There has been much said and written about good and bad ideas in the startup world. Most notable and influential to me being Paul Graham’s essay around its counter intuitiveness  and Peter Thiel’s Venn diagram of ‘looks like a bad idea’ and ‘is a good idea'.
With a general consensus that coming up with good ideas is hard, how would one go about generating them?
It occurred to me that there were several ways. More specifically there are at least 3 ways to increase the number of good ideas one (or a team) generates. (I’m sure its synthesis of different ideas I have read and heard, but I cannot recall a single point of origin. If you know of any pre-existing work that my idea may have come from, please send me a note so I may correctly attribute it)
To more easily visualize, let’s say you or your team generates 100 ideas, and each of them are placed on a ‘quality’ axis. Your plot of ideas would end up looking like a bell curve. Few really horrible and great ideas in the far edges (left and right respectively) and most of the ideas around the center- thus a normal distribution.
If you draw a line vertical line at point x, the area right of the line and below the curve would be the total number of ideas above quality x. So how do you increase that area? Statistically speaking there are 3 ways to do so. 2 a bit obvious but the third not.
1.Increase the sample size : total number of ideas
Say that efforts are put in and now the idea count increased by 100 and the total is now 200. The curve would have moved up vertically resulting in an increase in ideas above quality x.
2.Shift the mean to the right : increase overall quality of ideas
Say more efforts were put in, say training, that allows for ideas that are generated to be of better quality. So 100 ideas generated after these efforts would end up plotting a curve that is shifted to the right, relative to the previous curve. This again, increases the total number of ideas that has a quality higher then x.
3. Increase variance : intentionally increase the number of really bad ideas
The most unique and most counter intuitive way. Increase the number of bad ideas you generate. I mean not just bad. I mean really bad ideas. Let me explain.
So the final way to increase the number of ideas above quality x, is to find ways to decrease the number of ideas around the ‘average’ (mean : center of the curve) and increase ideas that ends up in the fare right side of the bell curve. Increasing the variance. Effectively flattening the curve.
But how do you do that?
Recall what Paul Graham said about good ideas. Great ideas look a lot like bad ideas. The point is that no one can really distinguish good and bad ideas. Yes, some experienced people may be able to do so with higher probability, but as history has shown, breakout success (unicorns) have been made with ideas that were considered ‘bad’ and its these microscopically small number of startups that are built on ideas that look bad but turn out to be great ideas that shoot through the stratosphere. The better the idea, the more extreme this gets. The better the idea the worse the idea initially looks like.
Back to our curve. Given the disguise ideas can wear, ideas that you would put on the far right of the curve could actually be on the opposite side of the curve, and vis versa.
Its really hard to come up with great ideas, but, many would argue, a bit easier to come up with bad ideas.
So, if you (or your team) start coming up with really bad ideas. I mean REALLY bad ideas, they may actually be great ideas in disguise, and in turn increase the number of ideas above quality x.
 I can’t find a direct reference to it since Peter Thiel drew the vendiagram during a YC talk and was not recorded. https://www.businessinsider.com/this-venn-diagram-shows-why-startup-investing-is-hard-you-have-to-be-on-bad-ideas-2012-9