Antifragility

In his book, Taleb stresses the differences between antifragile and robust/resilient. "Antifragility is beyond resilience or robustness. The resilient resists shocks and stays the same; the antifragile gets better."[1] The concept has now been applied to ecosystems in a rigorous way.[17] In their work, the authors review the concept of ecosystem resilience in its relation to ecosystem integrity from an information theory approach. This work reformulates and builds upon the concept of resilience in a way that is mathematically conveyed and can be heuristically evaluated in real-world applications: for example, ecosystem antifragility. The authors also propose that for socio-ecosystem governance, planning or in general, any decision making perspective, antifragility might be a valuable and more desirable goal to achieve than a resilience aspiration. In the same way, Pineda and co-workers[18] have proposed a simply calculable measure of antifragility, based on the change of “satisfaction” (i.e., network complexity) before and after adding perturbations, and apply it to random Boolean networks (RBNs). They also show that several well known biological networks such as Arabidopsis thaliana cell-cycle are as expected antifragile.

An adaptive system is one that changes its behavior based on information available at time of utilization (as opposed to having the behavior defined during system design). This characteristic is sometimes referred to as cognitive. While adaptive systems allow for robustness under a variety of scenarios (often unknown during system design), they are not necessarily antifragile. In other words, the difference between adaptive and antifragile is the difference between a system that is robust under volatile environments/conditions, and one that is robust in a previously unknown environment.

Taleb proposed a simple heuristic[19] for detecting fragility. If ${\displaystyle f(a)}$ is some model of ${\displaystyle a}$, then fragility exists when ${\displaystyle H<0}$, robustness exists when ${\displaystyle H=0}$, and antifragility exists when ${\displaystyle H>0}$, where

${\displaystyle H={\frac {f(a-\Delta )+f(a+\Delta )}{2}}-f(a)}$.

In short, the heuristic is to adjust a model input higher and lower. If the average outcome of the model after the adjustments is significantly worse than the model baseline, then the model is fragile with respect to that input.

The concept has been applied in business and management,[20] physics,[5] risk analysis,[4][21] molecular biology,[7][22] transportation planning,[8][23] urban planning,[24][25] engineering,[11][12][26] aerospace (NASA),[13] megaproject management,[27] computer science,[11][14][15][28][29] and water system design. [30]

In computer science, there is a structured proposal for an "Antifragile Software Manifesto", to react to traditional system designs.[31] The major idea is to develop antifragility by design, building a system which improves from environment's input.

• Complexity theory and organizations
• Hygiene hypothesis
• Information management
• Structures of organizations
  • Nodal organizational structure
• Systems theory
  • Systems engineering
  • System accident