Standing by the school photocopier during a mid-afternoon break, I saw a 'guaranteed' lottery system ad on my phone that made me want to throw my coffee across the staffroom. As a math teacher, the blatant disregard for independent events felt like a personal insult. It is the kind of thing that makes me want to sit the entire internet down for a double-period lesson on probability. The ad claimed to have 'cracked the code' for the EuroMillions, a game where the odds of winning the jackpot are precisely 1 in 139,838,160. If someone had truly cracked a code with those odds, they would not be selling it for nineteen quid on Instagram.
I play the EuroMillions every week. My students would be absolutely delighted to know their 'sensible' math teacher spends a few pounds on Tuesday and Friday nights hoping for a share of a prize pool that can reach a maximum jackpot cap of 250 million Euros. But I do it with my eyes open. I know that the 50 main numbers and the 12 Lucky Stars—a range updated back in September 2016—are picked by machines named Stresa and Paquerette, and those machines do not have a memory. They do not care which numbers were 'due' to come up.
The Secret Notebook in the Desk Drawer
That afternoon in late winter 2025, specifically on one rainy Tuesday evening in February, I decided to stop shouting at my screen and start an experiment. I have always kept a basic frequency analysis spreadsheet—tracking which numbers appear most often—but the rise of AI-driven lottery platforms intrigued me. Were these neural networks actually seeing something I was missing, or were they just very expensive versions of my own bias?
I cleared out a corner of my desk drawer at school, tucked away behind the spare protractors and a box of confiscated fidget spinners, and started a notebook. For every draw over the next six months, I would run a parallel test. I would track my own 'hot and cold' frequency picks against the suggestions from three different AI lottery prediction tools. No more guessing; just data, documented with the physical scratch of a red pen on paper after the final bell had rung each week.
The Methodology of a Skeptic
To keep things fair, I followed a strict routine. Every Monday and Thursday, I would log into the AI platforms. These tools use historical draw data to identify what they call 'patterns' or 'clumping.' Meanwhile, I would update my own spreadsheet, looking at the same data but through the lens of simple frequency. I wanted to see if the AI could outperform the most basic math I teach my Year 11s.
Right then, here is the thing about frequency analysis: it is a record of what has happened, not a blueprint for what will happen. In the EuroMillions, you are choosing 5 numbers from 1 to 50 and 2 Lucky Stars from 1 to 12. Every time those balls drop into the tray, the probability resets. It is like flipping a coin. If you flip a coin and get heads ten times in a row, the eleventh flip still has a 50/50 chance of being heads. The coin doesn't 'remember' the previous flips, and neither do Stresa or Paquerette.
The Routine: Late Winter to Easter
By the week before the Easter holidays, my notebook was starting to look like the work of a beautiful mind—or a very tired teacher. I would sit in the quiet hum of my empty classroom, the smell of dry-erase markers still hanging in the air, and update the results. I was looking for any edge. If the AI suggested a 'high-probability' sequence, did it hit more often than my simple frequency picks?
I was thinking, 'I literally taught Year 10 about the Gambler's Fallacy this morning, and here I am looking for patterns in a plastic drum.' It is a strange cognitive dissonance. We want there to be a pattern because the human brain is a pattern-matching machine. We hate the idea that a life-changing amount of money depends on absolute, chaotic randomness.
During this period, I noticed that the AI tools were much better at avoiding 'unlikely' combinations—like 1, 2, 3, 4, 5—which technically have the same probability as any other set but are shared by thousands of people. However, in terms of actual hits? The results were stubbornly average. Whether it was the AI's neural network or my basic spreadsheet, we were all just dancing around the same statistical mean.
Why 'Hot Numbers' Are a Mathematical Trap
Here is a contrarian angle that most lottery sites won't tell you: focusing on 'hot' or 'overdue' numbers actually increases your expected loss. Now, that sounds counterintuitive, doesn't it? If a number hasn't appeared in twenty draws, surely it is 'due'? No. That is the heart of the Gambler's Fallacy.
When you focus on numbers that are 'hot' (frequent) or 'cold' (overdue), you are often playing the same numbers as thousands of other people who are following the same logic. In a pari-mutuel system like the EuroMillions, if those numbers do actually come up, you have to share the prize with more winners. Your expected value—the average amount you stand to win—actually goes down because the prize pool is diluted. You aren't increasing your odds of winning; you are just decreasing the amount you'll get if you do. I've written more about this in my notes on Probability of EuroMillions Patterns in Historical Draw Data.
The Turning Point: Exams and Realization
Mid-June arrived, and as exams started, my stress levels peaked. I found myself retreating to the notebook as a form of therapy. It was a world where numbers behaved, even if they didn't do what I wanted them to. One evening, after marking thirty calculus papers, I realized something profound about the AI tools I was testing.
The AI wasn't predicting the future; it was providing a sophisticated map of the past. It was excellent at identifying 'clumping'—the tendency for certain numbers to appear near each other in historical data. But clumping in a random system is an artifact of the sample size, not a law of physics. If you look at enough data, you will find clusters. It doesn't mean those clusters have any predictive power for the next draw on Friday night.
I am not a financial advisor, and I certainly have zero professional gambling credentials. I am just a teacher who likes a bit of fun. But it became clear that the value of these AI tools isn't in 'winning' the lottery—it is in managing the way we play. They help you pick numbers that are less likely to be shared, and they provide a structured way to engage with the game that doesn't involve picking your cat's birthday for the thousandth time. I documented this realization in The Spreadsheet in My Desk: My Semester Testing AI Lottery Patterns Against Pure Probability, which covers the specific draw-by-draw breakdown.
The Final Tally: Late July Reflection
By late July, as the school year finally wound down and the Manchester rain gave way to a few days of actual sunshine, I closed the notebook. Six months of data, hundreds of entries, and not a single jackpot. I wasn't surprised, of course. With odds of 1 in 139 million, I could run this experiment for ten thousand years and still not hit the big one.
But did the AI do better than my frequency charts? Marginally, yes—but not in the way you'd think. The AI was better at picking 'balanced' sets—odd vs. even, high vs. low—which tend to mirror the long-term distribution of random draws more closely than human 'random' picks. Humans are terrible at being random; we tend to spread numbers out too evenly. Real randomness is clumpy. It has streaks. It is messy.
The experiment taught me that honesty about probability is more valuable than any 'system.' I spent low-three-figures on tickets and tools over those six months, which I consider my entertainment budget. If you are looking for a way to pay your mortgage, the lottery is a terrible plan. It has a negative mathematical expectation, meaning on average, you will always lose money. You should always talk to a professional if you find yourself spending more than you can afford, as lottery games should stay as a hobby, not a financial strategy.
Final Thoughts from the Teacher's Desk
I still play. I still have my spreadsheet, and I still check the AI platforms occasionally. But I do it for the same reason I solve a crossword: it is a little puzzle for the brain. The AI tools are fun, and they offer a fascinating look at how we try to use technology to tame chaos. Just don't let anyone tell you they've 'guaranteed' a win. In a world governed by probability, the only guarantee is that the math will always win in the end.
Next time you see a flashy ad for a lottery secret, remember my rainy Tuesday in February. Remember that every draw is a fresh start for 50 balls and 12 stars. And maybe, just for fun, pick a few numbers that aren't 'hot.' At least if you win, you might not have to share it with the rest of the class.