Pages

Saturday, February 03, 2024

Wine Rating Scale

I am so tired of the numeric wine-rating scale popularized by Robert Parker. It is meaningless, every bottle wants to be 89-93. Some ideas for improvement.

Overall

That would be the Robert Parker-style score. But even in this category, we need to re-calibrate. 100 points is way too many. 20 would probably suffice. 

1: Actively disagreeable.

  • 2: Utterly bland and uninteresting. Someone who is utterly indifferent to wine and in the mood for an alcoholic beverage might drink it, but for anyone with even a modicum of wind appreciation, not worth the calories.
  • 3: Would drink as a last choice if in the mood to drink wine and no alternative.
  • 4-15: Relative ratings for wines most people are likely to drink--max price $50/bottle.
  • 16-20: for the elite.

As an Exemplar of Its Category

Assuming the wine fits a well-established category (e.g., New Zealand Sauvignon Blanc), how well does it express the qualities expected of the category? Sort of like how a dog show judges an animal's features not in the absolute, but relative to its breed.

Under the Right Circumstances

Most likely factor to influence this would be food pairing. E.g., this wine is generally rough, but with really spice food, it holds up. This category is a bit loose, if not careful, it could be an excuse for grade-inflation.

Value

Graded on the curve of price.

Multi-Factor Console for Updating Critical Business Parameters Executed by Software

Major business policies, especially in financial services, are often driven by some simple parameters that are executed in software. E.g., for a lender, what interest rate is being charged.

There is typically a dashboard accessible to a small number of users with authorized elevations, to enter these values. In general, that works fine, but there may be concerns about data-entry mistakes or deliberate sabotage (less probable). I think a good solution involves:

  • Purpose-built UI (console), limited of course to an appropriate, small user group.
  • Require multi-factor -entry to implement a change: i.e., 2 or more members have to independently submit the update (exact number is configurable, depending on the sensitivity--but in general, I think 2 is the magic number).
  • Add some UX niceties for workflow—e.g., showing pending updates where only 1 entry has been made, reminders if 2nd entry is overdue.
  • Email the entire group whenever a change is initiated and completed.