After reading ‘From Academia to policy makers: A Methodology for Real-Time Forecasting of Infrequent Events’ I thought of ways that the skills we acquired in our course ‘Building Near Futures’, could assist future research towards building more accurate conflict forecasting models.  The researchers in this article compared the use and accuracy of multiple forecasting models. They introduced the methodology for forecasting of real-time events (FORTE), were they compare and discuss the results from different modeling choices. A personal discovery made while reading this article is the large amount of prediction indicators that are used in these models. The researchers mentioned that more than a 100 different indicators could be used but they found and selected 86 predictor variables that gave their models the most accurate outputs. The large amount of indicators makes sense, but they also mentioned how the value of each indicator can change based off of the certain years they used to test their models as well as how far into the future they use the model to predict conflict events.: ‘There are greater variations in the quality of predictions; hence, the predictions are less stable.’

An interesting next step for research in this field would be to find ways to combat the un-stableness of predictions farther in the future. Some of the readings from our course ‘Building Near Futures’ came to mind while trying to think of possible solutions to this ‘un-stableness’ problem. Both the book ‘Speculative Everything’ and the article ‘Designing an Experiential Scenario’ discuss the importance of imaginative exploration and context-specific thinking about alternative possible near futures. Using the Experiential Futures Ladder framework, developed by the authors of the latter reading, one can imagine possible futures to increase the odds of achieving desirable futures.

As I read and further research into the methods and practices that have been and are currently being used for conflict forecasting models, I will try and spot areas of improvement and possible methods that could adopt this method of future thinking. This would allow to adjust models and use indicators that would be more valuable when predicting events that will happen further in the future. This will allow more time for policy makers and users of these models to proactively address these situations and prepare or stop them from occurring at all.

Building Near Futures for Better Performing Forecasting Models / Jantz Andersson / Knowledge Integration and Project Planning: Future Governance (2023-24) by is licensed under a