YQ, I didn't say you were wrong - just take care using
that as a reason.
And for my views on QM - I've never said it wasn't an accurate
predictive framework - when I say it's "market research", you appear to react as if
that is not an accurate framework. Statistical modelling is actually quite effective when specifics are unknown.
I should know - my wife's a program director of a rock radio station, and there are great parallels between her methods and QM. You cannot predict the behavior of a specific listener. In fact, if you "measure" a specific listener, you screw up the state - people who
never listen to you claim they do, and people who
always listen to you claim they don't. And no matter who you talk to, they all
hate the music.
But oddly enough, when the statistical models are applied (ala QM), the numbers come, exactly as predicted. If you own several stations in a given city - adding one single song to one of your stations will cause a predictable shift in all stations - drawing numbers
from some, and pushing numbers
toward others, and you can even predict what gender and age the shift will be in.
Quite literally, there are at least
three stations in our market that are
pawns - they have no audience, and they seemingly play
random songs. One of them (a competitor to my wife's station) played Metallica - into Led Zepplin, into some RAP song (and I mean hardcore rap). NOONE can
stand to listen to that station - if you love Metallica (white male "rocker", 18 to 35 years old), you hate Led Zepplin (white male "classic rocker", 30 to 54 years old). If you like Led Zepplin, you
hate rap (urban male, 12 to 39 years old). Those are the statistics of the population - and yet this Clear Channel station, which is being programmed by some Guru 2500 miles away, is playing
exactly songs that will
piss off the listeners of the
previous song. Care to guess why?
Yep, because that playlist is designed to solidify their Country and Top-40 stations. And it does - the ratings of this pawn station are less than .1% of the market. There are 20 stations, and Arbitron (the ratings company) rates the top 18. This pawn station
never even makes it into the top 18. What it does do, though, is
shift the probabilities in the market. With it's introduction - and noone is listening to it, we have proof - since it's introduction, both the Country and Top-40 stations have enjoyed a combined 3% increase in male listenership. The 3 "real" (non-pawn) rock stations have (combined)
lost 3%. And, I mean it happened
overnight, exactly with the introduction of this... pawn.
So, fight back - Citadel introduces a pawn of its own, a sports radio station on the AM band. This station, which has an effective signal radius of about 10 feet, is designed to screw with a Clear Channel "Talk Radio" station. And guess what - this sports pawn
beats the ClearChannel pawn, with a .4% share (compared to .1%). They could do "custom" radio broadcasts, personalized "per listener" - and in a week, they'd have done *everybody* who listens. "Today's show is for you, Bob!" And yet, the impact was huge. The talk-station took a dump of almost 4%, and the two "younger oriented" rock stations picked them up, right on cue... my wife got the 3% she wanted, and her competitor got 1%. All the result... of probabilities.
You'll never find a person who
likes the playlist of a given radio station - but from if you apply a probability model to "who will like it" (and by "who", I mean segregated right down to age, gender, and income) - you can predict
how many of each type will like it, but not
who, exactly. Sound familiar?
Your reactions to my jokes about QM "not being real" indicates that you think the listeners, above, have no cause for their behavior - as if each specific listener is truely governed by randomness, probabilities, and "waves". Right? Err... no, each person does *exactly* what they do for *very exact* reasons. You certainly do, you have legitimate cause for every action you take. There is no randomness *at all* in the *choices* they make. It might appear random to an observer because he cannot measure what those choices are based on, but
that limitation changes nothing. Their choices are exact, not random, and
are deterministic in each one of their eyes.
I mean, how often do you act in a truely random fashion? Ever? No, you have cause for each action you take - even when
trying to act random, you'll still have a trail of causality - but only you will know what it is.
So, just like my wife's radio station trying to figure out how to get the most male listeners aged 18 to 34, with a secondary target of males 25-49 - she cannot predict what *each one* will want to listen to. She can model them as a population, though, and get pretty damned close. QM is no different - and just like my wife's technique does not imply true random behavior, QM does not imply true random behavior either.
QM simply implies that we can't predict what the behavior will be in a specific case, because the level of cause is beyond our ability to measure. It does not imply that such level of cause doesn't exist - because if it does, then we as people have no cause for our behaviors, either. After all - I can model us with probabilities, right?
So, hopefully I've clarified my views on QM.