@Shog9 Sure.

dark, almost bitter chocolate drizzled on top

dammit. Now I'm all hungry

@hobodave Yes, I recalculated it in excel. I also have week 5.

But I don't think you can do this.

But it will definitely be not the most-far-off-week.

Oh, sorry, it isn't formatted of course

But yes, the best week is 5, with 0.733

@rumtscho what is that a chart of

@Jay The correlation of rank achieved in a certain week to overall rank (for some kind of sports competition @hobodave gave me the data for)

It means that if somebody does well in Week 5, it is probable that they will do well overall too. Week 3, not so much. (Assuming that this happens all the time in the same way, for real confirmation the data for many years should be studied, not the data for one year only)

thanks @rumtscho

@hobodave, if you need to make this more often, I can give you the r code, it is two lines

But if you don't have R installed, you will have to use R web, and you will have to have a public Internet address where the data lies (in csv)

Here is what I used, if you should need it

@rumtscho I don't use R. Your calculations were identical to Excel's CORREL btw

@hobodave Including the numbers?

Let me see if I haven't forgotten to change the method then

@rumtscho identical

Yes, I have

It takes pearson by default

interesting

@hobodave that's with a method for ranked data

Using the wrong method for correlation in statistics is a bit like using the correct formula but with wrong units with physics

The direction will probably be similar, but if you calculate in calories instead of joules, the numbers will be off

so this is the formula i want?

cor.results <- sort(round(cor(womens.nc[,3:ncol(womens.nc)], method="kendall")[,1], digits=3))

This is the formula you want, yes

It is because the ranks have no meaning

hmm don't the ranks have to be normalized to 1-60 though?

Using the default method (the one Excel uses) assumes that the ranks have meaning like integers - so that the person at rank 4 is exactly twice as bad as the person at rank 2, and the person at 40 is exactly 10 times as bad as the one at 4

according to this they do: statistics.laerd.com/statistical-guides/…

some of the ranks I gave you have numbers much greater than 60

Aren't they? I didn't pay attention

The list isn't complete.

There were more than 60 women.

yea it's a partial list

OK, let me see how to get that working

I only gave the top 60 overall, but some of them may have been as much as rank 200 in a given week

I wonder if I can get the full dataset

my R knowledge is somewhat rusty, haven't used it since November, and back then I learned it for the first time

@hobodave I think it went on on pages 2 etc ;)

@Mien yea but I mean an automatic way

@hobodave Being lazy again? ;)

What does the html code say?

@hobodave what do you want your final statement to be?

"The higher your place in week 5 when compared to everyone, the better your chance you will place high overall"

@rumtscho what do you mean?

or

"The higher your place in week 5 when compared to the ones who are going to win overall, the better chance you will place high overall"

The calculation we made now is good for the first statement

meh I dunno, I don't think it matters that much anymore :)

the gist of it is 5, 6, and 1 are good matches and 2,3,4 are not

If we normalize the partial list, we will get a cor. coefficient which lets you make the second statement

Depends on your definition of "good", but yes, they are probably the better choice

@rumtscho It's about someone who will have have to choose 1 program out of those six.

I'm guessing she wants to compare a bit overall.

So the first statement is the most important I guess.

And I even think it's not that important anyway.

yea

You're overthinking things again.

I'm just being a nerd because she asked me which she should do

a hand-wavy approx. is sufficient

I also gave her the 70/80/90th percentile scores for each week

Is she interested in results or fun?

results

Do you have a reason to think the results will be the same this year?

Are the programs different in the different weeks? Or the trainers?

I know for a fact that the results this year will have absolutely no bearing on last year

@mfg - I completely don't see how robert's answer there applies to your Q

@rumtscho et al, in re; cooking.stackexchange.com/q/21419/15

cool, i didn't want to go through the comments :)

the question applies specifically to foods that are massively consumed by veg*ns

so is rice, but we don't tag rice as veg*n

they have, on average, more experience eating and cooking spongiform foods

that is not the purpose of tagging @mfg

i was on that meta discussion, and i made that point on my own

tags are not meant to be used as an ad-hoc notification service to target users

it is meant to target a demo? right?

is that not what robert said?

That is not what he said, and I am inclined to agree with @rfusca here

Your question is not about being vegetarian, vegan or modifying a dish to be such. It's simply about an ingredient commonly used by those groups.

I would suggest instead, tagging questions that are specifically targeted at a group with an appropriate tag. "Is it possible to make vegan omlets?" would certainly support a vegan tag

specifically targeted at a group with an appropriate tag

thats verbatim from meta.cooking.stackexchange.com/questions/21/…

so how is that not what Robert said?

your interpretation of the verb targeted is rather loose

I would support a vegan tag for the "is is possible to make vegan omlets" example

His example provides the clear intent

whether you define "targeted" as @ SEO traffic or @ users who can answer will have interest, it still works

Because in this case, being a vegan dish is important part of the question.

his example would get closed as a recipe request

i don't think it would..

Depends largely on the wording of the body.

@rfusca agreed

but thats not really pertinent to this conversation

@rfusca true

There's a difference between being coincidentally "targeted" at a group and specifically tagging something to "target" a group. You're doing the latter @mfg.

@hobodave please elaborate

Veganism has a corollary Categorical problem in defining its tags as Atheism does

I would say that a tag should describe what a question is about, not who is it for

You're trying to use the tag as a notification of sorts. Maybe someone can come up with a better counterpoint, but let me try: I know molecular gastronomists probably have a high level of knowledge regarding the Maillard reaction, and I have a question about searing a steak. This would never get tagged molecular-gastronomy just because I want them to see it.

@rumtscho exactly

The vegan omelet is about veganism.

Tofu isn't by definition about veg*n.

each has a corps of interests and a sub-domain of topics pertinent to its discussion, yet remain neither a religion nor a diet

the question is about the dynamics of cooking in one of those sub-domains, tofu

tofu isn't a sub domain of veg*n

and tags aren't meant to be used as a hierarchical classification system

It is even slightly discriminating to tag questions in a way which connects a subject matter with a defined group. I can imagine what would happen if I tagged a question about picking strawberries with Mexican, on the assumption that Mexicans know most about it.

you don't tag a question about rice asian-cuisine, unless it specifically refers to some aspect of asian cuisine

what i'm getting at is that tofu is descriptive of veg*n diets, correct?

no

<---- non veg*n, eats tofu

if someone says 'I eat tofu regularly' you would infer what about their diet from his/her statement?

that they like tofu

I appreciate your obstinance @rfu

lol

I accept that the rest of your differ

I don't think he's being obstinate. I eat tofu very regularly, as well as steak, and chicken.

and respect all the reasonings

regular meaning, bi-weekly

right, as opposed to vegans who are warned to limit their intake of tofu and other soy products

limit it to no more than once per day

I think we're headed off on a tangent here

the doctor told me to limit my fat...should we tag all fat related questions with 'rfusca' ?

I conceded my point when I said I accept and respect the time to go over this

thanks for talking about it @mfg

Argh, just when I wanted to hop in on the discussion :(

@rfusca I suppose if I wanted to create a synonym for mfg=bbq, then yes

I think the smarter thing to do here is include tofu in the contest

@hobodave no prob, much better here than belaboring in the comments

@hobodave But then I would leave out 'vegetables', and also include quorn and the likes.

@hobodave sounds like too meta a solution

no it wouldn't

Although tofu is in fact a form of vegetables.

inclusion does not imply exclusion

@Mien quorn would be included as vegetarian

@mfg According to your definition ;)