CellarTracker and the problem of groupthink.

Sorry if this has already been discussed here before, but my admittedly lazy search yielded no hits in the first two pages.

CellarTracker is a fantastic tool with many benefits, but I have noticed that, similar to Godwin’s law on Hitler and the net, it seems like the greater the number of wine reviews on CellarTracker, the closer the aggregate score will approach 90.

I have learned to downplay scores over the years, and pay more attention to descriptions of those whose palates that align with mine, but from an academic standpoint, I find it fascinating that we usually end up close to 90. Once again, I’m not knocking CellarTracker, but if 1945 DRC had a production run of 600,000 instead of 600 (insert Rudy joke here), I bet that it would have an aggregate score of 90.

I’ve noticed the scores from a slightly different take. I once went through a listing of all of my different wines and found that only a small handful had an aggregate rating of less than 90 points. Part of it is confirmation bias and part of it is people don’t want to admit they have a bunch of sub-90 point wines in their cellar.

Here’s a couple with about 300 notes…

There’s definitely an anchoring effect around 90. I don’t look at scores though, just notes, so this isn’t really a problem for me. Especially if I know who wrote the note and what their palate is like, I can learn far more from the note than I ever could a score.

I’ve also noticed something similar. However, what I have noticed is that the points scores tend to be quite sticky - meaning that people tend to score the wines in or around where they currently are. I also don’t use scores but rather the descriptions but I always find it amusing to see what scores are given.

Works for me as I always drink 90pt wines.

Hahaha. A 89 haut Brion. Maybe there are outliers to the trend?

Sorry if this has already been discussed here before, but my admittedly lazy search yielded no hits in the first two pages.

Just look at any thread involving CT. The usefulness of the scores have been discussed many times.

Before I posted my Sojourn tasting note thread, I decided to check my past vintage notes on the wines I was posting.

My averages over several vintages (eyeballed).

Home Ranch 92
Gaps 92
Sang 92
S.C. 90

You didn’t look hard enough.

And about that 90? We all revert to the mean, just as in the long run, we are all dead.
Something to keep in mind when somebody is bragging about his new & shiny…

I’m not sure how I average out on CT but I have posted my fair share of below 90 scores on wines that average 92-96.

I also do own some wines that average between 86-89.9. But not many.

I have also bought wines with CT scores of 92 and above and pro scores higher still and not cared for that specific wine while still enjoying lower scores wines by that same producer. Such as the Becklyn estate vs the Moulds. I will keep buying the estate and free up the Moulds for others to enjoy.

the other issue, same as with Zagat and Yelp, is that “expert” or “informed” opinions carry the same weight as novices. If one has never tasted aged Bordeaux or grand cru Burgs, does the tasting note score mean as much?

Excellent point. But what does even expert or informed mean? One reason I pay more attention to wines rated here than on CellarTracker is that I know who is posting. Once in a while I will see a familiar name on CellarTracker, but not always. For example, Alan, I will pay a lot more attention to a post by you than by someone who is new to Burgundy or someone who always rates bigger wines more highly.

Agree Anthony that people are afraid to put scores out there that are too different from the average score. I don’t have that problem and have scored wines 8-10 points different than what the average is and have had a couple of people over the years comment on it and ask why so different than everyone else. I think that is part of the great things of an honest professional wine reviewer that they are not afraid to score wines how they seem to them (insert RP score here for Caymus). It takes guts sometimes to score the way you see it

The lack or paucity of ratings has often camouflaged the best WineHunter finds.

Summation of all scores in CT:

