a) This strategy (the co-op model) is overwhelmingly likely to fail
b) But every once in a great while, it will succeed
c) And some random fluke for whom it succeeded will think it's a great idea and write a blog post like this one, so
d) The existence of this blog post is not strong evidence against a)
However, given this scenario, we would expect the distribution of S (number of successful projects a given person started under the co-op model) to be heavily peaked at 0, with a few outliers at 1, and falling to ~0 at 2. This author has started three successful projects under the co-op model, suggesting that this is not just survivor bias -- the right sort of person can, at least occasionally, find success under this model.
No, survivorship bias doesn't say anything about probabilities or even the correctness of conclusion, it's a selection bias. It certainly doesn't imply probability "falling to ~0 at 2".
Anyway it doesn't really matter since this isn't a study or an attempt to prove something, I just think it's funny how well this article fits that definition so I pointed it out.
If it is probable that a person will be able to repeatedly accomplish something, looking primarily at successes isn't survivorship bias — it's just the way the numbers naturally fall out.
a) This strategy (the co-op model) is overwhelmingly likely to fail b) But every once in a great while, it will succeed c) And some random fluke for whom it succeeded will think it's a great idea and write a blog post like this one, so d) The existence of this blog post is not strong evidence against a)
However, given this scenario, we would expect the distribution of S (number of successful projects a given person started under the co-op model) to be heavily peaked at 0, with a few outliers at 1, and falling to ~0 at 2. This author has started three successful projects under the co-op model, suggesting that this is not just survivor bias -- the right sort of person can, at least occasionally, find success under this model.