EuroJackpot PyLab

Coding the lottery. Keeping it human.

C(50,5,4) Min-Covering - a “4-hit safety net” dataset

2025-12-15
Dataset: Download

What problem are we solving?

EuroJackpot main numbers are “5 out of 50”. The full space is huge, and random play is… well, random.

So here’s a spicy question:

Can we build a fixed set of tickets so that no matter what the winning 5 numbers are, at least one of our lines will match 4 of them?

No jackpot promise. No “I cracked the lottery” nonsense. Just a strong coverage guarantee.

The definition (plain language)

We want a set of 5-number lines such that:

  • for every possible winning set W (5 numbers),
  • there exists at least one line in our set that overlaps with W in 4 numbers.

In combinatorics terms, this is a covering design:

C(50, 5, 4)

What you get here

This page includes a CSV file containing 33,572 lines (5 numbers per line).

If someone plays the entire set, then for any winning 5-number draw:

  • you’re guaranteed to have at least one 4-hit on the main numbers.

That’s the deal.

Before you get carried away

A “guarantee” can mess with your head if you don’t keep it grounded:

  • It does not cover EuroNumbers.
  • It does not guarantee profit.
  • It can still be expensive to play at full size.

Treat it like a tool: useful, measurable, and very easy to misuse.

How to use the CSV

Each row is one line (ticket), usually as:

  • st1, st2, st3, st4, st5

Good uses:

  • compare this cover against your filters,
  • build smaller “slices” that still keep decent coverage,
  • study how coverage behaves on the 1..50 grid.

Why this belongs in the Lab Blog

Because once you have a cover set, you can start asking better questions:

  • Which filters destroy coverage?
  • Which filters keep coverage but cut cost?
  • Can we build “cover-aware” reduced pools?

That’s where the fun starts.

Here are the real stats for this covering (911 draws so far)

=== Norms (random tickets baseline) ===
Cover size : 33572
Total combos: 2118760
Norm p(>=1 five-hit): 1.585%
Norm p(>=1 four-hit): 97.251%
Every draw had at least a 4 number hit
Norm mean #4-hit lines per draw: 3.565

=== Original (real cover vs real history) ===
5-hit draws: 1.317%    lift vs norm: 0.831
4-hit draws: 98.683%    lift vs norm: 1.015
mean #4-hit lines per draw: 3.622    lift vs norm: 1.016