statistics

probability effects



1.

probability does not self-correct. e.g. after a series of 10 tails, the probability of the result on the toss of a 

	coin is still 50% heads, 50% tails.


each event is independant, there is no reversal to correct for past events.




2.

random events such as tossing a coin, etc. do average out to close to the theoretical average over a large number of events,

	as an example 49.2% heads, 50.8% tails after 1000 events




3. random walks do __not return to the zero level. A movement that starts at one point and rises and falls randomly,

	tends to rise steadilty in stages or fall to large negative values in stages, it does not fall back towards

	the average



4.

if two sets of events are statistically related, it may be because both are caused by a third factor, it does

not necessecarily indicate that one set of events causes the other

	e.g. 

		hot weather

			-> high sales of summer clothes

			-> large amount of motor vehicle traffic



it could still be a relationship of one variable against the other though




5.

correlation



	+1	two data sets perfectly correlated, a large value in one set occurs when a large value in 

			the other set also occurs


	0	independant events, there is no relationship between the two sets of values



	-1	two data sets perfectly negatively correlated, a large positive value in the first data set is

			associated with a large negative value in the other data set.




Notes:


1. A tossed coin, as an example, is a spinning metal disk, it has no memory of past events, and the heads/tail outcome is 

	purely dependant on the which side the coin is facing up when it hits the table or the floor





