The sniper fallacy

The sniper fallacy lurks wherever you are dealing with a comparatively big amount of data and it does not evaluate clean sufficient. In this fallacy, data is first tested on many things and only after that, a hypothesis is formulated. This is then selected the same way that it fits the data tested.

Toss

Let us take a simple example of this fallacy: At a school with 1,000 students, we let each student toss a coin ten times. For every roll of “number” gets a student a point. The probability that an individual students 10 times in a row raises number is about 1 to 1,000. On average therefore, likely one of the thousands of students ten times throws numbers and gets 10 points. A sniper fallacy It would be to this one pupil in retrospect to be chiefly talented coin thrower since he has thrown the most points.

poker

It was purely by chance that these students threw ten times the number and outcome is not an indication on the coin toss talent of the student to. This coin toss example is still obvious, since in advance is quite clear that the outcome of a coin toss is random. But the sniper fallacy comes just at far less obvious things to fruition, as the following example shows:

The sniper fallacy in practice: leukemia and power lines

One of the best-known examples of such a fallacy comes from a Swedish study from 1992. This study examined over a period of 25 years, all Swedish households, who lived near a power line in a variety of diseases. In the end it turned out that children who lived near high-voltage lines, about four times as often with leukemia, like other children. A very effective press conclusion of the study was quickly found high-voltage lines lead to a dramatically increased leukemia risk in children.

As a result a lot of money was spent on root cause analysis and, countless other studies on the subject in order. However, was able to confirm the link between leukemia and power lines, no further study. In all subsequent studies the children showed the same leukemia rates as children who did not live in the vicinity of such lines under power lines.

 

http://resipsapoker.com/wp-content/uploads/2015/06/poker.jpghttp://resipsapoker.com/wp-content/uploads/2015/06/poker-150x150.jpgSteven HuffmanOnline PokerPoker Sucksfallacy and poker
The sniper fallacy The sniper fallacy lurks wherever you are dealing with a comparatively big amount of data and it does not evaluate clean sufficient. In this fallacy, data is first tested on many things and only after that, a hypothesis is formulated. This is then selected the same way...