Random Chance and Startups

Craig Dunk
4 min readApr 21, 2023
Doubt kills more dreams than failure. But Random Chance is, like, the grim reaper for dreams
Random chance from https://www.smbc-comics.com/?id=3506

I think about the role randomness — and incomplete information — plays in startups (and all businesses? and life?) and I tend to think it plays a bigger role — both good and bad — than people assume on a day to day basis. In some ways I get myself comfortable with entrepreneurship — which is associated with risk for many people — by recognizing that even “safe” choices have unforeseeable downsides and that for the most part entrepreneurship is just making risks visible. When it comes to chance I also kind of like the cheeky compliment to the effect “you must be very skilled to have been born in a productive and growing society with a great education system in an affluent neighbourhood” as a way to acknowledge the birth lottery and parody the idea that success entirely comes from within and doesn’t depend on others.

There are a few unproductive ways of taking this — one is a fatalism to see the future as unchangeable which just seems pessimistic to me no matter the circumstances you start in — and another is to be dismissive of peoples accomplishments from presence or absence of a particular privilege. I wanted to share a few ways that I think _are_ productive to deal with randomness and incomplete information.

Directed Opportunism is a concept I was introduced to in the Art of Action. Basically a form of planning that acknowledges distributed decision making, the role of competing agents, and the role of random chance. It is similar to a great leadership video on an alternative submarine leadership approach. What both have in common is that they emphasize being prepared by being very proficient when it comes to routine items and then emphasizing clarity and simplicity on the strategic side. This creates a situations where it is safe and effective to respond to changing situations. We have had a small version of this already in this startup: there has been preparation and planning and reasoning about the high level direction when I became aware, along with the design group that had built a bioreactor, of an incubator where the application deadline was coming up very quickly. We were in a position to assemble some information and complete that submission in the short time required because it was clear it aligned to our intent and we had the skills (the bioreactor design, experience building pitch decks, etc.) so we could prioritize it. Which is the open startup news: we have applied to Techstars Farm to Fork. It is of course a program that is oversubscribed and we might not end up participating this year — but it isn’t building in the open if you preserve your dignity by waiting until you find out if it works 🙂. This has further triggered an exploration of other accelerators, venture creation labs, and programs — since we are at that early stage that I think most benefits from these programs. Recommendations welcome.

Thinking in Bets is another book that outlines a way of still being effective when randomness and incomplete information is in play in a dynamic environment. Which basically describes Poker — and the author comes from great success in the competitive poker world. This is immensely popular in the entrepreneurship space and I think rightly so. It is rooted in stronger math of game theory and more exciting analogy that the old term “analysis paralysis”. Plus it goes beyond analysis into the idea of thriving. I regularly find myself pulling interesting framings from this book. The easiest one to apply if you are a leader is just to check quickly ‘how should my actions change based on the outcome of this” where this could be new information or an experiment you are running etc.

The last interesting way of trying to make good decisions in random, incomplete, and dynamic environments comes from the VC community. It may be widely practiced but I heard it as a story about FJLabs though I don’t have first hand experience. One would expect any VC to analyze investments they made and whether they worked out, and most I am confident go further to the full confusion matrix (false positives, false negatives) and look at investments they chose not to make and see what they missed — though maybe as regrets rather than systematic analysis. Apparently some VCs then go further still and add another axis to the matrix of “did we get the theory of the outcome correct”. This system would penalize a success if, for instance, they thought a business would work because educators would love it, if instead the customer based turned out to be researchers. And similarly calling a failure but for the wrong reasons would also be considered a decision making. What I like about this practice is that it is more directly testing the decision making and directly incorporating variance. Plus I am a big fan of continuous improvement and this seems like a powerful tool if done well.


Intended takeaways

  • Randomness (and its partner incomplete information) plays a huge role in startups — but it need not result in fatalism or random decisions
  • Art of Action book recommendation (And submarine video) highlights the role of preparedness combined with clear intent.
  • Thinking in Bets book recommendation highlights how to focus input actions on outcome reasoning. And is a cooler and more accessible presentation of decision making and game theory than most.
  • Borrowing the confusion matrix concept from Machine Learning and adding a “theory of outcome” axis could be interesting if you like to geek out on decision making. But deploying the concept in retros is easy for anyone.

Intended takeaways for the build-in-the-open startup

  • The project is in review for Techstars Farm to Fork and are looking at other accelerators and programs too. Let me know if there are programs or people I should be talking to for a venture at this early stage.



Craig Dunk

Tech leader, speculative fiction fan, parent to adult children, and a big fan of camp fires.