# How do you start solving a statistical problem? Topic - Goodness of fit criteria.

#### yuramoros

##### New Member
At the university, I was given the following task in statistics in topic Goodness of fit (and probability):

Previously, people often used tables of random numbers (for example, when they did not have access to a computer with the necessary statistical software). Such tables can be found in almost any textbook on statistics. M. Kadyrov's book "Tables of random numbers" contains 2000 four-digit pseudo-random numbers. The first digit turned out to be zero for 160, three for 247, six for 191, nine for 185 (the remaining 1217 numbers began with other numbers). Should I use such a table?
I don’t know how can I start solving it?

#### noetsi

##### No cake for spunky
Why would you use a table that has a limited number of values in it, when you can use a computer that can generate far more.

Computers don't truly generate random numbers, however. I don't know if the old charts in books (which I have never seen in any statistics book) were able to get around this issue or not. No computer can be totally random.

" The first digit turned out to be zero for 160, three for 247, six for 191, nine for 185 (the remaining 1217 numbers began with other numbers)."
I think the question here is how likely is it that you would generate this distribution if the results were truly random. Among other things you have to know what distribution you are drawing from.