Statistics in develop process

[_Al3x]

Hi everyone

I'm making a solitaire game. I told a friend of mine that I would make a method to try to resolve a game. Then I'll make it resolve thousand games and make some statistics for the fun of it. He told me that statistics are useless in the develop process and asked me to answer why.

I can't find a proper answer other than if the method worked once, you don't need to see anything else. I mean, I fixed a deck so the method resolves a fixed game, so if the method is right, should work just fine, no need to test it a thousand times. Just once with a non fixed deck (or a fixed one with impossible solution).

That would be my answer, what do you think? 

[gimbal]

To me your reasoning doesn't answer the actual question in any way. Okay fine, then you know that your game can be solved. But why do you want statistics?

[_Al3x]

To me your reasoning doesn't answer the actual question in any way.

For me neither! That's the reason of the topic! Thank you for your reply!

I wanted statistics just for fun, to find out how many times out of 100 or 1000, the randomness of .shuffle() gave me a resolvable game. Again, JUST FOR FUN

But he told me that in GENERAL, statistics aren't usefull for the develop process. And that's the question, WHY

I know statistics may give you the average use of a software, so it MAY be relevant, but not for the design or develop stage of making a software.
What do you think?

Games published by our own members! Check 'em out!

[Jimmt]

Statistics ARE important. If your game only resolves 1 time out of 1000, it's not going to be very fun, is it? (although you may be able to get this stat from your code)

[_Al3x]

If your game only resolves 1 time out of 1000, it's not going to be very fun, is it?

True, you might need to tweak something here and there, but what about in a GENERAL talk? Not only games?

(although you may be able to get this stat from your code)

What do you mean? How would you do that?

Thanks for your reply!

[gimbal]

Statistics ARE important. If your game only resolves 1 time out of 1000, it's not going to be very fun, is it? (although you may be able to get this stat from your code)

Semantics I guess. You can call them statistics, but I'd rather call them test results EDIT: but thinking about it for more than half a second I do see the point actually. The fact that the game fails to be solved such a large number of times IS a useful statistic.