Report

WINETASTER ON 03/04/14 WITH 7 JUDGES AND 9 WINES BASED ON RANKS, IDENT=N
Copyright (c) 1995-2014 Richard E. Quandt, V. 1.65

FLIGHT 1:
Number of Judges = 7
Number of Wines = 9

Identification of the Wine: The judges' overall ranking:

Wine A is Clos de Tart 2003 ........ 6th place
Wine B is Bonnes Mares 2002 (Moine Hudelot) tied for 8th place
Wine C is Bonnes Mares 1999 (Laurent) ........ 4th place
Wine D is Bonnes Mares 2003 (Bouchard) ........ 1st place
Wine E is Clos des Lambreys 2003 ........ 3rd place
Wine F is Clos de Tart 1999 tied for 8th place
Wine G is Clos de Tart 2002 ........ 7th place
Wine H is Clos des Lambreys 2002 ........ 5th place
Wine I is Clos des Lambreys 1999 ........ 2nd place

The Judges's Rankings

Judge Wine -> A B C D E F G H I

Alexa 8. 5. 6. 3. 1. 4. 9. 7. 2.
Burt 5. 9. 3. 2. 6. 7. 8. 1. 4.
Bob 7. 2. 3. 5. 4. 9. 1. 8. 6.
Mike 3. 8. 9. 1. 4. 5. 7. 6. 2.
Frank 3. 7. 5. 4. 8. 9. 1. 6. 2.
Ed 5. 8. 2. 1. 3. 7. 9. 4. 6.
Dick 6. 9. 1. 3. 2. 7. 8. 4. 5.

Table of Votes Against

Wine -> A B C D E F G H I

Group Ranking -> 6 8 4 1 3 8 7 5 2
Votes Against -> 37 48 29 19 28 48 43 36 27

( 7 is the best possible, 63 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 = 0.2762

The probability that random chance could be responsible for this correlation
is quite small, 0.0507. 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

Ed 0.7448
Dick 0.7167
Burt 0.5272
Alexa 0.3598
Mike 0.3333
Frank -0.1000
Bob -0.2167

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 D is Bonnes Mares 2003
---------------------------------------------------
2. ........ 2nd place Wine I is Clos des Lambreys 1999
3. ........ 3rd place Wine E is Clos des Lambreys 2003
4. ........ 4th place Wine C is Bonnes Mares 1999
5. ........ 5th place Wine H is Clos des Lambreys 2002
6. ........ 6th place Wine A is Clos de Tart 2003
7. ........ 7th place Wine G is Clos de Tart 2002
---------------------------------------------------
8. tied for 8th place Wine B is Bonnes Mares 2002
9. tied for 8th place Wine F is Clos de Tart 1999

We now test whether the ranksums AS A WHOLE provide a significant ordering.
The Friedman Chi-square value is 15.4667. The probability that this could
happen by chance is 0.0507

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 0.70 for significance at the 0.05
level and must exceed 0.60 for significance at the 0.1 level

Alexa Burt Bob

Alexa 1.000 0.083 -0.150
Burt 0.083 1.000 -0.433
Bob -0.150 -0.433 1.000
Mike 0.483 0.333 -0.467
Frank -0.417 0.150 0.317
Ed 0.367 0.783 -0.183
Dick 0.383 0.733 -0.100

Mike Frank Ed

Alexa 0.483 -0.417 0.367
Burt 0.333 0.150 0.783
Bob -0.467 0.317 -0.183
Mike 1.000 0.217 0.317
Frank 0.217 1.000 -0.133
Ed 0.317 -0.133 1.000
Dick 0.167 -0.100 0.917

Dick

Alexa 0.383
Burt 0.733
Bob -0.100
Mike 0.167
Frank -0.100
Ed 0.917
Dick 1.000

Pairwise correlations in descending order

0.917 Ed and Dick Significantly positive
0.783 Burt and Ed Significantly positive
0.733 Burt and Dick Significantly positive
0.483 Alexa and Mike Not significant
0.383 Alexa and Dick Not significant
0.367 Alexa and Ed Not significant
0.333 Burt and Mike Not significant
0.317 Mike and Ed Not significant
0.317 Bob and Frank Not significant
0.217 Mike and Frank Not significant
0.167 Mike and Dick Not significant
0.150 Burt and Frank Not significant
0.083 Alexa and Burt Not significant
-0.100 Bob and Dick Not significant
-0.100 Frank and Dick Not significant
-0.133 Frank and Ed Not significant
-0.150 Alexa and Bob Not significant
-0.183 Bob and Ed Not significant
-0.417 Alexa and Frank Not significant
-0.433 Burt and Bob Not significant
-0.467 Bob and Mike Not significant

COMMENT:

We thought it would be interesting to have a tasting of wines from three
adjacent vineyards in Burgundy for three matched years. So, what we have
is Bonnes Mares, Clos de Tart and Clos des Lambreys for 1999, 2002 and 2003.

Clos de Tart is a single owner vineyard, hence it is designated as "Monopole."
Clos des Lambreys is almost single owner, but just misses because a small portion
is owned by Domaine Taupenot-Merme. Bonnes Mares had many owners. Clos de Tart
and Clos des Lambreys are in the commune of Morey Saint Denis; Bonnes Mares
is situated mostly in Chambolle Musigny, with a small part in Morey Saint Denis.
It would have been nice if all three bottles of Bonnes Mares had come from a
single owner, but that was beyond the host's powers. The actual Bonnes Mares were
Laurent (1999), Moine Hudelot (2002) and Bouchard (2003).

We anticipated that all wines would be good, and that was indeed the case. One
person remarked that, taken alone, each of these wines would be quite fantastic.In
spite of that there was a surprising amount of agreement in the group, with the
probability that the degree of agreement could have arisen by chance only a hair over 5%.
An interesting point to note is that the disagreement among the tasters was worst for the
2002 Clos de Tart, which was ranked first by two tasters and worst or next to worst by
four. Also interesting is the fact that the very best wine is by a negociant (Bouchard).
We were surprised by how badly Clos de Tart did agaist the other wines and performing
the significance test that it is worse than the other six wines yields a test
statistic of 1.29, marginally significant at the 0.05 level. Testing the hypothesis that
2003 is significantly better than the aggregate of the other wines (being mindful of
the fact that these two tests cannot be performed validly on the same set of data) yields
a test statistic of 1.37, which is highly significant. The wines aged beautifully and we
surmised that they can be aged for 25 years without bad results. Given the stylistic and
differences, it is hard to draw distinctions among them.

Return to previous page