Graph changes when freezing or saving

[joost]

I seem to have problems freezing and/or saving empirical CDF graphs. The CDF graphs change when I freeze or save them. Actually, when I take the frozen graph (which differs from the original) and freeze it a second time, the graph changes again. Each time Eviews plots a different function. The x- and y-axes remain the same, however, and the underlying data from which the function is derived cannot have changed either, I think.

Sometimes the difference between the original graph and the frozen/saved graph is small, sometimes large. The same happens with several other graph members of the distribution family, i.e. empirical survivor graphs and empirical log survivor graphs. I've tried both manually saving/freezing the graphs and from within a program. Neither works. I've tried saving the graphs in different units of measurement (e.g. inches, points, pixels, etc.), but that doesn't help either.

I use Eviews 8.

Is this a known problem? I can't find any reference to this issue on the forum. Or am I overlooking something obvious? Hoping for the last...

[EViews Gareth]

I haven't seen this before, and can't replicate it.

Is your copy of EViews up to date? (Help->About EViews)

[joost]

I'm running Eviews 8.1, Standard Edition - Jan 12 2015 build. Not sure what other info I can give you that may help.

[joost]

Perhaps it helps if I give an example. The attached picture shows (reading from left to right) the empirical CDF of matrix Z_CONNECTICUT, the freeze thereof, and the freeze of the freeze. As you can see, each time the distribution function differs. Might it have to do with my graphics defaults? Or other settings that I have overlooked?

[EViews Gareth]

Could you provide the workfile used to generate those graphs?

[joost]

Of course, see attached. I have stripped down the workfile to only contain the important objects. So it contains the matrix Z_CONNECTICUT and I added three graphs (freeze1, freeze2, freeze3) which show you what happened when I copied the CDF graph of Z_CONNECTICUT. Each of the three times that I froze the CDF graph, it led to a differently shaped function. Does something similar happen when you try this?

[EViews Gareth]

Perfect. I can confirm I see the same thing happening with your workfile. We'll fix it.

[joost]

Thank you, looking forward to the solution!

[joost]

So not an easy fix, I gather?

[EViews Gareth]

We have a fix. It will be in the next patch.

[joost]

Bril! When do you expect the next patch? I'm asking because I currently rely on screen prints to 'export' the graphs from Eviews to other software, which is kind of tedious...

[joost]

Thank you Gareth, works like a charm now!