we need to define terms like "consistently" and "trend". but even once you do that we still have the problem that you're already explaining away exceptions. this theory is not disprovable because there is no outcome that you would say actually disproves it. you would say we just need more data.