Danger Finds Bourbon




Recent Activity:
Tasting NotesKoval 4 Grain                                                           April 8, 2014
Tasting NotesOld Blowhard 26yr                                                April 6, 2014
Tasting NotesMellow Corn                                                           April 6, 2014
Tasting NotesOld Scout 10yr                                                        April 6, 2014
Tasting NotesWillett "pre-fire" guest review!                     March 5, 2014
AnalysisTable of computer generated whiskey picks       March 4, 2014
OpinionsSite is International!                                                     March 1, 2014
ExperimentBrutally random blend                                            February 28, 2014
Tasting NotesWilliam Larue Weller 2005                                February 28, 2014
Tasting NotesHirsch Rye (25yr)                                                   February 28, 2014
Tasting NotesBlanton's Straight from the Barrel                 February 28, 2014
Tasting NotesWillett Family Reserve Rye (24yr)                 February 28, 2014
Tasting NotesThomas H. Handy Sazerac 2013                       February 13, 2014
- PhotosSnow and rye mix, apparently                                      February 13, 2014
Tasting NotesGeorge Dickel Rye                                                February 13, 2014



(all whiskey taken neat unless otherwise specified)


I’m a professional research scientist by day and an amateur whiskey drinker by night, and this site is where I combine those two passions. My approach is to record basic information, tasting notes and rankings to better appreciate each US whiskey I try individually, and then to use those reviews as data to get insight about comparisons between different whiskey. Either through the notes above or my analysis of them below I hope you find my thoughts useful and that you find the whiskey you are looking for. - Danger






What whiskey should I buy? Let the quality vs price plot guide your decision.

Explanation of the Value-Breakdown plot: Along the horizontal axis is the whiskey's rank, that is, the quality of that whiskey. The numbers stand for each of the seven different whiskey ranks in the tasting notes spreadsheet (see the 'whiskey rankings' tab for more detail). I have added jitter to the quality axis to break apart some of the clumps for easier visibility. Along the vertical axis is how much a 750ml bottle of that whiskey costs (at the time I wrote the review). So, the best values (low cost paired with high quality) are labels in the lower right quadrant, while the worst values (high cost paired with low quality) are labels in the upper left quadrant. Finally, if a whiskey's bubble is blue I have tried it at least twice and am more confident in my review, if it is orange I have tried it only once. The plot shows that the trend, understandably, is that the better quality a whiskey the more it costs.






Have you ever wondered which flavors are most common in bourbon? The following graph shows the frequency of the most common flavors in my reviews. This is an indication of which flavors are most prevalent in bourbon and what flavors I am most prone to detecting.




If you are interested in the flavor profiles from particular bourbon producers, the graphs below show the top few flavors from the producers I have sampled at least three times.




I have written an algorithm to automatically compare whiskies to each other based on my tasting notes. Look up a whiskey you like in the table below and see what other whiskies the algorithm recommends for you. Next to each pick is a rough estimate of the algorithm's confidence in that suggestion. If you want more insight into the algorithm look here.

Algorithm Whiskey Picks




We can also look at the correlations (the strength of linear relationship, with values near 0 meaning no correspondence between variables) among important whiskey factors such as quality (ranks I give), price, how long my reviews are, age, proof and the number of flavors detected in the whiskey [left]. The raw distributions of these factors in relation to one another are also informative [right].

Additionally, partial correlations with respect to price (that is, correlations with the effect of price eliminated) are given in parentheses in the correlation table. It is interesting to note that both age and proof have significant correlation reductions with rank when price is accounted for, indicating that price is the dominant factor correlating with rank. Also, the correlation between age and proof is completely eliminated, meaning that there is no real correlation at all between age and proof of a whiskey (after being bottled and sold), rather it is evidence that proof and age are factors responsible for setting price points but that age does not affect the proof at which a whiskey is sold or vice versa. Interestingly, the number of separate flavors correlates with rank (even corrected for price) but not with proof or age, suggesting that there is no simple linear relationship between 
the number of separate flavors in a whiskey and the age or proof of the whiskey. (3sigma outliers removed)



Acknowledgements:


Contact me: DangerBourbon [at] gmail.com