CFAR - An Impressive Failure
webCredibility Rating
Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.
Rating inherited from publication venue: Center for Applied Rationality
A CFAR blog post using a nonprofit case study to illustrate rationality principles around failure, pre-commitment, and experimental design; tangential to core AI safety but relevant to epistemic and organizational best practices.
Metadata
Summary
CFAR highlights Effective Fundraising's failed grant-writing experiment as a model of rigorous experimental practice, praising their public pre-commitment to success criteria, advance contingency planning, and transparent resource tracking despite sending 25 applications and receiving zero grants. The piece argues that well-structured failures yield valuable information and deserve recognition alongside successes.
Key Points
- •Public pre-commitment to explicit success criteria (raising $50,000 in 20 applications or one year) helped prevent rationalization and sunk-cost bias.
- •Having a clear Plan B before starting reduced emotional attachment to a failing strategy and made pivoting easier.
- •Transparent tracking of time and money enabled learning and allowed others to understand where resources went.
- •Negative experimental results are framed as equally valuable to positive ones when the experiment is well-designed.
- •The example illustrates applied rationality principles: calibrated stopping rules, accountability, and epistemic honesty.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| Center for Applied Rationality | Organization | 62.0 |
Cached Content Preview
An impressive failure | Center for Applied Rationality A graph of how Effective Fundraising spent their six months. I'd like to highlight a particularly impressive failure I heard about recently -- and by "impressive," I don't mean, "Wow, I'm impressed at how utterly they failed." I mean, "I'm impressed at how gracefully and wisely they failed." Effective Fundraising is a young nonprofit startup, a two-person team founded in 2013 to write grants for effective charities -- in other words, charities that have a demonstrably high positive effect on the world per dollar donated to them. EF chose The Humane League and the Against Malaria Foundation as their initial focus, and just posted their six month review of the fruits of their labors. In a nutshell, they sent out 25 applications and received zero grants, with the exception of a Google Adwords grant of free ads. So why am I impressed? Because EF's experiment did so many things right. For example: They pre-committed publicly to their criteria for success : "We announced to multiple people in public forums that if after 20 applications or after one year we had not raised $50,000, we would stop grant-writing. Studies have shown time and again that if other people know about your goals, you’re more likely to follow through." (It also, I'm sure, helps ward off the temptation to keep plugging along, rationalizing away the lack of results.) They took precautions to make sure they wouldn't feel committed to sticking with a losing strategy , if indeed it did turn out to be losing: "One reason it’s hard to admit a failed experiment is because you may be uncertain or unhappy about your other options. In fact, most people don’t even have a Plan B. Effective Fundraising grants had a clear idea from the beginning what our Plan B was if the experiment didn’t work so that if it didn’t we knew we’d still be secure and happy." They kept track of how they spent their time and money : "We found this personally useful, and it allowed us to see areas of improvement as well as let others see and understand where our hours went." So hats off to you, EF -- you've done credit to the true but often-overlooked point that negative experiments can be just as valuable as positive ones. If only other organizations and people could fail as well as you. Find out more about our Workshops Or donate to CFAR
1a39a81f52e17ba6 | Stable ID: ZGVkYzEyZG