The success of longshot sports betting doesn't automatically stop military and defense market outliers; these are entirely different beasts operating on distinct information streams. What works for predicting a football upset isn't the trigger for a defense contract outcome, because the size of data and its invalidation vectors are poles apart. You can't just slap a sports betting model onto geopolitical events and expect it to predict, say, a coup or a major arms deal with any accuracy. My experience tracking even public tenders here in Morocco shows that the information asymmetry in defense is on a completely different level than a football match; one is about public stats, the other about highly classified intelligence and state actors.