WINETASTER ON 03/02/15 WITH 8 JUDGES AND 8 WINES BASED ON RANKS, IDENT=N Copyright (c) 1995-2015 Richard E. Quandt, V. 1.65
A Tasting of 2005 Grand Cru Burgundies
FLIGHT 1: Number of Judges = 8 Number of Wines = 8
Identification of the Wine: The judges' overall ranking:
Wine A is Chambertin Rousseau tied for 5th place Wine B is Clos de la Roche Ponsot tied for 7th place Wine C is La Grande Rue Monopole tied for 5th place Wine D is Bonnes Mares Vogüe ........ 4th place Wine E is Clos de Tart ........ 3rd place Wine F is Musigny Vielles Vignes Vogüé ........ 1st place Wine G is Chambertin Clos de Bèze Rousseau tied for 7th place Wine H is Vosne Romanée Leroy ........ 2nd place
The Judges's Rankings
Judge Wine -> A B C D E F G H Orley 6. 7. 1. 5. 2. 3. 8. 4. Frank 7. 8. 6. 2. 3. 1. 4. 5. Tom 4. 6. 3. 7. 8. 1. 2. 5. Mike 8. 6. 7. 3. 4. 1. 5. 2. Zaki 3. 7. 4. 2. 8. 1. 6. 5. Bob 5. 4. 8. 6. 2. 3. 7. 1. Jerry 2. 1. 7. 6. 3. 5. 8. 4. Dick 5. 3. 4. 8. 7. 6. 2. 1.
Table of Votes Against Wine -> A B C D E F G H
Group Ranking -> 5 7 5 4 3 1 7 2 Votes Against -> 40 42 40 39 37 21 42 27
( 8 is the best possible, 64 is the worst)
Here is a measure of the correlation in the preferences of the judges which ranges between 1.0 (perfect correlation) and 0.0 (no correlation):
W = 0.1563
The probability that random chance could be responsible for this correlation is rather large, 0.2711. Most analysts would say that unless this probability is less than 0.1, the judges' preferences are not strongly related. We now analyze how each taster's preferences are correlated with the group preference. A correlation of 1.0 means that the taster's preferences are a perfect predictor of the group's preferences. A 0.0 means no correlation, while a -1.0 means that the taster has the reverse ranking of the group. This is measured by the correlation R.
Correlation Between the Ranks of Each Person With the Average Ranking of Others
Name of Person Correlation R Mike 0.3615 Bob 0.1708 Frank 0.0244 Zaki -0.1566 Tom -0.1677 Orley -0.2087 Dick -0.3571 Jerry -0.5061
The wines were preferred by the judges in the following order. When the preferences of the judges are strong enough to permit meaningful differentiation among the wines, they are separated by -------------------- and are judged to be significantly different.