100 0.1% Notes (7,391) / Avg. 100.0 pts. in 7,391 ratings
99 0.1% Notes (6,182) / Avg. 99.0 pts. in 6,182 ratings
98 0.3% Notes (15,549) / Avg. 98.0 pts. in 15,549 ratings
97 0.4% Notes (25,445) / Avg. 97.0 pts. in 25,445 ratings
96 0.9% Notes (52,101) / Avg. 96.0 pts. in 52,101 ratings
95 2.0% Notes (118,048) / Avg. 95.0 pts. in 118,048 ratings
94 3.3% Notes (193,170) / Avg. 94.0 pts. in 193,170 ratings
93 5.7% Notes (332,872) / Avg. 93.0 pts. in 332,872 ratings
92 9.1% Notes (529,259) / Avg. 92.0 pts. in 529,259 ratings
91 9.6% Notes (562,748) / Avg. 91.0 pts. in 562,748 ratings
90 13.4% Notes (781,247) / Avg. 90.0 pts. in 781,247 ratings
89 10.0% Notes (586,298) / Avg. 89.0 pts. in 586,298 ratings
88 9.8% Notes (569,960) / Avg. 88.0 pts. in 569,960 ratings
87 6.0% Notes (353,340) / Avg. 87.0 pts. in 353,340 ratings
86 4.2% Notes (244,697) / Avg. 86.0 pts. in 244,697 ratings
85 3.9% Notes (227,708) / Avg. 85.0 pts. in 227,708 ratings
84 2.4% Notes (141,566) / Avg. 84.0 pts. in 141,566 ratings
83 1.3% Notes (73,524) / Avg. 83.0 pts. in 73,524 ratings
82 1.1% Notes (65,748) / Avg. 82.0 pts. in 65,748 ratings
81 0.5% Notes (28,483) / Avg. 81.0 pts. in 28,483 ratings
80 1.1% Notes (64,796) / Avg. 80.0 pts. in 64,796 ratings
79 0.5% Notes (27,025) / Avg. 79.0 pts. in 27,025 ratings
78 0.3% Notes (20,320) / Avg. 78.0 pts. in 20,320 ratings
77 0.2% Notes (9,976) / Avg. 77.0 pts. in 9,976 ratings
76 0.1% Notes (8,534) / Avg. 76.0 pts. in 8,534 ratings
75 0.4% Notes (23,291) / Avg. 75.0 pts. in 23,291 ratings
74 0.1% Notes (6,996) / Avg. 74.0 pts. in 6,996 ratings
73 0.1% Notes (3,472) / Avg. 73.0 pts. in 3,472 ratings
72 0.1% Notes (3,863) / Avg. 72.0 pts. in 3,863 ratings
71 0.0% Notes (2,140) / Avg. 71.0 pts. in 2,140 ratings
70 0.2% Notes (11,024) / Avg. 70.0 pts. in 11,024 ratings
69 0.0% Notes (1,721) / Avg. 69.0 pts. in 1,721 ratings
68 0.0% Notes (1,369) / Avg. 68.0 pts. in 1,369 ratings
67 0.0% Notes (677) / Avg. 67.0 pts. in 677 ratings
66 0.0% Notes (565) / Avg. 66.0 pts. in 565 ratings
65 0.0% Notes (2,581) / Avg. 65.0 pts. in 2,581 ratings
64 0.0% Notes (485) / Avg. 64.0 pts. in 485 ratings
63 0.0% Notes (387) / Avg. 63.0 pts. in 387 ratings
62 0.0% Notes (490) / Avg. 62.0 pts. in 490 ratings
61 0.0% Notes (451) / Avg. 61.0 pts. in 451 ratings
60 0.1% Notes (3,071) / Avg. 60.0 pts. in 3,071 ratings
59 0.0% Notes (120) / Avg. 59.0 pts. in 120 ratings
58 0.0% Notes (120) / Avg. 58.0 pts. in 120 ratings
57 0.0% Notes (86) / Avg. 57.0 pts. in 86 ratings
56 0.0% Notes (73) / Avg. 56.0 pts. in 73 ratings
55 0.0% Notes (475) / Avg. 55.0 pts. in 475 ratings
54 0.0% Notes (85) / Avg. 54.0 pts. in 85 ratings
53 0.0% Notes (88) / Avg. 53.0 pts. in 88 ratings
52 0.0% Notes (120) / Avg. 52.0 pts. in 120 ratings
51 0.0% Notes (177) / Avg. 51.0 pts. in 177 ratings
50 0.0% Notes (2,254) / Avg. 50.0 pts. in 2,254 ratings
(unrated) 12.5% Notes (728,432) / Avg. 0.0 pts. in 13 ratings


So just slightly more 90 point scores than there are no scores at all.
It also means 86.6% of ALL notes are NOT 90 pts

Surprising to see that 59.3% af all CT notes with scores, are in the 88-92pts range. (3 mill.+)

First, at least to me, individual tasters you are familiar with, that are consistent in your view are more important than group scores. The score does take into consideration who is rating the wines and does not speak as well as the comments supporting the score, from tasters you have come to trust.

Second, generally speaking, most people write notes on wines they own and or already like, so scores are prone to being higher. Very few people post notes on wines they do not know, or are not familiar with.

Third, like all groups, there is a common think and many people, due to fear of not being invited to the party will not post scores out of a certain range that has been determined by other tasters, friends and critics.

Fourth, individual tasters you are familiar with, that are consistent in your view are more important than group scores. The score does take into consideration who is rating the wines and does not speak as well as the comments supporting the score, from tasters you have come to trust.

But if something’s expensive old and rare, the score will cluster around 95 even if drinks blind like a 90 .

I wanted to wait until a few more people posted before I weighed in on this… But I am 90 points on this thread…