Key quote from a paper on trying to build a company of all above average people: "Strong dependence on the initial conditions; in economic terms, this means the first few hires of your start-up can have a tremendous effect!"
This paper is fascinating: http://theory.stanford.edu/~sergei/papers/hiring-soda.pdf and even though they only look at Lake Wobegon strategies for hiring and firing there are some interesting implications regarding the length of time between hires; the number of interviews needed; the need for preprocessing steps before interviewing; the impacts of mean versus median; the clustering effects around whatever cutoff is used.

I'd love to see someone tackle extending their model (hopefully with less maths and more simulation) to scenarios where the quality score of employees can be improved.


"""The additional difficulty comes from the fact that, for the strategy
of hiring above the mean, a single new hire cannot
change the mean employee quality scores significantly.
However, when hiring above the median, a single new
hire can have a drastic impact on the median employee
quality score,
...
The number of interviews required to hire an employee for the Lake
Wobegon strategies starts small but grows quickly. As
interviews are themselves expensive, the fact that the
number of interviews grows in this way suggests a potential problem with our model.
...
In reality, however, the
interview process cannot be perfect, so the interview
score will differ from the applicant’s true quality. We
may therefore hire applicants whose true quality scores
lie below the threshold prescribed by the hiring strategy,
and similarly we may reject applicants whose scores
lie above this threshold. Furthermore, it may also be
unrealistic to assume that we know the exact value of
the threshold when interviewing applicants, as this may
require more information about our employees’ quality
scores than we can exactly determine."""
Shared publiclyView activity