> While stochastic models present us with numerous difficulties, an even more perplexing conundrum faces contemporary scientists and engineers who wish to model highly complex systems involving hundreds or thousands of variables and model parameters. Owing to their sheer number, many model parameters cannot be accurately estimated via experiment and are left uncertain. As a consequence of uncertainty, for each different set of possible values for the unknown parameters, there is a different model—possibly an infinite number of models.
> Confronting the problems of complexity, validation, and model uncertainty, I have previously identified four options for moving ahead: (1) dispense with modeling complex systems that cannot be validated; (2) model complex systems and pretend they are validated; (3) model complex systems, admit that the models are not validated, use them pragmatically where possible, and be extremely cautious when interpreting them; (4) strive to develop a new and perhaps weaker scientific epistemology.
At the moment I am in favor of option 3, though option 4 might be more appealing in the future.
This isn't an easy option to take. Recently I had an article accepted for publication where I basically argued that no models, including my own, were truly validated because none fit a non-naive data set well. (In another paper I argued that most data sets used for validation are too easy to match because they don't cover the parameter space well.) A reviewer recommended rejection, basically saying that because the model isn't validated, it shouldn't be published. So much for being intellectually honest!
The paper was eventually accepted after I made it more clear that none of the popular models work that well (some are absolutely terrible in my view), and that my model improves on the status quo in a few ways.
Note that this situation isn't exactly the same as that described in the link. In the case of my article, I think we can get enough data to validate a model. We just don't have that data at present.