Intuition testing - page 12
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
No problem, Urain. There is a library of statfunctions in kodobase, there is an inverse of Gaussian one, that's what we need.
It turns out simply: if we give a value uniformly distributed at [0;1], then by applying the inverse of normal distribution, i.e. Ф-1, we get N(0,1), i.e. a standard normal distribution.
Another thing is that it is mauvais ton to generate a normally distributed one like this.
Rosh has an article about another way of generating a normally distributed value.
Mathemat, I didn't expect you. After all, no one thought that 50 wasn't very much. I just increased it to 100, and voila, the result, as they say, is in your face! No tails, no tops, just a little bit on the side.
ZS. It may be tawdry, but it's a normal tawdry.Normal distribution
IlyaA, it is a mauvais ton because probably with large arguments (many sigmas) it is not so easy to approximate the inverse normal function. It is not elementary.
IlyaA, it is a mauvais ton because probably with large arguments (many sigmas) it is not so easy to approximate the inverse normal function. It's not elementary.
Wait a second, would you tell me the distribution is normal?Well, what's there to say. The blue bars are kind of close to the red ones. So it looks like normal. But you can't say anything for sure here, the most important thing is the behaviour in the area of big deviations.
Well, what's there to say. The blue bars are kind of close to the red ones. So it looks like normal. But there's nothing to say for sure, the most important thing is the behaviour in the area of big deviations.
I agree with the mauvais ton. And the behaviour in the >3 sigma range is VERY unlikely. Well where do you see a machine guess 80 numbers out of 100, for example. :) So it's all good here.
Wait a second, you tell me the distribution is normal?There's no point in arguing, there's a chi-square test, check it out.
Well, for a spherical horse in a vacuum, i.e. for a guaranteed normal distribution, yes, it is unlikely. Well, real returns aren't horses in a vacuum. There are usually 5, and 6 s.c.o., and even 10.
There's no point in arguing, there's a chi-square test, check it out.
Fight to the last. In short, the distribution is normal, or close to normal. I can put the data out there for you to check. I've already done one test. Now it's your turn.
Fight to the last. In short, the distribution is normal, or close to normal. I can put the data out there for you to check. I already did one test. Now it's your turn.I'm not beating myself up, I'm just suggesting an objective way of checking. I used to do such a check when I was just starting to study the markets, the result was negative at 0.85 significance level.
I'm not beating myself up, I'm just suggesting an objective way of checking. I used to do such a check when I was just starting to study the markets, and the result was negative at the 0.85 significance level.
>> OK. Here's the data.