/jp/ - Otaku Culture

All three in USA..

Using a standard statistical test for the difference in means (see Hogg, McKean, Craig, p.274), I computed a probability of 55.6% that any given nightshift curry witnessed proportionally more flavor than any given non-nightshift curry.

For you keeners saying "Hey! How can 95.8% and 85.4% be bigger than the 99.7% overall curry ratio?", I'll explain. I'm now averaging the curry ratio per meal, not the overall curry versus overall meals. Consider:

Meal 1: Curry. Curry ratio: 100%
Meal 2: Ramen. Curry ratio: 1%.
Non-Curry meals comprise 1.98% of the posts, but the average curry ratio is 85.5%.

I deliberately made the decision not to weigh the curry by the type of ingredients used, since each different type is delicious, regardless of what is in it. Admittedly, this point can be argued.