That reminds me. When we eventually replaced the mimeo with a Multigraph a co-worker came up with the slogan "Ink Don't Stink"! We all got a t-shirt.
Every Monday I tell my coworkers about how LUFC did over the weekend. They smile politely and are suddenly very interested in their work.
My coworkers and I spend most of our time discussing whether or not a particular outcome variable that is infinitely divisible and bounded at zero (with a skewed distribution) should be estimated using OLS or a ML (gamma distribution) estimator. See, we both know the OLS assumption of linearity in parameters is violated, but it's convincing others that we're not just adding needless complexity (and losing direct interpretability) to show off our abilities. Transforming the variable - the other option - leads to its own basket of problems. Thoughts?
I suppose it depends on whether the estimator is distributed on a lattice or underlying topology consistent with infrastructure evolution, and piece-wise linear function of two underlying state variables, provided these two variables have a bivariate lognormal distribution.
Roadkit is off the rails. What you said makes sense. Your 1st paragraph is the same formula me and my compadres use when we divvy up the take out pizza.
Even Trump defenders in my office make fun of the management of the white house staff. It still does not matter, is not big deal.
I've checked a couple of times each year to see how Leeds was doing. Hope you all get promoted and can become relevant again!