Report

WINETASTER ON 06/02/08 WITH 7 JUDGES AND 8 WINES BASED ON RANKS, IDENT=Y
Copyright (c) 1995-2008 Richard E. Quandt, V. 1.65

Today's tasting comprises two flights of relatively inexpensive 2005 Bordeaux wines. The same
judges participated in both flights.

FLIGHT 1:
Number of Judges = 7
Number of Wines = 8

Identification of the Wine: The judges' overall ranking:

Wine A is Ch. La Rose Mayat, Puissegon St.Emilion ........ 6th place
Wine B is Ch. Bellisle Mondotte, St. Emilion ........ 1st place
Wine C is Ch. de Lavagnac, St. Emilion ........ 2nd place
Wine D is Ch. Quercy, St. Emilion Grand Cru tied for 4th place
Wine E is Ch. Lauriol, Côtes de Francs ........ 7th place
Wine F is La Sacristie de Vielle Cure, Fronsac ........ 3rd place
Wine G is Ch. La Bourrée, Côtes de Castillon tied for 4th place
Wine H is Ch. St. Sulpice, Entre Deux Mers ........ 8th place

The Judges's Rankings

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

Bob 3. 1. 2. 4. 6. 5. 7. 8.
Mike 3. 6. 7. 5. 8. 1. 2. 4.
Lou 3. 5. 8. 6. 2. 1. 4. 7.
Ted 8. 1. 2. 6. 4. 7. 3. 5.
John 6. 4. 1. 3. 8. 2. 5. 7.
Ed 6. 5. 1. 2. 4. 7. 8. 3.
Dick 8. 1. 3. 5. 6. 4. 2. 7.

Table of Votes Against

Wine -> A B C D E F G H

Group Ranking -> 6 1 2 4 7 3 4 8
Votes Against -> 37 23 24 31 38 27 31 41

( 7 is the best possible, 56 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 = 0.1516

The probability that random chance could be responsible for this correlation
is rather large, 0.3857. 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 Price

John 0.7381 -0.4880
Dick 0.6707 0.3172
Bob 0.4392 0.3904
Ted 0.1796 -0.2684
Mike -0.2530 -0.4392
Ed -0.3234 -0.5123
Lou -0.5030 -0.3172

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 B is Ch. Bellisle Mondotte, St. Emilion
2. ........ 2nd place Wine C is Ch. de Lavagnac, St. Emilion
3. ........ 3rd place Wine F is La Sacristie de Vielle Cure, Fronsac
4. tied for 4th place Wine G is Ch. La Bourrée, Côtes de Castillon
5. tied for 4th place Wine D is Ch. Quercy, St. Emilion Grand Cru
6. ........ 6th place Wine A is Ch. La Rose Mayat, Puissegon St.Emilion
7. ........ 7th place Wine E is Ch. Lauriol, Côtes de Francs
8. ........ 8th place Wine H is Ch. St. Sulpice

We now test whether the ranksums AS A WHOLE provide a significant ordering.
The Friedman Chi-square value is 7.4286. The probability that this could
happen by chance is 0.3857

We now test whether the group ranking of wines is correlated with the
prices of the wines. The rank correlation between them is -0.4909. At the
10% level of significance this would have to exceed the critical value of
0.5240 to be significant.

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

Bob Mike Lou

Bob 1.000 -0.333 -0.167
Mike -0.333 1.000 0.381
Lou -0.167 0.381 1.000
Ted 0.262 -0.548 -0.452
John 0.548 0.143 -0.262
Ed 0.238 -0.714 -0.738
Dick 0.357 -0.048 -0.143

Ted John Ed

Bob 0.262 0.548 0.238
Mike -0.548 0.143 -0.714
Lou -0.452 -0.262 -0.738
Ted 1.000 0.143 0.214
John 0.143 1.000 0.190
Ed 0.214 0.190 1.000
Dick 0.762 0.548 -0.167

Dick

Bob 0.357
Mike -0.048
Lou -0.143
Ted 0.762
John 0.548
Ed -0.167
Dick 1.000

Pairwise correlations in descending order

0.762 Ted and Dick Significantly positive
0.548 Bob and John Not significant
0.548 John and Dick Not significant
0.381 Mike and Lou Not significant
0.357 Bob and Dick Not significant
0.262 Bob and Ted Not significant
0.238 Bob and Ed Not significant
0.214 Ted and Ed Not significant
0.190 John and Ed Not significant
0.143 Ted and John Not significant
0.143 Mike and John Not significant
-0.048 Mike and Dick Not significant
-0.143 Lou and Dick Not significant
-0.167 Bob and Lou Not significant
-0.167 Ed and Dick Not significant
-0.262 Lou and John Not significant
-0.333 Bob and Mike Not significant
-0.452 Lou and Ted Not significant
-0.548 Mike and Ted Not significant
-0.714 Mike and Ed Significantly negative
-0.738 Lou and Ed Significantly negative

COMMENT:

This is the first of two flights of tastings from 2005 Bordeaux wines.
They all tasted remarkably good and drinkable. They all had very soft
tannins and the top four wines were really quite outstanding. At least
two of the top four wines in this tasting had a high Merlot content
which may account for their relatively high rankings. The ripeness of the
fruit in these wines is also quite evident. But overall, the wines were quite
comparable, which is revealed by the facts that the overall correlation in the
group was quite low and that none of the wines was significantly preferred over
the others.

