Chapter 2

Ask robots for lottery numbers, and tickets repeat

A toy shop. A robot behind the counter has a speech bubble holding a green ticket with six dots. On the counter stand an open cloth bag with balls in it and one ball beside it, and a row of five tickets: four green ones with the same dots and one white one with other dots. A child looks on, and a man with a clipboard takes notes.

Imagine the toy shop runs a lucky draw. A bag holds balls numbered 1 to 49. You pull out six, and those six numbers are your ticket. Now you ask the shop robot for six numbers, again and again. Its answer changes, so it looks as mixed as the bag.

Your customers can do the same with AI. They can ask one question many times, and answers that change can look like a wide choice.

So I tested how wide the choice is, with lottery numbers. It is a plain number question, with no company, product or purchase in it.

The short version. A robot’s answers can change from one try to the next and still keep landing on a few choices. So if a report says the answers varied, ask two things. How many different answers were there? How often did the top one come up?

What I did

A fair bag is a useful yardstick. The published Polish Lotto counts looked much like one at this scale.

A fair bag almost never repeats a ticket

I calculated this yardstick; I did not draw from a real bag.

703 small boxes in rows. One is green and 702 are white.
Take 200 tickets from a fair bag, and about 1 round in 703 holds a repeat

Picture it. Each small box here is one round of 200 tickets from the bag. In the green round, some ticket came up twice. In the 702 white rounds, every ticket was new.

At Giulia’s: like 703 months of 200 invoices each, with two matching invoices in about one of those months.

Every robot kept returning to a top ticket

200 small tickets in rows. 136 are green and 64 are white.
One robot gave the same six numbers in 136 of 200 answers

Picture it. Each small ticket is one answer from that robot. Green tickets all hold the same six numbers. White ones hold others.

At Marta’s: like 136 of 200 parcels going out with the same six items inside.

None of these tickets is a lottery tip. They show what the robots repeated.

Answers can change and still stay narrow

Eight piles of small tickets side by side. Two tall green piles hold 101 and 86 tickets. Six short white piles hold 4, 4, 2, 1, 1 and 1.
One robot used 8 different tickets in 200 answers, and 2 of them made up 187

Picture it. Each small ticket is one answer, and each pile is one set of six numbers. The 2 tall green piles hold 187 answers. The 6 short white piles hold the other 13.

At Samira’s: like filling 200 ad slots from 8 different ads, with 2 of those ads taking 187 slots.

These answers did change, and a few of them side by side would look mixed. A bag can be uneven and still work by chance. Fill one with many balls that carry the same few numbers. Each pull is still chance, and those numbers keep coming up. So change on its own is no proof of a wide choice.

The words you use can change the count

One robot’s most common ticket appeared in 4 of 50 answers. With the word “lottery”, its most common ticket appeared in 36 of 50.

Picture it. Each small ticket is one answer from that robot. Green tickets hold the top set for that wording. The two green sets differ by one number.

“Give me six different random numbers between 1 and 49. Answer with the six numbers only, separated by commas.”50 small tickets, 4 of them green.4
“Pick six lottery numbers between 1 and 49.”50 small tickets, 36 of them green.36

At Joost’s: like a helper on a book-fair stall. Ask for “a novel” 50 times, and the top title comes up 4 times. Ask for “a prize novel” 50 times, and the top title comes up 36 times.

Do company names behave like these numbers? I have not run this test on company questions. The next chapter counts names in another way.

How to check this for your business

Portrait of Marta, who runs an online shop.Marta KowalczykOnline shop, PolandMade-up example

Marta gets a report on robot answers about gift shops. It says the answers changed each time, so many shops have a chance.

Marta asks two questions. How many different answers came back? How often did the top one come up? The counts will not tell her what buyers do. They will show her if “changed each time” meant many answers or a few.

Try this on Monday, with one report

Take any report someone sent or sold you about robot answers, and pick one question. Write two counts: how many different answers or names came back, and how often the top one came up. If the report gives neither, ask for both.

No report? Use your sheet from chapter 1. Count how many different names came up and how often the most common name appeared. Keep each question and robot separate.

References

Chapter 2. Ask robots for lottery numbers, and tickets repeat

Preprint. Dmitrij Żatuchin (2026). Ask a Language Model for Lottery Numbers: Concentration in Repeated Six-of-49 Outputs. arXiv:2610.00052, version 1. arXiv:2610.00052v1. Data and code. github.com/Rankfor/rankfor-open/tree/main/research/lotto-models.

What does AI say about your business?

Put the book into practice. Tell us your company and website to request an AI visibility audit. We’ll reply by email to agree what to check.

What the audit shows

  • Where your company appears in the answers, and next to whom
  • The pages AI cites in your market
  • What to publish first on your own site

Fields marked * are required.

Who is it for?

Your choice about cookies

We use cookies this site needs to work, and others only if you say yes. Choose what you allow, and change your answer whenever you like. Read our privacy policy