A visual demo of Probability

[packman]

Please do not be mean, I didn't sine up for this

Its called Galton Board. Every single particle released from the top has equal-ish movement ability, where the edges are "harder" to go (you need several bounces) and directly below is "easier" (you only need a bounce or two).

After a few thousand trials (every single grain of sand is an instance) we find that they end up almost exactly how math predicts that they would (the wavy line is that prediction).

Note that if we had say even more grains, for instance 10X more, the final result would be even closer to the line (you'd see less outliers etc).

[zipplewrath]

1. Chicago Museum of Science and Industry

They had a large display that worked like this. Thousands of marbles were released and would fall, ultimately stacking up in a large display case with the "bell curve" painted on it. Each drop took about a minute, and over and over it would stack up in the same basic curve. My father was particularly fascinated by it.

It inspired a science fair project for my sister. She got a blue die and a green die and would roll them over and over, recording each result of each die, as well obviously as the combined result. Ultimately the results were as expected. Each die rolled each number an equal amount of time, but 7 was the most common combination. The interesting result to me was that it really wasn't until about 3000 rolls that the plots really started to "look" right. For a long time, there were way too many 2's and 12's popping up.

Years later in college this lesson was sort of reinforced for me by a professor that published a paper about true randomness. It is always a problem apparently in large data sets to determine if the data set is truly random or is there bias of some sort indicating an underlying pattern or trend. The professor's paper determined that in fact if a data set is truly random, certain "patterns" should appear within it. Depending upon the size, patterns like "straights" or ordered sequences should appear, and with fairly predictable frequency based upon the size of the data set. When we see deviations in the predicted probabilities before the necessary population size is reached, that's really reinforcing that there will be patterns inside the randomness.

Forensic accounting uses a similar principal in that in large financial reports or other "books" certain digits should appear with more or less frequency than others. Counting all the individual digits used in the report can reveal whether the books have been "cooked".

[packman]

2. That was used in one of my favorite movies - "The Accountant"

Forensic accounting uses a similar principal in that in large financial reports or other "books" certain digits should appear with more or less frequency than others. Counting all the individual digits used in the report can reveal whether the books have been "cooked".

[IrishAfricanAmerican]

3. My wife, an accountant, loves that movie...

she never cared for the action genre. I think she liked how focused and efficient the character was. Plus the surprise twist at the end.

I thought it was Affleck's best role. I always found him a stilted and unemotional actor. Playing someone "on the spectrum" worked perfectly for him. I was hoping there would be a sequel.

[lastlib]

4. At first, I thought you were going to go off on another tangent......

I'm glad you didn't cos this really tripped my trig--ger.

[packman]

5. Oh, I thought mine were bad-but, here you were plotting all the time

[lastlib]

6. I have a knack for blowing the curve.....

It's my skewed mind-set, I think.