Hacker News Reader

Comment by marcus_holmes | Hacker News Reader

Simple science: you need a control group.

The whole "25 things that successful people do" completely ignores the entire question of whether unsuccessful people also do those things, and therefore whether those things have anything to do with success.

Survivor bias is so prevalent in any thinking about "what makes someone successful?" that yes, it would be much more effective at this point to consider "what makes someone unsuccessful".

I'm willing to bet that the things we consider to be signs of success are actually table-stakes, in that both successful and unsuccessful founders do them, and the thing that actually marks the difference between successful and unsuccessful founders is luck.

View on HN

> Simple science: you need a control group.

While controlled experiments are the ideal for hypothesis testing, as someone who does stats for a living it's worth point out observational studies do exist.

Entire fields, such as anthropology, depend entirely on observational statistics. The field has also come a tremendously long way in recent years with lots of very useful tools coming out of the Bayesian community (applying proper priors, multi-level modeling, inference on more complex models) and the causal inference community.

McElreath's Statistical Rethinking is a great text that deals mostly with what I would consider the current state of the art in observational techniques. Highly recommended if isn't already on your bookshelf.

I get that, but just studying one sample of the population, deliberately ignoring any other grouping in that population, and then drawing conclusions for the entire population, is still ludicrous.

I’m gonna go even further and say that using statistics and quantifying any of this is a futile effort. There are too many variables and noise-factors to make any sense in statistics.

Every company is unique and problems are different and in different domains at different external conditions.

Statistics would just be a sample size of 1 and std dev of infinity.

"We have found that 100% of successful people breathe at least once a minute"

Successful founders choose GCP over AWS or Azure.

I’m betting a ton read those books and articles and try to do the 25 things and still fail.

My hunch is that there would be an audience for this stuff. I think the key to this is the go to market model for such content.

There's a pretty established model regarding successful founders who then write books about themselves. The audience seems to be well known and if not presumably a successful founder could self-publish their way to success.

The model for mediocre founders talking about their failures is probably a tough sell to backers as the audience size is unknown and if the founder is unsuccessful they won't be able to self publish.

The best format would probably be a free podcast and where some host would invite other moderately/less than successful founders to talk about their ups and downs. The best bet for the host would be some VC funded moderately successful founder who's startup only 1.5x'd instead of 10x'd snd had to sell the company and now had free time.

Reply on news.ycombinator.com