Please check the y axis on my standard rating graph.

[mick46]

My 1442 is below 1435 and my high of 1524 is below the 1524 oon the y axis.

[richard]

Hi Mick,

The graph is currently sampled from your attempts.  So for example if you've done 1000 problems and the graph has 200 pixels across then only every 5th problem in your problem set will be represented on the graph. So if your high point is not one of the sampled points on the graph then it will not show up. The alternative to sampling is to interpolate the points, but then you end up having points on the graph that NEVER existed which can be just as confusing. Ideally there should be a top level overview that uses sampling or interpolation that then allows you to zoom in for a one-to-one mapping for a restricted time range.  It will be a while before I get to something like that however.

Regards,
Richard.

[drahacikfm]

Richard:  Seems you could make the rating graphs twice as wide as they are now.  The tables for tactical motif performance and unrated problem set performance use more than twice as much width as the graphs.  Then you could sample every 2 or 3 problems, instead of every 5.

And maybe make the y-axis label round numbers instead of the weird numbers they have now

[richard]

Drahacik: I've just changed this so that graphs are now using the available space a bit more efficiently.  I've also adjusted the sampling rate accordingly. I also noticed when making the change that the graphing library I'm using was also interpolating under some circumstances which explains a bug stormcrow had seen in his graph a few weeks ago. This change will appear after the next upload (hopefully sometime this weekend, if not earlier).

re: the labels on y-axis, yes they are a little strange but doing it that way does have the advantage that you can tell at a glance what your min and max ratings were (with the labels spread out evenly between).  I'm looking around for a more flexible graphing library so I'll have a look at this again when I've found what I'm looking for (and have time to switch to it).

[mick46]

Thanks!