A Topological Survey of the Lands Surrounding and Including the Last Outpost

By Jeff Jahr

It has always been my conjecture that a large part of what makes a zone astheticly pleasing is the zone's topology, or the 'lay of the land.' Certainly, a zone consisting of one long hallway-- no matter how interesting the text-- will be a less than exciting zone. The same holds true of zones that are nothing but wide open space. These types of zones do serve a purpose, but they rarely fall into that category of 'favorite zone'.

Since the all-hallway and all-space zones (two extremes) always miss the mark of 'favorite zone', I wondered just what sort of topology my favorite zones had. More to the point, I wondered if my two or three favorite zones had similar topology. This question prompted the following survey.

The survey encompassed the entire world as of August 6 1995, The 15th Day of the Month of the Dark Shades, Year 93. Rooms were tallied by zone number using the following categorization: (Note: Up and Down links were ignored in this survey.)

For each zone, the tallied data provided the amount to which each room type contributed to the whole zone. The percentages for each zone were plotted into a (1.0 x 1.0 x 1.0 x 1.0 x 1.0 x 1.0) 6-space, and a Cartesian distance algorithm was used to find the closest neighbor to each point. The correlation factor by which the final data was sorted is one minus the ratio of (the distance to the nearest neighbor / the maximum possible distance in 6-space). 0.0 denotes two zones that could not be less alike, while 1.0 denotes topologically equivalent areas. Note that when an zone is topologically equivalent to several other zones, its closes neighbor is given as the zone with the lowest zone number. Equivalence classes were not formed for this survey.

Note: Do to an error on the part of the Gods of the Last Outpost and the original creators of the areas involved, zone 0 (LIMBO) is actually a composite of three zones, Limbo, The hero clan-house, and the fofi clan-house. Zone 44 is a one-room plane area. 'Test' is a part of limbo, but is flagged as its own zone.

Although I am unwilling to proclaim one area 'more playable' than another on the basis of statistical evidence, I do believe that the data shows a striking correlation between areas of similar 'feel'. For example, notice that different zones made by the same author fall close to each other in the 6space. The same is true of turf areas like the swamp and the grasslands. My two favorite areas (one of which was written by me) also had a very strong correlation.

It is obvious to even the newbie that the topology of a zone can affect its playability. After seeing my favorite zones closely coupled in this survey, it is my conclusion that topology plays a much more subtle and important role in the playability of a zone than was previously realized. I encourage the readers to view the data and decide for themselves if my conclusion is valid.

View the correlation data.

View the raw door data.