[Is that the right way to cite this? It’s on Arxiv and Yaneer Bar-Yam is of course the name which stands out… and also named is Alfredo Morales, who sadly passed away before publication
Anyway, brought to my attention by Ivo Velitchkov on Mastodon who says:
https://mastodon.social/@kvistgaard@vivaldi.net/113992163851518045
“Science is in the business of explanation, not prediction. And #complexity science should know better than any other science that future is unpredictable. Yet, this paper claims it offers “practical tool for prediction”. https://arxiv.org/html/2501.07349v1“
The paper, though, is very interesting albeit it raises more questions in my mind than answers. From two data sets (order to a US manufacturing company, and references to a particular environmental concern in US legislation etc) it purports to show that the ‘sigmoid curve’ predicts both individual and whole-system behaviours – i.e. orders and references in legislation etc start slow, speed up, slow down, then stop. And that this “model is applicable for any system in which entities initiate, accelerate (engaging phase), decelerate (disengaging phase) and finally cease their activity. “
There may be some element of circularity there (no pre-registration or process note to show if other datasets were sought, if characteristics were pre-determined etc) but anyway it’s clear the model is all bottom-up causation with all factors determined within individual entities (and some rather odd references to ‘noise’). Clearly not true for systems for which is this not true, which of course is a truism. So no mention of Lindy effects, no real exogenous factors (economy? environment?), no contagion effects etc… well, if everything can be captured intrinsically then everything can be captured intrinsically….
I also wonder if there’s some sort of mirroring effect where this is just capturing the visibility of the environmental concern to lawmakers/the success of the legislation etc in dealing with the issues, and whether the manufacturer fits in to some natural lifecycle of customers… so in general I’m frustrated that there’s no holistic or contextual thinking brought in to what is otherwise Quite Intriguing].
Leila HedayatifarNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAleila@necsi.edu, yaneer@necsi.eduAlfredo J. MoralesNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USADeceasedDominic E. SaadiNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USARachel A. RiggNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAOlha BuchelNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAAmir AkhavanNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAEgemen SertDepartment of Computer Engineering, Middle East Technical University, Ankara, TürkiyeAabir Abubaker KarNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAMehrzad SasanpourDepartment of Physics, Sharif University of Technology, Tehran, IranIrving R. EpsteinNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USADepartment of Chemistry, Brandeis University, Waltham, MA 02453, USAYaneer Bar-YamNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAleila@necsi.edu, yaneer@necsi.edu
Abstract
Predicting dynamic behaviors is one of the goals of science in general as well as essential to many specific applications of human knowledge to real world systems. Here we introduce an analytic approach using the sigmoid growth curve to model the dynamics of individual entities within complex systems. Despite the challenges posed by nonlinearity and unpredictability in system behaviors, we demonstrate the applicability of the sigmoid curve to capture the acceleration and deceleration of growth, predicting an entity’s ultimate state well in advance of reaching it. We show that our analysis can be applied to diverse systems where entities exhibit nonlinear growth using case studies of (1) customer purchasing and (2) U.S. legislation adoption. This showcases the ability to forecast months to years ahead of time, providing valuable insights for business leaders and policymakers. Moreover, our characterization of individual component dynamics offers a framework to reveal the aggregate behavior of the entire system. We introduce a classification of entities based upon similar lifepaths. This study contributes to the understanding of complex system behaviors, offering a practical tool for prediction and system behavior insight that can inform strategic decision making in multiple domains.
keywords:
Sigmoid model, Accelerating and Decelerating phases, Lifepath
Predicting System Dynamics of Universal Growth Patterns in Complex SystemsLeila HedayatifarNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAleila@necsi.edu, yaneer@necsi.eduAlfredo J. MoralesNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USADeceasedDominic E. SaadiNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USARachel A. RiggNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAOlha BuchelNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAAmir AkhavanNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAEgemen SertDepartment of Computer Engineering, Middle East Technical University, Ankara, TürkiyeAabir Abubaker KarNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAMehrzad SasanpourDepartment of Physics, Sharif University of Technology, Tehran, IranIrving R. EpsteinNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USADepartment of Chemistry, Brandeis University, Waltham, MA 02453, USAYaneer Bar-YamNew England Complex Systems Institute, 125 Mount Auburn St., Box 380762, Cambridge, MA 02138, USAleila@necsi.edu, yaneer@necsi.eduAbstractPredicting dynamic behaviors is one of the goals of science in general as well as essential to many specific applications of human knowledge to real world systems. Here we introduce an analytic approach using the sigmoid growth curve to model the dynamics of individual entities within complex systems. Despite the challenges posed by nonlinearity and unpredictability in system behaviors, we demonstrate the applicability of the sigmoid curve to capture the acceleration and deceleration of growth, predicting an entity’s ultimate state well in advance of reaching it. We show that our analysis can be applied to diverse systems where entities exhibit nonlinear growth using case studies of (1) customer purchasing and (2) U.S. legislation adoption. This showcases the ability to forecast months to years ahead of time, providing valuable insights for business leaders and policymakers. Moreover, our characterization of individual component dynamics offers a framework to reveal the aggregate behavior of the entire system. We introduce a classification of entities based upon similar lifepaths. This study contributes to the understanding of complex system behaviors, offering a practical tool for prediction and system behavior insight that can inform strategic decision making in multiple domains.keywords: Sigmoid model, Accelerating and Decelerating phases, Lifepath
Predicting System Dynamics of Universal Growth Patterns in Complex Systems
Systems Community of Inquiry 