> Imperfect models don't scare me.

I think that is the point: models are always imperfect. One of the ideas of a scientific theory is that is should be predictive. And at one time, if a theory's predictions were not correct, that meant the theory was not valid. Evolution was dismissed as "not predictive." And while it is true that no one can predict what evolution *will* produce, there are still general principles that hold true. We can see how life has evolved, and imagine what it may look like in ten thousand years from now. 

The same holds true for weather predictions: we can't predict the weather with any degree of accuracy, as it is inherently unpredictable. But when a freeze warning is issued, people are motivated to act. In some parts of the country, people have tornado cellars and listen for tornado warnings. I remember one time a tornado warning was issued here (Upstate NY) and a bunch of us went to a hilltop overlook to see if we could see it. No one thought it was really a threat.

Some sciences such as physics, chemistry, rocket science, etc. conform well to mathematical predictions. Others, like ecology, evolution, weather, psychology, etc. not so well. That doesn't mean that simulations, statistics and sampling have no value here. It means we have to constantly remember we are dealing with probability. These are inherently unpredictable due to a profusion of variables. Scientific studies usually attempt to minimize variables so that the results are meaningful. Real life is about uncontrollable variables. Get used to it.

Relevant quote:

> Although the quantity and quality of single‐cell data have progressed rapidly, making quantitative predictions with single‐cell stochastic models remains challenging. The stochastic nature of cellular processes leads to at least three challenges in building models with single‐cell data: 1) because variability in single‐cell data can be attributed to multiple different sources, it is difficult to rule out conflicting mechanistic models that explain the same data equally well; 2) the distinction between interesting biological variability and experimental variability is sometimes ambiguous; 3) ) the non‐standard distributions of single‐cell data can lead to violations of the assumption of symmetric errors

Hsu, I. S., & Moses, A. M. (2021). Stochastic models for single‐cell data: Current challenges and the way forward. The FEBS Journal.

See also

Hening, A., Nguyen, D. H., & Chesson, P. (2020). A general theory of coexistence and extinction for stochastic ecological communities. arXiv preprint arXiv:2007.09025.

Legault, G. B. (2017). The impacts of demographic stochasticity on populations and communities (Doctoral dissertation, University of Colorado at Boulder).

             ***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software.  For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html