WINETASTER ON 04/03/00 WITH 8 JUDGES AND 6 WINES BASED ON RANKS, IDENT=N Copyright (c) 1995-2000 Richard E. Quandt

FLIGHT 1: Number of Judges = 8 Number of Wines = 6

Identification of the Wine: The judges' overall ranking:

Wine A is Musigny Leroy 1990 ........ 4th place Wine B is La Tache 1990 ........ 5th place Wine C is Romanee St. Vivant Leroy 1990 ........ 6th place Wine D is Richebourg Leroy 1990 ........ 3rd place Wine E is Clos de la Roche Leroy 1990 ........ 2nd place Wine F is Chambertin Leroy 1990 ........ 1st place

The Judges's Rankings

Judge Wine -> A B C D E F John 6. 3. 2. 4. 5. 1. Bob 1. 6. 5. 3. 4. 2. Orley 2. 3. 6. 4. 5. 1. Ed 6. 4. 5. 2. 1. 3. Frank 3. 2. 6. 5. 4. 1. Burt 4. 6. 5. 3. 2. 1. Bruno 1. 5. 4. 3. 2. 6. Dick 5. 3. 4. 2. 1. 6.

Table of Votes Against Wine -> A B C D E F

Group Ranking -> 4 5 6 3 2 1 Votes Against -> 28 32 37 26 24 21

( 8 is the best possible, 48 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 = 0.1482

The probability that random chance could be responsible for this correlation is rather large, 0.3132. 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 Burt 0.9429 Frank 0.4058 Orley 0.2571 Bob 0.2571 Ed 0.1429 Bruno -0.2029 John -0.2609 Dick -0.3189

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 Chambertin Leroy 1990 2. ........ 2nd place Wine E is Clos de la Roche Leroy 1990 3. ........ 3rd place Wine D is Richebourg Leroy 1990 4. ........ 4th place Wine A is Musigny Leroy 1990 5. ........ 5th place Wine B is La Tache 1990 --------------------------------------------------- 6. ........ 6th place Wine C is Romanee St. Vivant Leroy 1990 We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 5.9286. The probability that this could happen by chance is 0.3132 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 0.89 for significance at the 0.05 level and must exceed 0.83 for significance at the 0.1 level John Bob Orley John 1.000 -0.314 0.086 Bob -0.314 1.000 0.600 Orley 0.086 0.600 1.000 Ed 0.029 -0.143 -0.200 Frank 0.200 0.257 0.886 Burt 0.086 0.600 0.314 Bruno -0.943 0.371 -0.257 Dick -0.429 -0.486 -0.657 Ed Frank Burt John 0.029 0.200 0.086 Bob -0.143 0.257 0.600 Orley -0.200 0.886 0.314 Ed 1.000 -0.029 0.600 Frank -0.029 1.000 0.257 Burt 0.600 0.257 1.000 Bruno -0.086 -0.429 -0.029 Dick 0.657 -0.486 -0.086 Bruno Dick John -0.943 -0.429 Bob 0.371 -0.486 Orley -0.257 -0.657 Ed -0.086 0.657 Frank -0.429 -0.486 Burt -0.029 -0.086 Bruno 1.000 0.371 Dick 0.371 1.000 Pairwise correlations in descending order 0.886 Orley and Frank Significantly positive 0.657 Ed and Dick Not significant 0.600 Ed and Burt Not significant 0.600 Bob and Burt Not significant 0.600 Bob and Orley Not significant 0.371 Bob and Bruno Not significant 0.371 Bruno and Dick Not significant 0.314 Orley and Burt Not significant 0.257 Bob and Frank Not significant 0.257 Frank and Burt Not significant 0.200 John and Frank Not significant 0.086 John and Orley Not significant 0.086 John and Burt Not significant 0.029 John and Ed Not significant -0.029 Burt and Bruno Not significant -0.029 Ed and Frank Not significant -0.086 Ed and Bruno Not significant -0.086 Burt and Dick Not significant -0.143 Bob and Ed Not significant -0.200 Orley and Ed Not significant -0.257 Orley and Bruno Not significant -0.314 John and Bob Not significant -0.429 Frank and Bruno Not significant -0.429 John and Dick Not significant -0.486 Frank and Dick Not significant -0.486 Bob and Dick Not significant -0.657 Orley and Dick Not significant -0.943 John and Bruno Significantly negative

COMMENT: Fact 1: The tasting initially was meant to compare 8 wines, two of which were 1959 Leroy grand crus. After beginning the tasting, several people agreed that the 1959s were obviously different from the 1990s. So, it seemed much more sensible to create two flights of wines. All but one person preferred the 1959s to the 1990s. Fact 2: The wines of both vintages were of extraordinary high quality, every one in excellent condition. The Romanee St. Vivant was the weakest wine of the tasting, and it is generally regarded as the weaker of the Leroy wines in this vintage. However, Robert Parker still rates this wine 96 points, and nobody dis- agrees that this wine is of high quality. The groups consensus favorites of the 1990 vintage, Chambertin Leroy and Clos de la Roche Leroy, were also the wines as ranked as perfect 100 by Parker. One taster remarks that the common wisdom to the effect that Burgundys are victimized by excessive yields and diminished extract, are defined by these extraordinary wines in their richness and multi-dimensional complexity.

