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!