Return to previous page

WINETASTER ON 06/02/08 WITH 7 JUDGES AND 8 WINES BASED ON RANKS, IDENT=N
Copyright (c) 1995-2008 Richard E. Quandt, V. 1.65

FLIGHT 2:
Number of Judges = 7
Number of Wines = 8

Identification of the Wine: The judges' overall ranking:

Wine A is Ch. Recougne, Terra Incognita ........ 8th place
Wine B is Ch. Pey La Tour, Bordeaux Supérieur ........ 7th place
Wine C is Grands Chenes, St. Christoly Médoc ........ 6th place
Wine D is Ch. Marjosse, Grand vin Bordeaux tied for 3rd place
Wine E is Ch. Puy Bardens, Côtes de Bordeaux ........ 1st place
Wine F is Ch. le Pavillon de Boyrein, Graves tied for 3rd place
Wine G is Ch. Barreyre, Bordeaux Supérieur ........ 2nd place
Wine H is Ch. La Grange Clinet, Côtes de Bordeaux ........ 5th place

The Judges's Rankings

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

Bob 3. 4. 5. 6. 8. 7. 2. 1.
Mike 5. 6. 8. 3. 4. 2. 1. 7.
Lou 4. 5. 6. 8. 7. 3. 1. 2.
Ted 7. 8. 1. 3. 2. 6. 5. 4.
John 8. 7. 6. 2. 1. 3. 5. 4.
Ed 7. 6. 1. 4. 2. 3. 8. 5.
Dick 4. 1. 8. 3. 2. 5. 6. 7.

Table of Votes Against

Wine -> A B C D E F G H

Group Ranking -> 8 7 6 3 1 3 2 5
Votes Against -> 38 37 35 29 26 29 28 30

( 7 is the best possible, 56 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 = 0.0690

The probability that random chance could be responsible for this correlation
is rather large, 0.8477. 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 Correlation Price

John 0.6747 -0.2036
Mike 0.0488 0.2036
Ted -0.0723 -0.2275
Ed -0.2651 -0.1916
Lou -0.3095 0.8503
Dick -0.4431 -0.6826
Bob -0.6429 0.3832

1. ........ 1st place Wine E is Ch. Puy Bardens, Côtes de Bordeaux
2. ........ 2nd place Wine G is Ch. Barreyre, Bordeaux Supérieur
3. tied for 3rd place Wine F is Ch. le Pavillon de Boyrein, Graves
4. tied for 3rd place Wine D is Ch. Marjosse, Grand vin Bordeaux
5. ........ 5th place Wine H is Ch. La Grange Clinet, Côtes de Bordeaux
6. ........ 6th place Wine C is Grands Chènes, St. Christoly Médoc
7. ........ 7th place Wine B is Ch. Pey La Tour, Bordeaux Supérieur
8. ........ 8th place Wine A is Ch. Recougne, Terra Incognita

We now test whether the ranksums AS A WHOLE provide a significant ordering.
The Friedman Chi-square value is 3.3810. The probability that this could
happen by chance is 0.8477

We now test whether the group ranking of wines is correlated with the
prices of the wines. The rank correlation between them is 0.0060. At the
10% level of significance this would have to exceed the critical value of
0.5240 to be significant.

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

Bob Mike Lou

Bob 1.000 -0.238 0.690
Mike -0.238 1.000 0.214
Lou 0.690 0.214 1.000
Ted -0.333 -0.214 -0.452
John -0.595 0.405 -0.333
Ed -0.714 -0.333 -0.595
Dick -0.429 0.238 -0.476

Ted John Ed

Bob -0.333 -0.595 -0.714
Mike -0.214 0.405 -0.333
Lou -0.452 -0.333 -0.595
Ted 1.000 0.548 0.714
John 0.548 1.000 0.500
Ed 0.714 0.500 1.000
Dick -0.405 0.143 -0.143

Dick

Bob -0.429
Mike 0.238
Lou -0.476
Ted -0.405
John 0.143
Ed -0.143
Dick 1.000

Pairwise correlations in descending order

0.714 Ted and Ed Significantly positive
0.690 Bob and Lou Significantly positive
0.548 Ted and John Not significant
0.500 John and Ed Not significant
0.405 Mike and John Not significant
0.238 Mike and Dick Not significant
0.214 Mike and Lou Not significant
0.143 John and Dick Not significant
-0.143 Ed and Dick Not significant
-0.214 Mike and Ted Not significant
-0.238 Bob and Mike Not significant
-0.333 Bob and Ted Not significant
-0.333 Lou and John Not significant
-0.333 Mike and Ed Not significant
-0.405 Ted and Dick Not significant
-0.429 Bob and Dick Not significant
-0.452 Lou and Ted Not significant
-0.476 Lou and Dick Not significant
-0.595 Bob and John Not significant
-0.595 Lou and Ed Not significant
-0.714 Bob and Ed Significantly negative

COMMENT:

This second flight of wines was not quite as impressive as the first
flight. However, one taster thought the exact opposite. This taster
thought that the second flight had more grape variety and that in this
flight there was more cabernet. In any event, the wines were very enjoyable.
We note that the rank sums in this group were even more tightly bunched
than in Flight 1 (which does not necessarily reflect the quality of the
wines, but possibly the increasing randomness in the tasters' choices
resulting from the consumption of a second set of eight bottles).
It is important to note that the 2005 Bordeaux are ready to drink now
and are extremely enjoyable. Because of their prices, these wines are
tremendous values. Buy them now.

Return to previous page