I have a systematic way of approaching this kind of situation, where you have to rapidly estimate a thing, commit to the estimate and are judged by the quality of your estimates in the long run; my approach is to first make a guess based off my gut, and then to pause and make a bet with myself, did I guess high or low? If my gut then says that my first gut instinct was too high or low, I adjust from there. I can't guess great the first time, but this two-stage guessing works a lot better for me.

I'm sure I'm not the first to use this technique, but I don't know what it's called.