Data, records and curiosities
Consecutive identical draws triggered a ministerial investigation, a commission and a national argument. The commission cleared the lottery, and the second draw's eighteen winners explain more than the coincidence does.
Repeated numbers across a long history are ordinary. Repeated numbers in back-to-back draws are a different order of coincidence, and in September 2009 Bulgaria produced one.
On Sunday 6 September 2009, the Bulgarian national lottery — operated by the state Bulgarian Sports Totalizator — drew 4, 15, 23, 24, 35, 42. Nobody won the jackpot.
On Thursday 10 September 2009, the very next draw produced the same six numbers, in a different order.
This time there were 18 winners, each receiving about 10,164 leva from a jackpot pool of roughly 200,000 leva. A lottery spokeswoman told reporters it was the first such event in the lottery's 52-year history and that officials were "absolutely stunned to see such a freak coincidence" (BBC News, The Irish Times).
Three of the same numbers appeared again in the following draw on 13 September.
Svilen Neykov, Bulgaria's Minister of Physical Education and Sport, ordered an inquiry. A commission chaired by Konstantin Simeonov examined the draw.
Its conclusion, reported at a press conference on 17 September 2009 by lottery head Damyan Damyanov, was unambiguous: no manipulation. Simeonov's summary was "We cannot talk about any manipulation." The Associated Press report of the finding recorded no evidence of fraud (San Diego Union-Tribune / AP, Reuters).
Mathematician Mihail Konstantinov gave the odds of the same six numbers coming up twice running as about 4,200,000 to 1, and stressed that this was "not impossible."
The detail most coverage skipped is the most informative one. Deputy lottery chair Maria Yaneva ruled out manipulation and pointed out something obvious in hindsight: the record number of winners was driven by players who deliberately re-entered the previous draw's numbers.
That is a behavioural fact, not a mathematical one, and it demolishes the fraud hypothesis on its own terms. A manipulator rigging a draw wants few winners, ideally only their own tickets. Selecting the single most heavily-played combination in the country — the one thousands of players had just copied from Thursday's newspaper — maximises the number of people you must share with. It is the exact opposite of what the Pennsylvania fixers in 1980 did, and the opposite of what Eddie Tipton did. Real lottery fraud is quiet and narrow.
It also illustrates a point that costs ordinary players money every week: your number choice does not change your odds, but it does change who you share with. Eighteen people split a pool that one person would otherwise have taken. See how prize pools are split and the number sharing risk tool.
The probability that the next draw exactly repeats the current one is 1/M, where M is the number of combinations.
Reporting of the game's format is inconsistent. Bulgarian outlets described it as Toto 6 of 42; other accounts describe a 6-from-49 game, and the Bulgarian Sports Totalizator has historically run both. We could not confirm the format against the operator directly, so here is the arithmetic for each:
| Format | Combinations M | P(next draw repeats this one) |
|---|---|---|
| 6 from 42 | 5,245,786 | 1 in 5,245,786 |
| 6 from 49 | 13,983,816 | 1 in 13,983,816 |
Konstantinov's quoted 4.2 million is in the same region as the 6-from-42 figure, which is the more likely reading.
Now the question that matters. The chance is not "what were the odds of this happening on this night" — it is how often should a consecutive repeat occur somewhere, ever. Over D draws a game contains D − 1 consecutive pairs, so:
E[consecutive repeats] ≈ (D − 1) / M
Bulgaria's lottery had run for 52 years. At two draws a week that is about 5,400 draws, giving:
5,400 / 5,245,786 = 0.00103 — roughly 1 in 970
For that single game, over its whole history, a back-to-back repeat was about a one-in-a-thousand event. Genuinely unusual.
But there are hundreds of six-number lottery games worldwide, most of them decades old. A hundred games each accumulating 5,000 draws gives an expected count of about 0.1; several hundred games with longer histories pushes it toward and past 1. Over the world's lotteries and their combined centuries of draws, "at some point, somewhere, a game repeated itself consecutively" is not a stretch. It is why patterns are guaranteed: rare events become certainties once you count all the opportunities rather than one of them.
Note how much rarer this is than an ordinary repeat. For any two draws in a game's history to match, the birthday problem makes the threshold √M rather than M — about 2,700 draws for a 6-from-42 game. That kind of repeat is expected. A consecutive one is roughly two thousand times rarer, which is exactly why this case drew a ministerial investigation and the Israeli repeat of 2010 drew mostly astonishment.
Three independent reasons, none of them requiring you to trust the commission:
The case has since been used as a worked example in the academic literature on coincidences — Pollanen's 2024 paper A Double Birthday Paradox in the Study of Coincidences analyses the 6 and 10 September 2009 draws directly.
The Bulgarian repeat was a real and unusual event, properly investigated and correctly cleared. What it demonstrates is not a flaw in the lottery but a flaw in intuition: people reason about coincidences by asking "what were the odds of that", when the question that determines whether to be surprised is "how many chances were there for something like that".
Run it yourself in the draw simulator, or check a real draw history for repeats with the frequency analyser.
Last verified: 2026-08-29