A personal opinion on statistics inspired by the excellent Kos diary "Vaccinate Yourself Against Statistics" which I recommend follows. Yeah, following is just my opinion, based on that dwem LaPlace's writing on probability and a stubborn streak about real facts.
A Snarky Treat(ise) on Statistics
Winning Powerball versus Committing Suicide
Powerball Odds ***Compared to U. S. Committing Suicide Odds ****
175,223,510 24,530 times more likely to commit suicide than win Jackpot
5,153,632 721 times more likely to commit suicide than win $1 million
648,975 91 times more likely to commit suicide than win $10,000
19,087 2.67 times more likely to commit suicide than win $100
12,244 1.71 times more likely to commit suicide than win $100 red ball
706 (only) 10 times more likely to win $7 with red ball than commit suicide
360 (only) 20 times more likely to win $7 without red than commit suicide
110 (only) 65 times more likely to win $4 without red than commit suicide
55 (only) 130 times more likely to win $4 with red ball than commit suicide
"LIkely" based on 14 suicides per 100,000, the tenth leading cause of death in US.
Suicide rate from http://www.bbc.co.uk/...
Which reported U.S. Center for Disease Control Statistics. May 2, 2013
Powerball odds from http://www.powerball.com/...
Calculations as follows: 14 per 100,000 equals 1 per 7,143 persons commit suicide.
Comparison of odds = powerball odds (per 1) divided by suicide odds of 7,143 (per 1).
Huh? In addition to the obvious, shows that comparison of statistical subjects has close to zero meaning for an individual, a way in which statistics mislead. For example, for a person who will definitely not commit suicide the comparison of statistics is invalid, his number will not change them. The powerball odds stay the same, the suicide odds stay the same. If he does win powerball then his odds of winning were 100%, there was just no way to know that in advance. If the individual does commit suicide, his odds were 100% in retrospect. At some point that person knew in advance.
What are the odds that the sun will rise tomorrow? Rise versus not-rise, like flip of a coin, 50/50. I hear in the Anthropology Museum in Ciudad Mexico that the Aztec priests realized this 50/50 statistic and prayed without ceasing, that the sun would rise tomorrow, which worked, as the sun kept rising. Heart rending. Same as with suicide as above, there are other factors than the 50/50. Statistics are to inform opinions, not twist them.
As seen here, comparing statistics does not work. The struck by lightning trope among thousands of other stats "more likely" is B.S.. When do stats work? In my opinion, only on an agreed common base, like cards, then only in a special way of guessing. Same for surveys, sports. Now if they rise the same way every day...for awhile...same base...
One person believed that if she has a certain gene her "risk of getting alzheimer's 'increases'. Huh? Wait a minute. No, no, some reasoning missing here. You can figure it out, and other "stats" too.
So more people have been killed by falling sandcastles than sharks? Did I get that right? Hm, if true what am I to conclude? Something fishy here. Some can check that out, saw it on BBC net. I won't bother.
May your chi-square be one or two dots. May your sun rise tomorrow.