# Let's make a bell curve. **URL:** \ **Category:** The Game Room\ **Created:** [March 11, 2010, 6:44pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137 "2010-03-11T18:44:04Z")\ **Posts on this page:** 14\ **Page:** 3
**Author:** ![garygnu](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/garygnu/32/2924_2.png) [@garygnu](https://boards.straightdope.com/u/garygnu)\ **Post date:** [March 16, 2010, 3:10pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/41 "2010-03-16T15:10:20Z")
> [@MeanOldLady](#): > > Why so much love for 9 as opposed to 1? Many people said they had chosen 9 specifically to dick with the results. I guess people figured 1 would be over-represented on its own. ---
**Author:** ![MeanOldLady](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/meanoldlady/32/10737_2.png) [@MeanOldLady](https://boards.straightdope.com/u/MeanOldLady)\ **Post date:** [March 16, 2010, 3:13pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/42 "2010-03-16T15:13:12Z")
I would have reckoned 1 would have dicked with the bell as much as 9. Pretty much, if we switch the numbers around in your post, it’d seem equally likely and plausible to me. But apparently not… ---
**Author:** ![Jackknifed\_Juggernaut](https://avatars.discourse-cdn.com/v4/letter/j/9de053/32.png) [@Jackknifed\_Juggernaut](https://boards.straightdope.com/u/Jackknifed_Juggernaut)\ **Post date:** [March 16, 2010, 3:18pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/43 "2010-03-16T15:18:09Z")
Unless I’m missing the point of the exercise, wouldn’t simply asking everyone “pick a number” be a better way to yield a bell curve. At the very least, you might figure out a general behavior bias. ---
**Author:** ![MeanOldLady](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/meanoldlady/32/10737_2.png) [@MeanOldLady](https://boards.straightdope.com/u/MeanOldLady)\ **Post date:** [March 16, 2010, 3:21pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/44 "2010-03-16T15:21:13Z")
> [@Jackknifed\_Juggernaut](#): > > Unless I’m missing the point of the exercise, wouldn’t simply asking everyone “pick a number” be a better way to yield a bell curve. At the very least, you might figure out a general behavior bias. Perhaps, but it is more interesting to see what people pick when they know the goal is to make a bell curve. To me, anyway. ---
**Author:** ![Zeldar](https://avatars.discourse-cdn.com/v4/letter/z/b77776/32.png) [@Zeldar](https://boards.straightdope.com/u/Zeldar)\ **Post date:** [March 16, 2010, 3:31pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/45 "2010-03-16T15:31:34Z")
Maybe everybody who’s participated so far was thinking, “Oh, wow. There will probably be at least 1000 other votes. I’d better make a case for these outlier numbers that nobody will mention. Surely everybody will be picking 4, 5, and 6.” How close will this ever get to 1000 players, you suppose? ---
**Author:** ![Jackknifed\_Juggernaut](https://avatars.discourse-cdn.com/v4/letter/j/9de053/32.png) [@Jackknifed\_Juggernaut](https://boards.straightdope.com/u/Jackknifed_Juggernaut)\ **Post date:** [March 16, 2010, 3:52pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/46 "2010-03-16T15:52:16Z")
> [@MeanOldLady](#): > > Perhaps, but it is more interesting to see what people pick when they know the goal is to make a bell curve. To me, anyway. I agree that it is interesting. But as we’ve seen, you get individual biases that get into the results such as trying to be different, not knowing what a bell curve is or what it represents, etc. ---
**Author:** ![garygnu](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/garygnu/32/2924_2.png) [@garygnu](https://boards.straightdope.com/u/garygnu)\ **Post date:** [March 18, 2010, 11:52pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/47 "2010-03-18T23:52:18Z")
> [@Jackknifed\_Juggernaut](#): > > …you get individual biases that get into the results such as trying to be different… That’s the point. ---
