<Edit> I thought I'd make a more detailed critique of Twyst's post instead of just a simple "no." </Edit>

Your methodology is flawed. You're not looking at the right thing. You include a lot of erroneous data. Allow me to critique your test.

Twyst wrote:

I tested 5 people 100 times.

Did you test five people a hundred times each for a total of five hundred tries or did you test five people for a total of one hundred tries?

Twyst wrote:

68% of the time people will pick either the center or the right side box (or left if their left handed)

Which box they take does not matter. This data set is irrelevant.

Twyst wrote:

50% of the time they picked the other box after removing one false box

25% of those times they picked the correct box

This is irrelevant. You're only testing how likely the box they do choose is correct.

Twyst wrote:

50% of the time the correct box was picked the first time

So the experiment shows that you have a 50% chance no matter which box you chose at first. A third box is irrelevant and is the true illusion, making you think you have a 1 in 3 chance. However, if you decide to switch boxes, your chances of picking the correct box is reduced to 25%.

First, your math is flawed. If half the time the subjects switch boxes and when they switch boxes 25% of the time it's the correct box, that does not mean that the other box is right only 25% of the time. In order to figure that out, you'd need to determine how often the first box was right. If the first box is right 50% of the time, when they kept the first box it would also have to be right only 25% of the time. Your math simply doesn't work out.

Second, your numbers are too neat. Real probability doesn't come out like that. Flip a coin one hundred times. You will almost never get fifty heads and fifty tails. Your total will be close to the optimum fifty, but rarely bang on. This is a good allegory as you claim each box has a fifty percent chance of being right. Even if both boxes had a fifty-fifty chance of being right, the chance it would show up as

exactly 50% is 8.8817841970012523233890533447266e-16-to-1 against. In case you're wondering, that means no way in hell.

Third, the numbers you have given are very unlikely statistically, but they

are very psychologically compelling. They appeal to human tastes. 50% is exactly half the total possible. 25% is exactly one quarter. That you claim to have tested five people a hundred times is less compelling in this way as this is one part of the experiment that you admittedly made up and thus would naturally be prone to the vagaries of human though patterns. None the less, when taken in context with the other numbers, they are suspect.

Finally, I am forced to concluded you made this data up. You did no real tests and simply fabricated the supposed results. Your math that you use o draw conclusions is incorrect. The probabilities you concluded did not equate with the data you posted. The numbers you listed show clear signs of fabrication. This would explain why they don't mesh with your conclusions. Simply put, you tried to fix both the data set and the conclusions but did not think to make them agree. The probability of the data you suggested showing up, even if there was a probability of fifty-fifty, is astronomically small. When viewed as a whole the only conclusion I can reach is that you are a lying liar who lies, and lies poorly at that.

Having concluded that you are a liar, nothing you further say in this matter can be admitted. Everything you say from now on will be suspect as, having been lied to by you once, there can be no trust that you are not lying now. One can only assume that if you were to post corrected data from another supposed test, that there would be a strong possibility that you merely used this critique to generate a more compelling lie the second time around. Simply put, STFU & GTFO.