Wednesday, November 29, 2006

Drunkard Club: Monty Python's Holy Ail

I review books and other written materials here, why not booze, too? I'll use a 5-star system for these reviews, with no 1/2 marks allowed.

A little while ago, I bought a 500 mL bottle of Monty Python's Holy Ail, made by a brewery in the UK named Black Sheep. While I doubt their claims that it was "tempered over burning witches", it did taste quite good. An enjoyable misspelled accompanyment to a meal.

I give it 3 stars.

A word about my rating system. Carlo and I have discussed (useless) rating systems for things like video games many times in the past. We've come to the agreement that the average rating across many reviews should be approximately equal to the mathematical mean of the total possible set of numerical reviews - that is, if the review scale is out of 10, the average review should be 4 [(0+1+2+3+4+5+6+7+8+9+10)/11]. Keeping the average review score low allows finer distinction among the very best reviews - which is where you should care about it; why would you care that something terrible is significantly better than something else terrible? No, you're interested in fine distinctions between the best and the very best. Most reviewers end up with average scores well above the mathematical mean, indicating that an excellent review really isn't much better than an average review, and leaving extraneous room for distinction at the bottom of the scale. I'll try to avoid that here, and hopefully my booze reviews will end up with a mean score of about 2.5.
The argument could be made that reviewers screen the things they're reviewing before consumption and review, which eliminates most of the shit at the bottom. The counter to this argument is that they should then be adjusting their expectations such that the mean still comes out in the middle. In other words, if I avoid drinks I expect to be bad, I should allocate stars to drinks with that in mind - a score of 1 star in my system should thus be a drink that's below average among drinks I've reviewed, and may not necessarily be awful on an absolute scale. Zeros will be reserved for those drinks that escape screening (Type I errors: false positives) and are really bad.

No comments: