The Mana World Forums

Freeyorp101 wrote:It's a trellis chart, it consists of 36 subcharts, each describing the relationship between its associated two attributes.

Yes

a correlation matrix;

No

In the map plot what are the abscissas?

I stand corrected, it seems.

For the Map chart, X describes base experience sum, Y describes job experience sum, radius describes instance count.
012-4 has the most instances despite not being as popular for experience as 027-1.

There's a description of all of this in the [?] by the title. I'll get around to adding axis labels eventually.


---Freeyorp

To be exact the status diagram is a kind of multiple 3D bar chart or better mutiple level curves chart.

Freeyorp101 wrote:[...]radius describes instance count.
012-4 has the most instances despite not being as popular for experience as 027-1.

The area of the bubbles should be proportional to instance count or it gives a wrong impression because differences are exaggerated by the square power.
See for example: http://flowingdata.com/2010/11/23/how-t ... le-charts/, http://en.wikipedia.org/wiki/Bubble_cha ... _correctly

radiant wrote:Not that I mind, but outliers really do stand out on the chart; from the visible data it appears that I'm single-handedly responsible for every "instance" of an event in those logs where luck is greater than 90. Luck in the 80s also appears to have a unique contributor, though one who's somewhat more active than I am.

You cannot say that or even give a reliable estimate of the probability your hypothesis is true because you don't know the size of the player sample. Also you would require that the class "luck≥90" has also a sufficient size compared to the player sample size which will never occur if you are alone in this class.

Freeyorp101 wrote:For the Map chart, X describes base experience sum, Y describes job experience sum,

It is likely that job experience grows in a somewhat linear way with overall experience. It explains why bubbles are roughly aligned.
The plot needs to be better readable, to have a better image, In order to acheieve that goal we must find variables and scale which separate the bubbles.
Plotting against a sum should be good in that purpose, but makes the chart heavily dependent on the number of concerned players. It increases the the scatter so much that you are obliged to use a log scale which make the chart more compact and deletes the advantage you could get.
I don't know what you are seeking with that chart so it will be difficult to suggest something that makes sense. Among the variables that (in my opinion) can have some interest are: relative to level experience gain (level grows logarithmically with exp), various averages: the preceeding variables' one, av. Player level, weighted av. mob level on map, av. Vitality (after the cross status grid it seems to be well distributed) ...
Edit1:
Also, plotting these data on TMworld map will help developers to know where to focus development effort.
Edit2:
Something weird: The bubbles along the X axis. Characters who didn't get the focus skill probably. But why a cumulated exp so high?

Excellent point, Nard! Bubble area is now proportional based on sqrt, rather than log (as with the other axes) [commit].

The difference is remarkable! Have another screenshot while the scrubber is being worked on:
---Freeyorp