Choose your own prize car

  • Thread starter Thread starter Touring Mars
  • 32 comments
  • 1,214 views
nah!
if you test with five times and dont get the car you want when you press the magic button all five times then it doesnt work.... getting 199 out of 200 doesnt make it work either... when is the right time to press the magic button? when the symbol is under the car you want seems to be popular... surely everyone here has enuff dexterity to do that! and if you miss the symbol you tend to go "D'oh i missed it" cause its easy enuff to see...

gumpy did a big test on this, you should read it.

so what i am saying is if its not 100% of the time that you press it at "the right time" then how could it be anything but random (or 1 in whatever the number of outcomes are).
 
Yes, I concede that it is not a perfect method, but to say that it should work 100% of the time or not at all is wrong, simply because it still involves an element of skill, and sometimes you press X too early or too late, so a success rate of 100% is unlikely...
 
I did read Gumpy's test. It was bollocks. There was no negative control (ie 200 times of doing nothing). You need a negative control to test the H0 hypothesis and to compare the results of pressing the "magic button".

As Mars says, humans are fallible - there is operator error. The effects of this can be analysed correctly with a ROC curve.

I'm doing a negative control at the moment - 200 times of doing nothing. The current results, whilst only 1/4 complete, are interesting. It reads:
Top Car: 16/50 (32%)
Right Car: 9/50 (18%)
Bottom Car: 7/50 (14%)
Left Car: 18/50 (36%)

This is the importance of the negative control... Assume you're trying to get the left car - pressing X works 36% of the time. Hey, that's more than 25%, it works!

I'll keep you all posted.
 

Latest Posts

Back