I am using regression on lottery numbers, below is the head of my database. I was wondering if anyone had a suggestions on what the best algorithm would be to use for the regression model.

Data Head:

      Date  b1  b2  b3
1984-09-01   9   4   8
1984-09-03   1   5   3
1984-09-04   8   6   6
1984-09-05   5   6   9
1984-09-06   2   1   9
  • $\begingroup$ so what is x, y? $\endgroup$
    – Peter
    Jun 15, 2019 at 7:05
  • $\begingroup$ Balls 1,2,3 = Y, Date = X. $\endgroup$
    – user76091
    Jun 15, 2019 at 8:03
  • $\begingroup$ What makes you think that the lottery numbers depend on the date? ML is not magical, if the features don't have any explanatory power it's not going to learn anything. $\endgroup$
    – Erwan
    Jun 15, 2019 at 16:53

1 Answer 1


This answer is based on additional information provided in a comment:

what is x, y?
Balls 1,2,3 = Y, Date = X

Lotteries and other games (such as Casino games) are designed so that one is (or should be) unable to predict what happens. For instance in roulette, you could win a lot of money by applying the St Petersburg paradox. However, casinos restrict the sequence of playing so that you have no chance to do this.

WRT lotteries, the data generating process is supposed to be 100% random. This means that your X (the date) does not carry structural information about the outcome (y). Consequently, there is no way to make a reasonable prediction of what the numbers of next week would be.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy