from green to black

the probability fallacy

take 30 men, make them do a test. record the results. plot
a graph. assume a normal distribution. interpret. voila!

results show a 2 peak, or fat tailed graph. the peaks are
10 seconds apart from each other. the fat tail consists
only of left rowers.

what inferences can be derived?

current is consistently resulting in 10 seconds of headwind
to right rowers. or current results in left having 10seconds
of extra help. (1)

left is stronger than right. current bias is negligible and
the results are sound. (2)

i will argue in favour of premise (2) for the 2 inferences,
noting that sample size is NOT random and that 10seconds is
a significant disparity.

conclusion, bad boat drift
recommendation focus on the essentials of a quick pull.

go figure the argument.

if all left should add 10secs because current not negligible
then left rowers like ian is clocking 2:29 and glen 2:33
they must be damn lousy and unfit.
- conclusion is improbable.

if right should subtract 10 seconds then those who did above
2:30 will no go below the average of 2:20 anyway. they will
still be weaker than left on timing.
- constants do not explain differences, if changing the result
does not change the overall results, change is not significant.

what is the likelihood of a 10second advantage current?
- in singapore, in a closed river system/reservoir. rofl!

and nobody wants to change, or believes they need to change yet.

posted by green unseen @ 9:43 AM