statistics

[Advocate]

Statistics are a measure of ignorance - how uncertain is your predictive power.

Using statistics to make actual decisions is gilded guess-work. Statistics aren't facts and they aren't usually knowledge (justified belief), they're guesses and approximations, however rigorously obtained.

[RCSaunders]

Using statistics to make actual decisions is gilded guess-work. Statistics aren't facts and they aren't usually knowledge (justified belief), they're guesses and approximations, however rigorously obtained.

The statistics themselves are not approximations. The statistics are factual representations of actual cases. It is the interpretation of statistics that are approximations and as you said, when used to make predictions or choices are nothing but guesswork. It's the whole reason induction is an invalid means to establishing anything, especially in science.

[Scott Mayers]

"Statistics" is only a guessing mechanism that is useful for practical purposes only. If I toss a coin, the two states possible of which we are concerned with are that it is either heads up or tails up. That we account for it using fractal numbers (fractions) is misleading. So the 1/2 probability of each often gets presumed to be 'precise' when it is only a statement about the possibilities if taken one event at a time.

They are abused all too often though when INTERPRETATION of what these numbers mean with respect to reality are imposed. Main Quantum Mechanics interpretations, for instance, have some that interpret reality itself as 'tossing dice' for a specific instance that assumes the instance itself is not actually repeatable. This makes such use arrogantly 'mystifying' and 'weird' in a way that does injustice to science and reasoning in general.

[RCSaunders]

"Statistics" is only a guessing mechanism that is useful for practical purposes only. If I toss a coin, the two states possible of which we are concerned with are that it is either heads up or tails up.

That is not statisitcs, that is probability.

Statistics would be if you tossed a coin 100 times and wrote done which way the coin landed each time. That record is a statistic, and is just a fact, the record of what actually happened. The attempt to interpret from that what the odds are that a coin will land in a particular state is called, "statistical probability," and is an interpretation.

It's a semantic thing. You have the principle right.

[Scott Mayers]

"Statistics" is only a guessing mechanism that is useful for practical purposes only. If I toss a coin, the two states possible of which we are concerned with are that it is either heads up or tails up.

That is not statisitcs, that is probability.

Statistics would be if you tossed a coin 100 times and wrote done which way the coin landed each time. That record is a statistic, and is just a fact, the record of what actually happened. The attempt to interpret from that what the odds are that a coin will land in a particular state is called, "statistical probability," and is an interpretation.

It's a semantic thing. You have the principle right.

Statistics is the induced parent class of math that includes probability as a major subset. I've never heard of one divorcing the relationship.? The collection of data is required to induce the probabilities. But both lack meaning without its use for PREDICTING the future and thus are only assured 'factual' where they are at the extremes, ....like 0% or 100% probable, for instance. This is itself an interpretation where we discuss what they mean with respect to things like the Copenhagen interpretation of Quantum Mechanics.

I've discussed this before here and elsewhere using the Monty Hall Problem and/or Bell's Theorem where used to justify whether the Copenhagen interpretation is or is not correct. I showed how the Monty Hall problem cannot speak of accurate representation without recognizing a 'hidden' cheat. The way the problem is resolved requires assuming validity without respect to the fact that it cheats by hiding essential information using this (not here) to show that Bell's Theorem, which is used to supposedly 'prove' the Copenhagen interpretation, is flawed because the math is just another form of the Monty Hall problem. The use of Bell's theorem was to attempt to prove that you can use a statistical operation to prove that there is NO 'hidden' factor to quantum entanglement changes at a distance. By showing the relationship of these two problems, one can show that the hidden factor has to be true in both cases to maintain consistency and why it could not prove the Copenhagen interpretation.

[I hold to the Everett Interpretation, or, with more inclusiveness, the "Many Worlds" interpretation. It justly permits parallel worlds to justify the slit experiments rather than the Copenhagen version which assumes that all probabilities exist but 'collapse' when observed into one specific instance.]

Statistics ARE useful. But to say that X occurs, say, 2/3 of the time would require proving that reality has all odds fair and only IF we have parellel worlds where taking all probabilities as 'possible' distributes these as occurring in 2 of 3 worlds occuring simultaneously.

[RCSaunders]

I hold to the Everett Interpretation, or, with more inclusiveness, the "Many Worlds" interpretation.

Well, that doesn't surprise me. I hold with the, "reality trumps all statistical probabilities," interpretation. Out of all the possible fictional worlds there might be, this is the one that is.