Town Size, map generation changed?

[Jando]

I'm getting odd results when making new maps, size of all towns is roughly the same and lower than the median city size.

With these settings (Screenshot: http://i.imgur.com/BHN9t2P.jpg) I'm getting this city list (Screenshot: http://i.imgur.com/7THwwZ2.jpg). Multiple maps generated with these settings all show the same effect. 

[MCollett]

size of all towns is roughly the same and lower than the median city size.

That's almost certainly because you are asking for a median size larger than the maximum size, as set in the experimental Settings panel.  That maximum defaults to 3000: if you want 30 towns with a median of 5000, you should change it to at least 15*5000=75000.

Best wishes,
Matthew

[Jando]

Thanks for the answer, I didn't even know that there is another setting in the Experimental panel.

I'm trying to get a map with a mixture of differently sized towns, i.e. some towns around 5000 inhabitants, a few larger ones, but also some small villages with a few hundred inhabitants.

[MCollett]

I'm trying to get a map with a mixture of differently sized towns, i.e. some towns around 5000 inhabitants, a few larger ones, but also some small villages with a few hundred inhabitants.

To get your villages you need a median size quite a bit smaller than 5000, since the smallest town that can be produced is half the median size. 

Best wishes,
Matthew

[jamespetts]

To get your villages you need a median size quite a bit smaller than 5000, since the smallest town that can be produced is half the median size. 

Best wishes,
Matthew

I don't think that that's so in Experimental any more - Richard Smith replaced the system for distributing town sizes with something following Zipf's law a year or two ago, which works if "big cities" is set to zero.

[MCollett]

I don't think that that's so in Experimental any more - Richard Smith replaced the system for distributing town sizes with something following Zipf's law a year or two ago, which works if "big cities" is set to zero.

The minimum I quoted is the Zipf's law one.  If the nth largest town has typically or on average a population 1/n times that of the largest town, then the minimum size is half the median. 

Best wishes,
Matthew

[jamespetts]

Ahh, I see - thank you for the clarification!