**Author:** ![Drain\_Bead](https://avatars.discourse-cdn.com/v4/letter/d/e47774/32.png) [@Drain\_Bead](https://boards.straightdope.com/u/Drain_Bead)\ **Post date:** [March 18, 2010, 11:57pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/48 "2010-03-18T23:57:39Z")
I picked 3 because I figured people would be picking the middle and ends a lot. I probably should have picked 4 or 6, because it seemed like a lot of people had the same thought as me. ---
**Author:** ![Lumpy](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/lumpy/32/446_2.png) [@Lumpy](https://boards.straightdope.com/u/Lumpy)\ **Post date:** [March 19, 2010, 2:40am UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/49 "2010-03-19T02:40:42Z")
> [@Chronos](#): > > Rather than trying to guess what everyone else would do, I decided to take an approach that, if everyone followed it, would naturally produce a bell curve: I rolled dice. Specifically, I did (3d6)/2, rounded down, which got me a 4. Why would that produce a bell curve? If everyone had used a truly random method such as dice to pick a number, that would have produced a flat line (all possibilities about equally likely to be chosen). Since the OP specifically asked for choices to yield a bell curve, the best number to choose is 5. ---
**Author:** ![Shark](https://avatars.discourse-cdn.com/v4/letter/s/73ab20/32.png) [@Shark](https://boards.straightdope.com/u/Shark)\ **Post date:** [March 19, 2010, 12:36pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/50 "2010-03-19T12:36:29Z")
I picked 3 because I figured it should be within 2 of the middle, and I thought that more people would pick to be above average than below and I wanted to counteract that assumed bias. ---
**Author:** ![Hoopy\_Frood](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/hoopy_frood/32/8452_2.png) [@Hoopy\_Frood](https://boards.straightdope.com/u/Hoopy_Frood)\ **Post date:** [March 19, 2010, 3:32pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/51 "2010-03-19T15:32:26Z")
> [@Lumpy](#): > > Why would that produce a bell curve? If everyone had used a truly random method such as dice to pick a number, that would have produced a flat line (all possibilities about equally likely to be chosen). Since the OP specifically asked for choices to yield a bell curve, the best number to choose is 5. Not true. 3d6 produces a bell curve in the long term. Always. Unless the dice are loaded. You have less than a 0.5% chance of rolling a 3 or an 18. You have around a 12.5% chance of rolling a 10 or 11. Only if you roll one die you have a flat line. ---
**Author:** ![Lumpy](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/lumpy/32/446_2.png) [@Lumpy](https://boards.straightdope.com/u/Lumpy)\ **Post date:** [March 19, 2010, 8:43pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/52 "2010-03-19T20:43:54Z")
> [@Hoopy\_Frood](#): > > Not true. > > 3d6 produces a bell curve in the long term. Always. Unless the dice are loaded. You have less than a 0.5% chance of rolling a 3 or an 18. > > You have around a 12.5% chance of rolling a 10 or 11. > > Only if you roll one die you have a flat line. Oh, I getcha. I’d thought you meant something like using d6 in a way that simulates a single d9. Yes, if you’re rolling any dice multiple times it tends to a bell curve; we had a thread about that not that long ago. ---
**Author:** ![Ponch8](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/ponch8/32/9883_2.png) [@Ponch8](https://boards.straightdope.com/u/Ponch8)\ **Post date:** [March 19, 2010, 8:50pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/53 "2010-03-19T20:50:24Z")
It looks a lot like a Chanukkah menorah turned on its side. ---
**Author:** ![Zeldar](https://avatars.discourse-cdn.com/v4/letter/z/b77776/32.png) [@Zeldar](https://boards.straightdope.com/u/Zeldar)\ **Post date:** [March 19, 2010, 8:54pm UTC](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137/54 "2010-03-19T20:54:53Z")
To paraphrase a popular expression: “That bell won’t ring.” [Previous page](https://boards.straightdope.com/t/lets-make-a-bell-curve/532137.md?page=2)