Yeah, I'm a Bulgarian, and I remember that... And you know what's even funnier? The second turnaround of the numbers had higher number of winners, because a few people had actually CONTINIOUSLY (not just between those two draws, but for a long time before that...) had used EXACTLY that strategy - they play the same numbers as whatever the previous weeks' numbers were, in the hope that by some odd chance, that would happen... and it actually did in that one instance!
Given that the duplicate lottery draws were back to back, this has nothing at all to do with the birthday paradox. If the odds of the lottery are 1 in 14 million, then the odds of a single draw duplicating the previous draw are 1 in 14 million. Period. The birthday paradox is completely irrelevant. The duplicate draws were on September 6 and 10, which tends to suggest 2 draws per week. It is noted elsewhere that the lottery was 52 years old at the time. So the total number of draws might be as high as 2 draws x 52 years x 52 weeks/year = 5,408. Dividing the overall odds by this number gives us odds of about 1 in 2600. That still seems unlikely, but then, as you correctly point out, one has to consider all the lotteries all over the world. So, basically, when one considers all the draws that have ever been held, over all of the lotteries in the world, the law of large numbers applies, and it is not all that surprising that this could happen somewhere. When considering just the single draw, though, it's an amazing coincidence.
