Just for information: last September the number of posts per day went up by 25%. Actually, the third week of September saw an increase of 50% compared to previous months.

And that's a factual increase in number of posts, I leave it to your imagination what that means in number of visitors and pageviews.

A road which suddenly gets 50% more traffic gets jammed, and unfortunately (and frustratingly) that's what happens with fuckfrance too, temporarily.

Just hoping the suicide rate stays stable globally.

It's Tozer's fault.

Tozer's "Saturday Night" thread: That thing has over 109,000 views!

It's Tozer's fault.

I meticulously tried to avoid that.

I meticulously tried to avoid that.

Saturday night fever is too addictive...
Try harder next time

Are there any solutions or is this just something we should get used to.

I mean software wise, not about buying a new server.

your "magnificient" thread of blown-up musslims is responsible too..
BTW it makes you the equal of those shit site showing dead soldiers or civilians, only because they dont share our values... congrat .. showing your "intelligence " as usual..losers

Dumbdefrance,

When I wrote that I wished for the global suicide rate to stay the same, I wasn't thinking of you.

What a dishonest way to present figures, why I'm not surprised.
Your months blocs are not proportional.

Should you make a graph with the number-of-posts axis starting at 0 (and not 1,500) the September peak would not look like if it was twice as large as June, when June represents in fact 80% of September.

I leave it to your imagination what that means in number of visitors and pageviews.

Well yeah, because it says absolutely nothing about it.

And what about previous years?

Do we also have to leave to our imagination if there were not similar season peak in September 2003 or 2004 with no server slowdowns?

Who do you think we are to believe such bullshit?
Republican voters?

Should you make a graph with the number-of-posts axis starting at 0 (and not 1,500) the September peak would not look like if it was twice as large as June, when June represents in fact 80% of September.

Which makes September 125% of June as I wrote, you number-illiterate. And because I plotted the Y axis starting at 1500, I inserted a horizontal line mentioning that.

I'm not forging statistics like you, to support I point I was convinced of in the first place. Instead I post data and then a brief analysis based on those. The analysis may be correct or not, but it is honest, straightforward and derived from correctly presented data that can be easily verified.

I am out K for today Atlantic, but i agree with this:

Should you make a graph with the number-of-posts axis starting at 0 (and not 1,500) the September peak would not look like if it was twice as large as June, when June represents in fact 80% of September.

Its not a fair presentation of numbers by Prescott, only rookies post numbers like that.....starting at 1500 when he is trying to compare the numbers, LOL, a rookie mistake by Prescott

starting at 1500 when he is trying to compare the numbers, LOL

Newbies are so much fun. Atlantic found himself a girlfriend. Oh, here's some Danish stats for you, Ms Roswell:



OMG, it doesn't start at zero! Oh zees is zee first time in zee wee wee life a stat has a bottom line I do not le agree wizz! Oh le not good!

ROFL, you are so much fun guys.

Hey again prescott, sorry i didnt answer back yesterday, but for some reason i got banned shortly after i posted this response to you.

I am out K for today Atlantic, but i agree with this:


"Should you make a graph with the number-of-posts axis starting at 0 (and not 1,500) the September peak would not look like if it was twice as large as June, when June represents in fact 80% of September."

Its not a fair presentation of numbers by Prescott, only rookies post numbers like that.....starting at 1500 when he is trying to compare the numbers, LOL, a rookie mistake by Prescott

.....

Your answer here, with a figure that too is cut on the y-axis dosent prove shit, that only have one variable, its not suppose to compare alot of numbers with each other, its just presenting the development in one variable, for which reason there is no need to have the y-axis starting at zero.

I still claim you are a rookie when it comes to presenting numbers Prescott, that last answer dosent convince me otherwise, on the contrary.....!

am out K for today Atlantic, but i agree with this:


"Should you make a graph with the number-of-posts axis starting at 0 (and not 1,500) the September peak would not look like if it was twice as large as June, when June represents in fact 80% of September."

Its not a fair presentation of numbers by Prescott, only rookies post numbers like that..... starting at 1500 when he is trying to compare the numbers , LOL, a rookie mistake by Prescott

Compare was the key word for you yesterday Prescott, is this figure comparaing alot of numbers?



- Just wanted to highlight to important word before you ban me again for correcting you

Atlantic and EU_Citizen are perfectly right, Presscott is showing again his usual cheap batave marchand de tapis crook mindset, Ks to you

OMG, it doesn't start at zero!

It's ok not to start the Y axis at zero with a curves chart, it doesn't mis-represent the datas, it just ommits an area empty of datas.

But with eras or volumes graphs, it's a different story.
The volume of the boxes are supposed to be proportional to the values. The zone you have ommitted would not be empty but show MOST of the datas.

I remember when I first posted the graph below, you bashed me because months with no datas were not showing.
I then corrected the graph, which made even more obvious the point I was trying to make. (The March 2003 peak)



Why don't you correct your own graph and show us what it would like if boxes were proportional to the datas like it should be.

If you don't, I will.

Yawn.

Your examples are totally unvalid Pressie, the reference figures for taux de chomage are perfectly known by everybody, hence the starting y axis figure, not the case at all for number of posts on FF, where any variation could be made to look the way you did it. End of story

If you cant see this is not an optimal way to present numbers, when you try to compare how large the "september contributions" are relative to the rest of the months, you have a problem Prescott, its textbook example......on how not to do it!

Hey, BNP is doing pretty well...

Quibble, quibble, quibble.

There -- that's what statistics are good for.

I'm not forging statistics like you, to support I point I was convinced of in the first place.

You did worse!

You misrepresented datas to dishonestly support a point that you know to be false, in a very Neo-Con kind of way.

We all know that the recent server's slowdown are not related to a supposed traffic increase but to a recent change of configuration (hard and soft), probably due to financial difficulties.

The system looks so crippled that we could even be entitled to challenge your datas themselves.
In the current situation, how can we be sure that the data-base itself is not corrupted?

Anyway, let me sho you something:

Your datas (correct or not) presented as curves.
As you can see, the Y axis can start as 0 or 1500 it doesn't affect the form of the curve and proportions are respected:



Now let's use bars.
As you can see, only an Y axis starting at 0 allows the bars to be proportional to the values.

Starting the axis at 1500, omitting the graduations (I did it too in my examples, but just to hide that I don't know the exact figures that you have carefully hiden) was a deliberate attempt to misrepresent reality:


You even used a graph with a larger 3D effect than my bars example to increase the misrepresentation to the maximum.

Now, come on, post the 2004 figures, I'm very curious to see your interpretation of season variations for that year, and the effect on the quality of service!

He's right.

Anyway, beautiful day on Paris, isn't it?

Ouai vraiment cool, Hé Ccold, cesse de faire la chipie svp, c'est qui qui a raison donc ?

Lies, Damn Lies, and Statistics. ...

did any of you guys work for the state of florida board of elections in 2002?