The world as a kind of clockwork, at its core simple and predictable - that was a modern idea in the eighteenth century. In such a world, science provided a method of being able to see into the future on the basis of data and laws - in principle as precisely as required, as long as the data quality is correct. Great thinkers were supporters of this idea, Isaac Newton for example or John Stuart Mill. The idea of a predictable world is without question quite magical. However, today, around 300 years later, we know that it is wrong. Because complex systems of the kind that unfortunately dominate the world work very differently.
Complex systems require unpleasant things like statistics, and require tedious distinctions like that between correlation and causality.
Countless influences and factors interact in them.
Some of them are sensitive to the smallest changes in parameters.
The opaque whole cannot be properly reduced to manageable parts.
The beautiful black and white world of truths and certainties is frayed into ugly gray tones of probabilities and error estimates.
When “model predictions” fail
This is disappointing. The resulting frustration can still be felt today. One can recognize them, for example, in the accusation against the epidemiological modelers that their “model predictions” did not come true. Also in the demand to make the models “better”, to integrate more parameters, more factors. This gives rise to the hope that exact predictability is something achievable if you just make a little effort. That you can leave the statistics, the imponderables, behind and gain security - because who would want to make decisions on the basis of uncertain statements?
One can try to admit that the study of complex systems can hardly produce more classic collateral than selling a declaration of bankruptcy.
One can refuse to take note of scientific results until science “has proven beyond doubt” that anything is the case.
The more productive attitude is to recognize that our modern understanding of our complex world is also leading to a changed approach to scientific results.
And that's where the much-invoked difference these days between scenarios and predictions comes into play.
In a space of possible futures
When we have to make decisions in situations that are characterized by uncertainty and incomplete knowledge, it is usually not an option to simply wait until we know more. This applies to acute catastrophes such as stock deals. Instead, you replace your ignorance with assumptions and think about what would result from it. What would happen if everything was in your own favor? What would be the worst that could happen? Then: what risk am I willing to take? In short, you collect scenarios and then look for a robust strategy that will lead to an acceptable result in as many of the likely cases as possible. No prognoses and predictions are required for this procedure. You operate in a space of possible futures. Needless to say, everything is doneto prevent a worst-case scenario from occurring.
Anyone who wants to criticize political decisions can do so constructively at this point. He can complain, for example, that he would have understood something different from the government by an acceptable result. Or that decision-making processes have not been made transparent enough. However, the criticism that certain singular model predictions did not occur, or in general: that decisions were made in a situation of uncertainty, does not help here in its methodological lack of information, but only increases - whether negligently or deliberately - the already high level of diffuse public dissatisfaction .
This also applies to the fundamentally very justified criticism of the empirical data situation, which is still often unsatisfactory. If such a criticism, as it was presented to the public by prominent authors of the data in the DIVI intensive care register last week, is bursting with statistical inaccuracies and errors, then little is gained. Instead, it creates the impression that the credibility of science is publicly undermined in favor of personal vanity. A recent study from Canada shows how extremely dangerous this is. Accordingly, anti-intellectualism plays an important role in the public's response to the pandemic. Not feeding them any further should be important to all of us, also with a view to future crises.