Talk:General Discussion/@comment-25189003-20141009091951/@comment-25289127-20141009135320

Erm... no. I think you misunderstood me.

What I'm trying to say is that a computer does every calculation on its own. So general statistic doesn't work the same way it normally does. You can't say "I have a 95% chance so I should have success around 95 times" - at least not for computers. A computer does the same calculation 100 times with 95% chance of success each and that can lead to a different result.

Using a high chance on a single calculation will most likely result in a success and if you do that 100 times it's possible that a 95% rate actually results in a success 99 or even 100 times.

The opposite goes for low chances.

So it is possible that increasing your chance from 90% to 95% (so only 5% more) is more effective than increasing your chance from 15% to 25% (10% increase).

Of course it's not like that if you do normal statistical calculation but it can work this way for computer. Usually you'll say that 10% of course is more useful than 5%. But for computercalculation it doesn't necessarily work that way.

And apart from that, to make things even worse, random for a computer isn't even random.

To randomise stuff on a computer you use a mathematical formula. There is a way to "calculate" random results. Of course as you would expect a formula that "calculates" random results, it doesn't really generate truely random results. If humans made it based on math it can't be really random.

So it is possible that you can get the most stupid results.

For example if you let the program generate four random numbers from 1 to 10, sometimes it happens that you just get 1111 or 1234. Sometimes it even happens three or four times in a row.

With other words it's possible that a 99% chance of success still result in only fails. Sounds stupid and the chance may be really small but it may happen. (」゜ロ゜)」

The last time I noticed this is when I read about how people when they do expeditions with a 96% chance for great success pretty much have a 100% rate of great success. That's just how computerchances work. *facepalm*