Intuition testing - page 16

 
Urain >> :

How do you generate normal from a uniformly distributed generator?

Add six Random's - you get close to normal. If you're interested, you can fit a theory to it.

:)

As for the development of intuition - I haven't seen anything better than the one in the trailer. Extract-concentrate, like... :)

Files:
 

Another interesting thing is the game with variable rules (see attachment). Just when you get used to one strategy, the rules change...

I was once (2005) hooked by its strange and surprising ability to evoke a kind of special "pinching" feeling which has never been aroused by games with a fixed set of rules.

As I understand it now, this reaction arises in life when there is a significant change in the context of life, when old patterns no longer apply.

// Like a divorce. Or a change of boss. Or a forex market during a crisis. :)

That is, it's specifically an emotional reaction to the realisation of the need for "global" retraining. A kind of "existential crisis". :)

At that time I was not even jokingly keen on the idea of games with variable rules. (And even now I am not indifferent to this theme, both theoretically and practically).

Since then, I have repeatedly tried to come up with and make something like that. Not seriously, but just in between. :)

I even implemented one game in the form of a program (in Delphi). I thought to post it here too, but hesitated: it's too big, rar-archive is more than mega (1197K).

Or may be you want to post it here? // Toy for superintellectuals, Rubik's cube is nervously smoking in the kitchen ;)

 
MetaDriver >> :

Or should I post it?


Of course you do. It's a shared affair. In a pinch, on Rapida.
 
MetaDriver >> :

Add six Random's and you get close to normal.


Put up the pictures. Look at the way the posts are designed here. If you have an idea, "show" it.
 

За 2-3 недели интенсивных тренировок колоду карт можно затереть вдрызг до такой степени и запомнить, что "угадать" значение карты по рубашке получится даже в полной темноте на ощупь.

It wasn't that one that was all right. I've been keeping an eye on it. And they don't look or feel any different. If only for the simple reason that the cards can be changed every time. Who makes you use the same ones? )


Can you be more specific?

What's more detailed than that? You take any three cards, memorise them and aim to either guess which card is pointed at or guess which of them is the pre-ordained one.

 
MetaDriver >> :

Another interesting thing is the game with variable rules (see attachment). Just get used to one strategy, the rules change...


The game is okay in principle, but has nothing to do with intuition. The rules change quite rarely and all the combinations are scored in advance. So it's enough to remember the sequence. For this game to be related to intuition, the rules have to change every time and randomly.

 
MetaDriver >> :

Add up six Random's and you get close to normal. If you're interested, you can fit a theory to it.


Was there a test on that statement? I checked it out, it's too narrow.

 
Urain >> :

Has there been a test on that statement? I checked it out. It's too narrow.


I mean, yes, that's what I'm saying. The man came up with the idea and others did the work for him.
 
Urain писал(а) >>

Как сгенерить нормальное из генератора равномерно распределённого?

MetaDriver
wrote(a) >>

Add up six Random's and you get close to normal. If you're interested, you can fit the theory to it.

You can only get a semblance of left-handedness that way. :-)

In order to generate any distribution from a uniformly distributed series of RNG it is necessary to first construct (or take in analytical form) its distribution function. And not the probability density function, but the integral PDF = F(x). It is known to vary monotonically from 0 to 1. Then we need to divide this interval into as many equal intervals as the RNG can generate. Note that points 0 and 1 must be excluded because these points correspond to infinite values of the variable x.

For example, a meta-quota RNG generates 32768 values (from 0 to 32767). Then, dividing the interval [0;1] into 32769 sections, we obtain that the value of RNG 0 corresponds to the point 1/32769, while the value of RNG 32767 corresponds to the point 32768/32769.

Now we can generate a uniformly distributed series and for each obtained value of RNG put F(x) = (RNG +1)/32769. Using the distribution function F(x), we can find the value of the argument x by its value. These values form a random series with the distribution F(x).

By selecting a suitable scale for the partition of the Oh axis, we can obtain the required values for x.

PS

It should not be forgotten that the RNG gives a discrete set of values, while F(x) is a continuous function. Therefore, all this will work correctly only if the Oh axis (or rather the interval from X1 to X2 where F(X1)=1/32769 and F(X2)=32768/32769) is appropriately partitioned into a discrete set of equal intervals. The lengths of these segments will precisely determine the required scale to translate the values of x to the one you want.

 

I'm sorry, I couldn't resist.

And on the subject I advise to read Norbekov. The Fool's Experience or the Key to Enlightenment seems to be in other books of this series about training intuition.