WINETASTER ON 02/03/14 WITH 8 JUDGES AND 8 WINES BASED ON RANKS, IDENT=Y Copyright (c) 1995-2014 Richard E. Quandt, V. 1.65
A Tasting of World Wide Cabernet Blends
FLIGHT 1: Number of Judges = 8 Number of Wines = 8
Identification of the Wine: The judges' overall ranking:
Wine A is Domaine du Castel 1997 Judean Hills Israel ........ 1st place Wine B is Leeuwin Estate Art Series 2001 West Australia ........ 4th place Wine C is Meinert 2000 Devon Valley Stellenbosch ........ 7th place Wine D is Ch. Musar 1994 Bekaa Valley Lebanon ........ 3rd place Wine E is Trefethen Res. 1998 Napa ........ 2nd place Wine F is Weinert 1995 Mendoza ........ 5th place Wine G is Ch. Lanessan 2003 Haut Médoc ........ 6th place Wine H is Domaine de Trevallon 2003 Les Baux de Provence ........ 8th place
The Judges's Rankings
Judge Wine -> A B C D E F G H Bob 3. 7. 8. 4. 2. 1. 5. 6. Ed 5. 3. 6. 1. 2. 8. 7. 4. Zaki 1. 2. 3. 4. 5. 6. 7. 8. Joe 1. 6. 3. 5. 4. 2. 7. 8. Mike 4. 5. 7. 6. 3. 1. 2. 8. Orley 4. 3. 1. 2. 5. 8. 7. 6. Frank 3. 6. 8. 4. 2. 5. 1. 7. Dick 2. 1. 7. 6. 3. 4. 5. 8.
Table of Votes Against Wine -> A B C D E F G H
Group Ranking -> 1 4 7 3 2 5 6 8 Votes Against -> 23 33 43 32 26 35 41 55
( 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 = 0.2716
The probability that random chance could be responsible for this correlation is quite small, 0.0334. 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 Dick 0.6467 1.0000 Joe 0.5061 1.0000 Bob 0.4286 0.1250 Zaki 0.4286 0.7500 Frank 0.3571 1.0000 Mike 0.2619 1.0000 Ed 0.0952 1.0000 Orley -0.1078 1.0000
Next, we show the correlation among the wine identifications of the judges, which also ranges between 1.0 and 0.0:
C = 0.7232
The probability that random chance could be responsible for this correlation is rather large: > 10%. 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 Domaine du Castel 1997 Israel --------------------------------------------------- 2. ........ 2nd place Wine E is Trefethen Res. 1998 Napa 3. ........ 3rd place Wine D is Ch. Musar 1994 Lebanon 4. ........ 4th place Wine B is Leeuwin Estate 2001 Australia 5. ........ 5th place Wine F is Weinert 1995 Mendoza 6. ........ 6th place Wine G is Ch. Lanessan 2003 Haut Médoc 7. ........ 7th place Wine C is Meinert 2000 Stellenbosch --------------------------------------------------- 8. ........ 8th place Wine H is Domaine de Trevallon 2003 Les Baux We now test whether the ranksums AS A WHOLE provide a significant ordering. The Friedman Chi-square value is 15.2083. The probability that this could happen by chance is 0.0334 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 Ed Zaki Bob 1.000 -0.071 -0.143 Ed -0.071 1.000 0.238 Zaki -0.143 0.238 1.000 Joe 0.476 -0.262 0.595 Mike 0.714 -0.452 -0.095 Orley -0.571 0.524 0.690 Frank 0.595 0.048 -0.095 Dick 0.333 0.095 0.595 Joe Mike Orley Bob 0.476 0.714 -0.571 Ed -0.262 -0.452 0.524 Zaki 0.595 -0.095 0.690 Joe 1.000 0.357 0.143 Mike 0.357 1.000 -0.643 Orley 0.143 -0.643 1.000 Frank 0.048 0.690 -0.405 Dick 0.381 0.548 -0.048 Frank Dick Bob 0.595 0.333 Ed 0.048 0.095 Zaki -0.095 0.595 Joe 0.048 0.381 Mike 0.690 0.548 Orley -0.405 -0.048 Frank 1.000 0.405 Dick 0.405 1.000 Pairwise correlations in descending order 0.714 Bob and Mike Significantly positive 0.690 Zaki and Orley Significantly positive 0.690 Mike and Frank Significantly positive 0.595 Zaki and Joe Not significant 0.595 Zaki and Dick Not significant 0.595 Bob and Frank Not significant 0.548 Mike and Dick Not significant 0.524 Ed and Orley Not significant 0.476 Bob and Joe Not significant 0.405 Frank and Dick Not significant 0.381 Joe and Dick Not significant 0.357 Joe and Mike Not significant 0.333 Bob and Dick Not significant 0.238 Ed and Zaki Not significant 0.143 Joe and Orley Not significant 0.095 Ed and Dick Not significant 0.048 Ed and Frank Not significant 0.048 Joe and Frank Not significant -0.048 Orley and Dick Not significant -0.071 Bob and Ed Not significant -0.095 Zaki and Frank Not significant -0.095 Zaki and Mike Not significant -0.143 Bob and Zaki Not significant -0.262 Ed and Joe Not significant -0.405 Orley and Frank Not significant -0.452 Ed and Mike Not significant -0.571 Bob and Orley Not significant -0.643 Mike and Orley Not significant
COMMENT: This tasting was scheduled to take place on a day on which a significant winter storm hit our town (once again!). Three of us have had the habit of never driving to a tasting but sharing a taxi. After several phone calls two more people wished to join in our motorized expedition. The taxi that showed up happened to be a Toyota Prius, obviously not designed for a driver plus five passengers. There was nothing to be done: e.g., no second taxi could be called since none were available in the ongoing snow storm. So four of us piled in in the normal way, and we stashed the fifth person in the luggage compartment---an inspired solution if there ever was one. The car's traction was good until we cam upon a slight hill, on which forward motion ceased. We immediately piled out and pushed and managed to push the taxi over the crest of the hill, from which point on it was easy sailing. This obviously worked up a healthy appetite and thirst, the latter being quenched immediately with a delicious Chassagne Montrachet. The wines were all excellent across the many regions and many vintages, all drinking very well. There was at least one big surprise, and that was the 1997 Castel which was an Israeli wine from the Judean Hills (Haute-Judee). Here we have an Israeli wine with considerable bottle age that does very well in the tasting. It was also a big surprise that the Trevaillon did so badly in the ratings. There is a possibility that the results were impacted by a flaw, which is that one of the tasters' wine F, the Ch. Musar, was evidently different from at least two of the wines labeled F in other tasters' glasses. One explanation for this phenomenon is that one of the glasses was not entirely free of residual taint. One taster felt, and many agreed, that although these wines came from many parts of the globe they had a considerable similarity. This suggests that a modern cabernet style will appeal broadly, but as a result this creates a demand for specialized variatals that don't fit under the big global grape umbrella. In any event, the degree of agreement among the tasters was very high. The alcohol content ranged from 12.5% (for the Domaine du Castel) to 14.5% (for the Meinert and the Trefethen). In each case the Cabernet Sauvignon was blended (mostly un unstated proportions) with a subset of Merlot, Petit Verdot, Cabernet Franc, Malbec, Cinsault, Carignan and Syrah, with the exception of the Meinert which was 100% Cabernet Sauvignon.
Return to previous page