering, replacement or missing values filled in. The window was inherited after collection; it was not documented as selected beforehand.

The Classic draw uses six distinct main numbers from 1–49, as described in the saved Loto-Québec game information. Only those main numbers enter this count. The separately recorded bonus, Gold Ball, Extra, Québec 49 and other auxiliary products are excluded.

The complete dated counting record links each observation to its official result in French. These saved sources support this historical window; this is not a latest-results feed.

What counts as a consecutive pair?

Sort the six main numbers from smallest to largest, then check the five gaps between neighbours. Count a pair only when the gap is exactly 1. Sorting describes numerical order, not the order in which the balls came out.

Call the number of pairs K. Its possible values are 0, 1, 2, 3, 4 and 5: K=0 means no pairs, K=1 means exactly one, and K=5 means five overlapping pairs across all six numbers. Overlapping pairs each count.

A maximal run is an uninterrupted sequence that cannot be extended using another main number from that draw. A run of length L contributes L−1 pairs. To count draws with at least one pair, count each qualifying date once. Apply the same once-per-draw rule when asking whether a run of at least three is present.

Can two pairs have different structures?

Yes. These are all four K=2 draws in the window. The dates link to French-language official results.

August 29, 2026: Main numbers: 16, 20, 23, 24, 25, 33. Pairs: 23–24 and 24–25 share 24. They form one maximal run, 23–24–25. K=2; run of at least three: Yes.
July 11, 2026: Main numbers: 5, 10, 11, 33, 34, 41. Pairs: 10–11 and 33–34 have no number in common. They form two separate maximal runs of length two. K=2; run of at least three: No.
July 25, 2026: Main numbers: 1, 8, 9, 24, 36, 37. Pairs and maximal runs: 8–9 and 36–37. K=2; run of at least three: No.
August 19, 2026: Main numbers: 6, 7, 10, 32, 33, 36. Pairs and maximal runs: 6–7 and 32–33. K=2; run of at least three: No.

August 29 shows why overlapping pairs count separately. July 11 shows why two pairs alone do not establish a run of three. These examples were chosen after observing the results to explain the classification, without claiming they are representative or recommended selections.

How many draws fell into each category?

The observed column counts these 30 draws. The theoretical column assumes that every unordered set of six distinct main numbers from 1–49 is equally likely, with no replacement. This uniform model is an assumption, not a finding proved by the observations.

On a narrow screen, scroll the table sideways to read every column.

LOTTO 6/49 Classic consecutive-pair counts, June 3–September 12, 2026; 30 draws, six main numbers only. Observed shares use 30 draws; theoretical probabilities assume uniformly selected six-number sets.
Consecutive pairs (K)	Observed draws / 30 (share)	Exact theoretical category probability
0	16/30 (53.33%)	7059052/13983816
1	10/30 (33.33%)	5430040/13983816
2	4/30 (13.33%)	1357510/13983816
3	0/30 (0.00%)	132440/13983816
4	0/30 (0.00%)	4730/13983816
5	0/30 (0.00%)	44/13983816

Shares use 30 draws and round independently to two decimal places, so the displayed percentages total 99.99%; the exact shares total 100%.

For example, K=3 occurred in no observed draw, but its model probability is nonzero. The same distinction applies to K=4 and K=5: absence in this window does not make a category impossible.

Why do 18 pairs correspond to 14 draws?

The ten one-pair draws and four two-pair draws contribute 10×1 + 4×2 = 18 pairs. Counting their dates instead gives 10+4 = 14 qualifying draws.

These answer different questions:

Pairs per draw: 18/30 = 0.6 pairs per draw, an observed average count, not a 60% chance.
Draws with any pair: 14/30 (46.67%) of the observed draws.
Draws with a run of at least three: 1/30 (3.33%) of the observed draws.

There was also just one maximal run of at least three in this dataset. A count of runs and a count of draws containing them are different units, even when their totals coincide. The 180 main-number appearances are not the denominator for these draw shares.

What does the theoretical comparison mean?

Each table fraction divides the number of sets in a category, the numerator, by all possible main-number sets, the denominator. C(n,j) means choosing j distinct items from n without regard to order; × means multiplication. Here C(49,6)=13983816 counts all sets.

For K=k, C(5,k) × C(44,6−k) counts the sets with exactly k consecutive pairs. Divide this by C(49,6) for the category probability. The counting idea is to group the six numbers into 6−k maximal blocks, including isolated numbers. The first factor counts possible block lengths; the second counts placements with unselected numbers separating the blocks.

For the question answered by the observed 14/30, the model comparator is P(K≥1)=6924764/13983816. Its numerator subtracts the no-pair sets from all sets: 13983816−7059052=6924764.

For the observed average of 18/30 pairs per draw, the corresponding model expected count is E[K]=30/49 pairs per draw. This is a model average count, not an event probability, a percentage chance or a quota for the historical sample.

A larger category contains more exact sets. It gives no particular combination a selection advantage. These category probabilities concern Classic main-number sets, not a complete ticket or winning any prize across all prize categories.

What can this window tell us?

It provides a dated, reproducible description of how pairs and runs were counted. The small inherited window does not establish long-term frequencies, representativeness, bias or a cause for the results. No significance test is being reported, and these observations prove neither uniformity nor independence.

The counts do not predict another draw or identify numbers to select. Their value here is that each total can be traced to the dates below, including every draw with no consecutive pair.

Dated results and counting record

Each date links to the official Loto-Québec result in French underlying the saved observation. Main numbers are sorted; the bonus is excluded. “None” means no qualifying pair or run, not missing data. Runs listed are all maximal runs of length at least two. “Run of at least three” records whether the draw contains one or more such runs.