You roll a 3...

[ValhallaGH]

If you're worried about character effectiveness, simply have all starting PCs fight a house cat; while the character is naked and unarmed. If they lose then they get to use point buy.

It worked spectacularly for two campaigns, and I only had one PC lose (something like three 8's, and nothing above 12). Several tried to lose (including casting Rage on the cat) but only one was defeated by the cat.

 

[sjmiller]

That gets me wondering about other combinations.

Well, as a professional mathematician, I wonder about the combinations you wonder about!

I was wondering how many combinations there are for rolling 3d6 six times (A lot I discovered). Then I was trying to figure out how to calculate the number of combinations for rolling 4d6, drop low, rolled six times. Still working on that one, as I do not that formula.

 

[Water Bob]

I was wondering how many combinations there are for rolling 3d6 six times (A lot I discovered). Then I was trying to figure out how to calculate the number of combinations for rolling 4d6, drop low, rolled six times. Still working on that one, as I do not that formula.

Since we're talking about Bell Curves, here, it seems that you might be more interested in standard deviation? Rolling a "3" or a "18" on 3d6 is pretty unlikely at 0.463%. Rolling a 10, though, happens about half the time (50% probability).

And, just for fun, I'm going to throw 3d6 10 times....



Interesting! The very first throw is a smack, dab, even 10, and I threw 10-11 (average 3d6 stat) five times out of ten--exactly 50%. Neat.

 

[Flatus Maximus]

Then I was trying to figure out how to calculate the number of combinations for rolling 4d6, drop low, rolled six times. Still working on that one, as I do not that formula.

The 4d6-drop-low method produces numbers in the range 3-18. So, while the possible ability score arrays are the same as with the 3d6 method, they occur with different probabilities.

 

[Dannyalcatraz]

I'm presuming it flattens the curve a mite.

 

[Flatus Maximus]

I'm presuming it flattens the curve a mite.

On the contrary, it shifts it up towards 18. The first jpeg is the probability distribution for the 3d6 method, and the second jpeg is for the 4d6-drop-lowest method. (The height of the bar is the probability of that score coming up.) It's more likely to get higher scores, which will produce "better" ability score arrays.