Levi & Joe & Thor & Company

originally posted by Rahsaan:
originally posted by Florida Jim: So when I say that what I got from that piece is that real, natural and industrial tend to overlap, I realize that I may have missed the point entirely.
Click to expand...

That seems to be the main value of the chart from my perspective as well.

But I'm not sure what the probability discussion adds.

When Nathan says that wines have mixed probabilities of ending up in various categories, does that depend on vintage or the palate of the person making the judgment?

Either way, the point about overlap is good.
Click to expand...

That means that there is a non-zero probability of being both a real wine and a natural wine, for example. Sometimes these probabilities are more even than others. Those are the places where it is hard to make distinctions.
 
originally posted by SFJoe:
Although no one calls him out on sliding everything onto a univariate parameter.

I'll just let it slide.
Click to expand...

The first principal component (or factor score) can be used to represent several variables as one variable. I'm imagining that the component I am using explains a lot of the variance. It's possible that it doesn't break out this way. I tried to be clear with my assumptions.

We essentially agree on this, I projected to one dimension for the sake of simplicity. Really figuring this out would require something more complex than I have the time or maybe even the brainpower for.
 
originally posted by VLM:

That means that there is a non-zero probability of being both a real wine and a natural wine, for example. Sometimes these probabilities are more even than others. Those are the places where it is hard to make distinctions.
Click to expand...

Hey Rahsaan,
I interpret these hieroglyphics to mean there are probably zero examples of naturally occurring distinctions in any given wine; really.
What do you think?
Best, Jim
 
originally posted by Florida Jim:
originally posted by VLM:

That means that there is a non-zero probability of being both a real wine and a natural wine, for example. Sometimes these probabilities are more even than others. Those are the places where it is hard to make distinctions.
Click to expand...

Hey Rahsaan,
I interpret these hieroglyphics to mean there are probably zero examples of naturally occurring distinctions in any given wine; really.
What do you think?
Best, Jim
Click to expand...

It was supposed to show that Angeli, Barthod and Lascombes are easy to categorize as "Natural", "Real", and "Industrial". Baudry is a more difficult case.

This is meant to be schematic, not representational of any specific wine or producer.
 
Why didn't you use a Venn diagram or some kind of star chart? (Creating and numbering a single X-axis is a technique associated with the Gouty Guy, as if you don't know.)
 
originally posted by Jeff Grossman:
Why didn't you use a Venn diagram or some kind of star chart? (Creating and numbering a single X-axis is a technique associated with the Gouty Guy, as if you don't know.)
Click to expand...

I thought about a Venn diagram and discussed it with Blackwood and Lipton over the weekend. In the end, I'm comfortable with this representation because ti more closely mirrors the way I think.

It is a fair point.

I've never had a real problem with scoring wines, per se. I don't personally do it, but I think it could be done.

Dimensionality reduction is a common and (mostly) good thing in complex models.
 
originally posted by VLM:
originally posted by Florida Jim:
originally posted by VLM:

That means that there is a non-zero probability of being both a real wine and a natural wine, for example. Sometimes these probabilities are more even than others. Those are the places where it is hard to make distinctions.
Click to expand...

Hey Rahsaan,
I interpret these hieroglyphics to mean there are probably zero examples of naturally occurring distinctions in any given wine; really.
What do you think?
Best, Jim
Click to expand...

It was supposed to show that Angeli, Barthod and Lascombes are easy to categorize as "Natural", "Real", and "Industrial". Baudry is a more difficult case.

This is meant to be schematic, not representational of any specific wine or producer.
Click to expand...

Monkey boy, I love you dearly, but this is a bit too esoteric for somebody who has been away from school as long as I have. To be honest, I think the distinctions drawn by those three words have little practical application.
And virtually no value outside of wine geek circles.
Best, Jim
 
originally posted by Florida Jim:
originally posted by VLM:
originally posted by Florida Jim:
originally posted by VLM:

That means that there is a non-zero probability of being both a real wine and a natural wine, for example. Sometimes these probabilities are more even than others. Those are the places where it is hard to make distinctions.
Click to expand...

Hey Rahsaan,
I interpret these hieroglyphics to mean there are probably zero examples of naturally occurring distinctions in any given wine; really.
What do you think?
Best, Jim
Click to expand...

It was supposed to show that Angeli, Barthod and Lascombes are easy to categorize as "Natural", "Real", and "Industrial". Baudry is a more difficult case.

This is meant to be schematic, not representational of any specific wine or producer.
Click to expand...

Monkey boy, I love you dearly, but this is a bit too esoteric for somebody who has been away from school as long as I have. To be honest, I think the distinctions drawn by those three words have little practical application.
And virtually no value outside of wine geek circles.
Best, Jim
Click to expand...

But it could draw still some more outrage on YAOWB with some predictably Johnsonian refutations :).
 
originally posted by VLM:
originally posted by SFJoe:
Although no one calls him out on sliding everything onto a univariate parameter.

I'll just let it slide.
Click to expand...

The first principal component (or factor score) can be used to represent several variables as one variable. I'm imagining that the component I am using explains a lot of the variance. It's possible that it doesn't break out this way. I tried to be clear with my assumptions.

We essentially agree on this, I projected to one dimension for the sake of simplicity. Really figuring this out would require something more complex than I have the time or maybe even the brainpower for.
Click to expand...
Sure. We don't know what most of the axes are. I just have my doubts that the fpc captures most of the variation because of my assumption that lots of the variation is orthogonal. But what am I, a statistician?
 
Yet Another Other Wine Board.

The meaning (though I'm sure there is no real need to explain it) was that if one refers to these wines as real wine then it is certain to provoke the heated rejoinder from certain producers that their wine is also real. And they may go so far as to kick the bottle to prove it.
 
You must log in or register to reply here.
Back
Top