# A Probability Question, from Leo Jia

November 21, 2014 |

If

X= "I touched the stove"

and

Y= "I got third degree burns"

then I would NOT disregard the data set as "too small to be indicative".

## Larry Williams writes:

Yes, yes…one sample size is adequate–there seems to be a connection readily seen.

## Leo Jia replies:

My issue is when X is not exclusive to Y, I have not much clue on how X happens and whether Y has a rational reason to be linked with X. This can be possibly because if it is too clear it might very well be different later.

It is quite like a situation where I am a monkey in the zoo. Most of the time I am kept very hungry. The zoologists play a lot of games with me, delivering food here and there at times. During the last five years, I discovered this thing (which I have no idea what is but you humans call it "stove") 10 times in my play field. The 9 times I touched it, it was warm but not harmful and dispensed quite a lot of food, although one time it had no food and was hot so I got burnt then. I am not sure if this is part of the game, but I am clever enough to remember this.

I keep in mind that the stove was not the only thing I encountered that dispensed food. There have been a lot of other situations, one of which is that, quite frequently, you human spectators throw in bananas andvcandies though also often times with garbage which causes some real pain.

So in this case, how do I take into account the stove case?

# Comments

Name

Email

Website

Speak your mind

2 Comments so far

1. Stefan Martinek on November 23, 2014 7:53 pm

B.F. Skinner in “Science and Human Behavior” classifies Intermittent Reinforcement as more addictive than alternatives (positive/negative reinforcements). Most feedback loops around us have a random element and this is why the sample size matters at the end.

2. anonymous on November 24, 2014 12:04 pm

Don’t trust the statistics. Trust your risk point.

You will never know the answer to “My issue is when X is not exclusive to Y, I have not much clue on how X happens and whether Y has a rational reason to be linked with X.”