Themis;373611 Wrote: > You need to sit down and think this through - you're very confused about > it.
Themis, part of what I do is teach statistics to Ph.D. students. > So, lets say the following: > 50% of the time the coffee is the same. > 50% of the time it is not the same, but half of this time (25% of > total), the CafMart makes a better coffee. > The audience identifies with 100% confidence the "better" samples. > > So ? We have 200 correct samples identified. Dunno where that "200" came from. But indeed, in the scenario you've specified the two coffee makers have identical performance, and so obviously the results are 50/50. In any scenario in which they do not have identical performance, the results will not be 50/50 (contrary to what you claimed previously). > More seriously : there are too many unknown variables in the test. The > 200 correctly identified samples don't tell us anything. But not > because they are 50% of the samples: they tell us nothing simply > because there are too many variables. No Themis, that is false. The test result tells us a lot - it tells us with what confidence we can reject the hypothesis that the two sources were identical. Everything in science - the whole edifice of scientific knowledge, technology, and innovation - rests on your "nothing". If people with attitudes similar to yours were in charge, we'd still be living in trees. Instead, we have nice stereo systems and slim devices to listen to. -- opaqueice ------------------------------------------------------------------------ opaqueice's Profile: http://forums.slimdevices.com/member.php?userid=4234 View this thread: http://forums.slimdevices.com/showthread.php?t=56712 _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/audiophiles
