WINETASTER ON 06/04/12 WITH 8 JUDGES AND 8 WINES BASED ON RANKS, IDENT=N Copyright (c) 1995-2012 Richard E. Quandt, V. 1.65
FLIGHT 1: Number of Judges = 8 Number of Wines = 8
Identification of the Wine: The judges' overall ranking:
Wine A is Gigondas Le Goubert 1998 ........ 7th place Wine B is Châteauneuf du Pape Perrin 1999 ........ 8th place Wine C is Gigondas Le Goubert 1995 tied for 3rd place Wine D is Gigondas Le Goubert 1999 ........ 2nd place Wine E is Châteauneuf du Pape, Chapoutier 1998 ........ 1st place Wine F is Châteauneuf du Pape Vieux Télegraph 1995 ........ 5th place Wine G is Châteauneuf du Pape Vieux Télegraph 2003 ........ 6th place Wine H is Gigondas Le Goubert 2003 tied for 3rd place
The Judges's Rankings
Judge Wine -> A B C D E F G H Alexa 4. 3. 8. 5. 1. 7. 2. 6. Mike 5. 8. 7. 1. 2. 6. 4. 3. Orley 5. 7. 6. 1. 2. 3. 8. 4. Burt 6. 8. 1. 2. 4. 3. 7. 5. Zaki 7. 8. 1. 2. 3. 4. 5. 6. Bob 5. 6. 7. 4. 1. 2. 8. 3. Frank 7. 8. 3. 4. 1. 6. 5. 2. Dick 7. 3. 2. 4. 1. 8. 5. 6.
Table of Votes Against Wine -> A B C D E F G H
Group Ranking -> 7 8 3 2 1 5 6 3 Votes Against -> 46 51 35 23 15 39 44 35
( 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 = 0.3757
The probability that random chance could be responsible for this correlation is quite small, 0.0037. 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 Frank 0.8193 Zaki 0.6946 Mike 0.6667 Burt 0.6667 Orley 0.6587 Bob 0.3571 Dick 0.2156 Alexa -0.1198
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 E is Châteauneuf du Pape, Chapoutier 1998 2. ........ 2nd place Wine D is Gigondas Goubert 1999 --------------------------------------------------- 3. tied for 3rd place Wine C is Gigondas Le Goubert 1995 4. tied for 3rd place Wine H is Gigondas Le Goubert 2003 5. ........ 5th place Wine F is Châteauneuf du Pape Vieux Télegraph 1995 6. ........ 6th place Wine G is Châteauneuf du Pape Vieux Télegraph 2003 7. ........ 7th place Wine A is Gigondas Le Goubert 1998 --------------------------------------------------- 8. ........ 8th place Wine B is Châteauneuf du Pape Perrin 1999 We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 21.0417. The probability that this could happen by chance is 0.0037 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 Alexa Mike Orley Alexa 1.000 0.310 -0.119 Mike 0.310 1.000 0.667 Orley -0.119 0.667 1.000 Burt -0.643 0.238 0.595 Zaki -0.357 0.333 0.452 Bob 0.024 0.452 0.833 Frank -0.024 0.619 0.452 Dick 0.333 0.071 0.000 Burt Zaki Bob Alexa -0.643 -0.357 0.024 Mike 0.238 0.333 0.452 Orley 0.595 0.452 0.833 Burt 1.000 0.905 0.286 Zaki 0.905 1.000 0.119 Bob 0.286 0.119 1.000 Frank 0.524 0.619 0.405 Dick 0.167 0.405 -0.095 Frank Dick Alexa -0.024 0.333 Mike 0.619 0.071 Orley 0.452 0.000 Burt 0.524 0.167 Zaki 0.619 0.405 Bob 0.405 -0.095 Frank 1.000 0.452 Dick 0.452 1.000 Pairwise correlations in descending order 0.905 Burt and Zaki Significantly positive 0.833 Orley and Bob Significantly positive 0.667 Mike and Orley Significantly positive 0.619 Zaki and Frank Not significant 0.619 Mike and Frank Not significant 0.595 Orley and Burt Not significant 0.524 Burt and Frank Not significant 0.452 Mike and Bob Not significant 0.452 Frank and Dick Not significant 0.452 Orley and Frank Not significant 0.452 Orley and Zaki Not significant 0.405 Bob and Frank Not significant 0.405 Zaki and Dick Not significant 0.333 Alexa and Dick Not significant 0.333 Mike and Zaki Not significant 0.310 Alexa and Mike Not significant 0.286 Burt and Bob Not significant 0.238 Mike and Burt Not significant 0.167 Burt and Dick Not significant 0.119 Zaki and Bob Not significant 0.071 Mike and Dick Not significant 0.024 Alexa and Bob Not significant 0.000 Orley and Dick Not significant -0.024 Alexa and Frank Not significant -0.095 Bob and Dick Not significant -0.119 Alexa and Orley Not significant -0.357 Alexa and Zaki Not significant -0.643 Alexa and Burt Not significant
COMMENT: All the wines were wonderful and drinking very well and they were characteristic southern Rhône wines. It should be noted that one of the wines, the 1999 Gigondas was available only in magnumn size; its identity was unknown to all tasters except the host of the tasting. This tasting and the success of the 1999 Gigondas highlight the magnum effect on all well made wines, because it enhances the freshness of the wine. With the exception of 1998 Châteauneuf du Pape Barbe Rac by Chapoutier, which is a renowned wine, in a great vintage, in every other vintage the Châteauneuf du Papes were beaten by the Gigondas. While this is an important finding, overall, the advantage of the Gigondas was not overwhelming: the sum of the ranks for the four Gigondas wines was 139 and for the Châteauneuf du Pape 148; not a statistically significant difference. It should also be noted that several tasters remarked on the similarity of the wines; in spite of that the overall agreement was very strong, with the Kendall W coefficient of concordance 0.3757, the probability of that occurring by chance being only 0.0037. It should also be said that we were drinking a single producer, single cuvée Gigondas (Cuvée Florence). This raises the observation that if one has the opportunity to buy premium Gigondas wines, one might be wise to consider it a preferred alternative to Châteauneuf do Pape if the price is right, as is likely to be the case. Note that the Côte du Rhône appellation is the largest production in France and the Gigondas and the Chateauneuf du Pape are the two best subappelations. Therefore it is not surprising that these wines are both very good examples of the southern Rhône. It should be also noted that wine B was a negociant wine from the Perrin family. As far as the vintages are concerned, the rank sums for the two vintages across the wines were 1995 74 1998 61 1999 74 2003 79 again not an overwhelming difference, with the possible exception of 1998. Our host designed a contest (in which he did not participate because of his knowledge of which wine was the magnum) in which each of the contestants had to guess which wine was a Gigondas and which was a Châteauneuf du Pape. Multiple winners were permitted; each person who guessed the identity of six or eight wines correctly (it is obvious that it is not possible to guess the identity of exactly seven wines) was deemed a winner and was awarded a bottle of California wine. Now it is obvious that one might get quite a few wines right by guessing randomly; the probability distribution of correct guesses under the hypothesis of random guessing is as follows: Number of correct guesses Probability 0 0.01429 1 0 2 0.22857 3 0 4 0.51429 5 0 6 0.22857 7 0 8 0.01429 So, under "random guessing," an outcome of 6 or 8 correct guesses has a probability of about 0.24, which was deemed sufficient to qualify someone as a winner. Two of the tasters achieved this result, with everybody else identifying correctly five or fewer wines. Nobody guessed the identities of all eight wines correctly.
Return to previous page