I've recently decided to branch out with a new review site: WhiskyAnalysis.com
No, I'm not going to be doing runtime testing of whiskies. In fact, I'm not generally going to be doing detailed reviews of individual whiskies at all. Instead, I'm using my statistical training to do a proper scientific meta-analysis of whisky flavours and quality, based on a controlled lexicon derived from the descriptions and scores given by actual whisky reviewers with extensive experience.
The goal of this new site is really to help you make sense of whiskies, to aid in helping you choose ones you may be interested in trying – based on your personal preferences. There is a similar problem in the whisky world as there is with flashlights, given all the marketing and hype that surrounds whisky production and sale. But in some ways, the need is even greater for whiskies, given the very subjective nature of taste. And unfortunately, even well-meaning enthusiasts have further muddied the waters, by perpetuating myths and fictions that were specifically created to mislead the whisky drinker.
Most people (including most whisky reviewers) simply assume that whiskies are too complex to be amenable to any sort of component analysis. It is normal to figure that taste preferences are just too idiosyncratic and variable, denying any attempt at classification. But that default assumption can be explicitly tested using scientific methods – and as I will show on this new site, it can be demonstrated to be wrong. You can actually integrate all the available information on whiskies (flawed though each piece may be) and still draw very meaningful inferences and classifications.
If you are interested to learn about statistical principles like Z-score normalization, cluster analysis and principal component analysis, this new site will be the place for you. But even you don't care about the "how" in the analysis, you will likely still find the results very helpful in your own selection.
At present, there are over 350 whiskies tracked in the online database, broken down by validated flavour clusters, with a properly constructed meta-critic score derived from up to 20 reviewers.
Cheers! :wave:
P.S.: Being a canuck, I can't help but use the Canadian/Scottish/Japanese spelling of "whisky". Not to fear, I've also registered the WhiskeyAnalysis.com domain for all you who follow the American/Irish "whiskey" spelling.
No, I'm not going to be doing runtime testing of whiskies. In fact, I'm not generally going to be doing detailed reviews of individual whiskies at all. Instead, I'm using my statistical training to do a proper scientific meta-analysis of whisky flavours and quality, based on a controlled lexicon derived from the descriptions and scores given by actual whisky reviewers with extensive experience.
The goal of this new site is really to help you make sense of whiskies, to aid in helping you choose ones you may be interested in trying – based on your personal preferences. There is a similar problem in the whisky world as there is with flashlights, given all the marketing and hype that surrounds whisky production and sale. But in some ways, the need is even greater for whiskies, given the very subjective nature of taste. And unfortunately, even well-meaning enthusiasts have further muddied the waters, by perpetuating myths and fictions that were specifically created to mislead the whisky drinker.
Most people (including most whisky reviewers) simply assume that whiskies are too complex to be amenable to any sort of component analysis. It is normal to figure that taste preferences are just too idiosyncratic and variable, denying any attempt at classification. But that default assumption can be explicitly tested using scientific methods – and as I will show on this new site, it can be demonstrated to be wrong. You can actually integrate all the available information on whiskies (flawed though each piece may be) and still draw very meaningful inferences and classifications.
If you are interested to learn about statistical principles like Z-score normalization, cluster analysis and principal component analysis, this new site will be the place for you. But even you don't care about the "how" in the analysis, you will likely still find the results very helpful in your own selection.
At present, there are over 350 whiskies tracked in the online database, broken down by validated flavour clusters, with a properly constructed meta-critic score derived from up to 20 reviewers.
Cheers! :wave:
P.S.: Being a canuck, I can't help but use the Canadian/Scottish/Japanese spelling of "whisky". Not to fear, I've also registered the WhiskeyAnalysis.com domain for all you who follow the American/Irish "whiskey" spelling.