Random Encounter Generation?

[Mezuka]

Which type of dice roll do you use? Do you prefer a bell curve distribution (as in 2d6) or a linear one (1d20)?

I'm partial to d66 results because of the 6 series of 6 results (11, 12, 13, 14, 15, 16). I group similar encounters within the same decimal. The second value is the number of creatures encountered. Example: Let's say 10s are wandering goblin scouts and patrols. 11 is 1 goblin, 12 is 2 goblins, etc.

 

[Emirikol]

Bell curve with elimination of ends as they are encountered, where as the center would be possible to encounter multiple times.

 

[Yora]

I like 2d6. You get different odds for different numbers, and with only 11 entries there's a decent chance for entries to come up more than once. That contributes to creatures on the list feeling like they are part of the local wildlife population.
List longer than that don't seem necessary unless you use them universally throughout a campaign, but I think random encounters work better if they are customized to specific environments.

 

[J.Quondam]

I've typically used bell curves (usually 2d6 or 2d8) tables, wherein some entries might contain a linear subtable (usually 1d4 or 1d6).

Thanks for mentioning d66, though. That could be a neat way to add a touch of variety where needed, rather than entire subtables which can be overkill.

 

[billd91]

Depends on how many options I want and how I want the table to work. Flat/bell distribution - whether or not that matters depends on how I apportion the rolls among my options because there's little I can do with a bell distribution that I can't do with a flat one.

 

[bloodtide]

My standard is 2d6, but I also mix in 2d8 and 2d12. For some I use the d100.

 

[aramis erak]

Which type of dice roll do you use? Do you prefer a bell curve distribution (as in 2d6) or a linear one (1d20)?

I'm partial to d66 results because of the 6 series of 6 results (11, 12, 13, 14, 15, 16). I group similar encounters within the same decimal. The second value is the number of creatures encountered. Example: Let's say 10s are wandering goblin scouts and patrols. 11 is 1 goblin, 12 is 2 goblins, etc.

aside from nitpicking that 2dX is a quantized bilinear, not a quantized approximation of a bell, nor a true bell (as a bell has changes in rate of change)

My preference when polys are not systemically problematic is for a flatspotted quantized trilinear... 1dX+1dY where X≠Y
a sample follows 1d4+1d6:

Roll	Histogram	Permutations (d4+d6)
10	◼	4+6
9	◼◼	3+6, 4+5
8	◼◼◼	2+6, 3+5, 4+4
7	◼◼◼◼	1+6, 2+5, 3+4, 4+3
6	◼◼◼◼	1+5, 2+4, 3+3, 4+2
5	◼◼◼◼	1+4, 2+3, 3+2, 4+1
4	◼◼◼	1+3, 2+2, 3+1
3	◼◼	1+2, 2+1
2	◼	1+1

Of these, I like the AD&D 1d12+1d8 rather well. 19 items, 9 to 13 flat spot, linear decrease. 1d20+1d10 is just too many entries...

I'm also fond of areas marked on a grid x×y... which has the added benefit of being shiftable in 4 directions rather than 2....

So, a d4×d6 with two steps of freedom either end of each is a -1 to 6 on the d4, and -1 to 8 on the d6, allowing collapsing overlapping tables.

If single die type only, d×d grit as well...