Return to previous page

Report

WINETASTER ON 04/03/00 WITH 8 JUDGES AND 2 WINES BASED ON RANKS, IDENT=Y Copyright (c) 1995-2000 Richard E. Quandt

FLIGHT 2: Number of Judges = 8 Number of Wines = 2

Identification of the Wine: The judges' overall ranking:

Wine A is Richebourg Leroy 1959 ........ 1st place Wine B is Musigny Leroy 1959 ........ 2nd place

The Judges's Rankings

Judge Wine -> A B John 1. 2. Bob 1. 2. Orley 1. 2. Ed 2. 1. Frank 1. 2. Burt 1. 2. Bruno 1. 2. Dick 2. 1.

Table of Votes Against Wine -> A B

Group Ranking -> 1 2 Votes Against -> 10 14

( 8 is the best possible, 16 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 = 0.2500

The probability that random chance could be responsible for this correlation is rather large, 0.1573. 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.

The correlation I measures the degree to which the identification of each judge is correlated with the truth. Here a 1.0 means that the judge identified the wines perfectly, and a 0 means that he identified none of them.

Correlation Between the Ranks of Each Person With the Average Ranking of Others

Name of Person Correlation R Correlation I John 1.0000 0.0000 Bob 1.0000 1.0000 Orley 1.0000 1.0000 Burt 1.0000 1.0000 Frank 1.0000 1.0000 Bruno 1.0000 1.0000 Ed -1.0000 1.0000 Dick -1.0000 1.0000

Next, we show the correlation among the wine identifications of the judges, which also ranges between 1.0 and 0.0:

C = 0.5625

The probability that random chance could be responsible for this correlation is quite small: < 5 %. Most people would say that unless this probability is less than 0.1, the judges' identifications are not highly related.

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 A is Richebourg Leroy 1959 --------------------------------------------------- 2. ........ 2nd place Wine B is Musigny Leroy 1959 We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 2.0000. The probability that this could happen by chance is 0.1573 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 1.00 for significance at the 0.05 level and must exceed 1.00 for significance at the 0.1 level John Bob Orley John 1.000 1.000 1.000 Bob 1.000 1.000 1.000 Orley 1.000 1.000 1.000 Ed -1.000 -1.000 -1.000 Frank 1.000 1.000 1.000 Burt 1.000 1.000 1.000 Bruno 1.000 1.000 1.000 Dick -1.000 -1.000 -1.000 Ed Frank Burt John -1.000 1.000 1.000 Bob -1.000 1.000 1.000 Orley -1.000 1.000 1.000 Ed 1.000 -1.000 -1.000 Frank -1.000 1.000 1.000 Burt -1.000 1.000 1.000 Bruno -1.000 1.000 1.000 Dick 1.000 -1.000 -1.000 Bruno Dick John 1.000 -1.000 Bob 1.000 -1.000 Orley 1.000 -1.000 Ed -1.000 1.000 Frank 1.000 -1.000 Burt 1.000 -1.000 Bruno 1.000 -1.000 Dick -1.000 1.000 Pairwise correlations in descending order 1.000 John and Bob Significantly positive 1.000 John and Orley Significantly positive 1.000 Orley and Bruno Significantly positive 1.000 John and Frank Significantly positive 1.000 John and Burt Significantly positive 1.000 John and Bruno Significantly positive 1.000 Bob and Frank Significantly positive 1.000 Bob and Orley Significantly positive 1.000 Frank and Burt Significantly positive 1.000 Orley and Burt Significantly positive 1.000 Bob and Burt Significantly positive 1.000 Bob and Bruno Significantly positive 1.000 Burt and Bruno Significantly positive 1.000 Frank and Bruno Significantly positive 1.000 Orley and Frank Significantly positive 1.000 Ed and Dick Significantly positive -1.000 Orley and Ed Significantly negative -1.000 Orley and Dick Significantly negative -1.000 John and Dick Significantly negative -1.000 Ed and Burt Significantly negative -1.000 Ed and Bruno Significantly negative -1.000 Bob and Dick Significantly negative -1.000 Bob and Ed Significantly negative -1.000 John and Ed Significantly negative -1.000 Frank and Dick Significantly negative -1.000 Ed and Frank Significantly negative -1.000 Burt and Dick Significantly negative -1.000 Bruno and Dick Significantly negative

COMMENT: The 1959s belie the conventional wisdom that Burgundies do not age well. The 1990s by contrast, were very drinkable now. Some of us were surprised that the 1959s were as good as they were over 40 years of the original vintage.

Return to previous page