1. ........ 1st place Wine F is Musigny Vielles Vignes Vogüe --------------------------------------------------- 2. ........ 2nd place Wine H is Vosne Romanée Leroy 3. ........ 3rd place Wine E is Clos dce Tart 4. ........ 4th place Wine D is Bonnes Mares Vogüé 5. tied for 5th place Wine C is La Grande Rue Monopole 6. tied for 5th place Wine A is Chambertin Rousseau 7. tied for 7th place Wine G is Chambertin Clos de Bèze Rousseau 8. tied for 7th place Wine B is Clos de la Roche Ponsot We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 8.7500. The probability that this could happen by chance is 0.2711
We now undertake a more detailed examination of the pair-wise rank correla- tions that exist between pairs of judges. First, we present a table in which you can find the correlation for any pair of judges, by finding one of the names in the left hand margin and the other name on top of a column. A second table arranges these correlations in descending order and marks which is significantly positive significantly negative, or not significant. This may allow you to find clusters of judges whose rankings were particularly similar or particularly dissimilar. Pairwise Rank Correlations Correlations must exceed in absolute value 0.74 for significance at the 0.05 level and must exceed 0.64 for significance at the 0.1 level Orley Frank Tom Orley 1.000 0.310 -0.071 Frank 0.310 1.000 0.095 Tom -0.071 0.095 1.000 Mike 0.214 0.786 0.024 Zaki 0.143 0.405 0.476 Bob 0.167 0.167 -0.333 Jerry -0.119 -0.476 -0.476 Dick -0.357 -0.548 0.357 Mike Zaki Bob Orley 0.214 0.143 0.167 Frank 0.786 0.405 0.167 Tom 0.024 0.476 -0.333 Mike 1.000 0.262 0.571 Zaki 0.262 1.000 -0.214 Bob 0.571 -0.214 1.000 Jerry -0.190 -0.286 0.595 Dick -0.143 -0.357 0.048 Jerry Dick Orley -0.119 -0.357 Frank -0.476 -0.548 Tom -0.476 0.357 Mike -0.190 -0.143 Zaki -0.286 -0.357 Bob 0.595 0.048 Jerry 1.000 -0.048 Dick -0.048 1.000 Pairwise correlations in descending order 0.786 Frank and Mike Significantly positive 0.595 Bob and Jerry Not significant 0.571 Mike and Bob Not significant 0.476 Tom and Zaki Not significant 0.405 Frank and Zaki Not significant 0.357 Tom and Dick Not significant 0.310 Orley and Frank Not significant 0.262 Mike and Zaki Not significant 0.214 Orley and Mike Not significant 0.167 Orley and Bob Not significant 0.167 Frank and Bob Not significant 0.143 Orley and Zaki Not significant 0.095 Frank and Tom Not significant 0.048 Bob and Dick Not significant 0.024 Tom and Mike Not significant -0.048 Jerry and Dick Not significant -0.071 Orley and Tom Not significant -0.119 Orley and Jerry Not significant -0.143 Mike and Dick Not significant -0.190 Mike and Jerry Not significant -0.214 Zaki and Bob Not significant -0.286 Zaki and Jerry Not significant -0.333 Tom and Bob Not significant -0.357 Zaki and Dick Not significant -0.357 Orley and Dick Not significant -0.476 Tom and Jerry Not significant -0.476 Frank and Jerry Not significant -0.548 Frank and Dick Not significant
COMMENT: As on April 2, 1012, the group once again was privileged to taste a fantastic array of top draw 2005 Burgundies. Despite one wine that in our previous tasting was thought thought to be corked, in this tasting there was no corked wine. The assessment of the group was that while all these grand cru Burgundies were drinking wonderfully well, with little discernable differ- ences between them at 10 years of bottle age, several of these wines will benefit from several more years in the cellar. It is clear that the top wine in this tasting and the corresponding tasting three years ago was the Musigny. Overall, the correlation within the group was small, with the Kendall W achieving only 0.1563, which is statistically not significant. However, the most interesting feature of the tasting was that it was nearly ideantical with a tasting held on April 2, 2012. Seven of the eight wines were identical in the two tastings for vintage and vineyard. The Nuits St. Georges in the earlier tasting could not be matched and was replaced by a Vosne Romanée from Leroy. In testing for intertemporal consistency one might proceed in two ways: (1) find the correlation between the two tastings, pretending that the non-matching winesd in each group are simply called "wine other than the seven matching wines" or (2) simply omit the non- matching wines and find the correspondences between the two sets of remaining sev en wines. We also note that seven of the eight tasters were identical in the two tastings; hence as in the wines, there was not a complete duplication of the experiment but nearly so. Using the first option, the rank correlation between the 2012 and 2015 tastings is 0.47; using the second option it is 0.36, and neither of these is statistically significant at any reasonable level of significance. It is also clear that these results are largely driven by Clos de Tart, which scored well in 2015 but was at the bottom in 2012. It is also driven by the generally low correlation between the rankings of the group members. It is all the more remarkable that the Musigny achieved such splendid results in both tastings.
Return to previous page