It seems to me there are a lot of arguments about how many points a wine gets. Parker gives it a 95, but WS gave it an 88, and so forth. The problem with the point system as applied to wines is that the scores are arbitrary. We have taken a technique used to grade school tests and apply it to wines. But there is a huge difference. For a school test, there are 100 questions. If you get 95 right, you get a test score of 95. Nothing arbitrary about that. The score is reprocible (i.e., no matter who scores the test, the score is the same). In many competitions, a maximum score of 100 (or 10, or 5 or whatever) is awarded for a perfect performance. Points are then subtracted for flaws. The number of points taken away for any given flaw are decided in advance, and the things that are considered a flaw in the performance are accepted by all those who participate. Although the scores will not be quite as clean as a school test, it is still relatively consistent and reproducible.

Wine scores, on the other hand have none of these features. My guess is that if you took 10 wine critics and asked them to score a wine blindly 10 times, you might get similar tasting notes, but you will get different scores. How do you decide between a 93 and a 96? It may depend on what else you have drank that day, your general health and emotional well being, and so on (not to mention bottle variation, differences in testing conditions, length ot time bottle was open, which order it was tested in a multi-wine line-up).

The most logical scale is simple: Either you think the wine was near perfect, or you thought it was excellent, or you thought it was pretty good, average, or poor. Even this is subjective, of course, but would eliminate arguments about a few points. Trying to split hairs with an axe is silly, and that's what we are doing when we break down nearly perfect wines to a 100 point wine versus a 98.

It seems to me, the only reason to have numerical scores is to sell wine. It's good marketing. And I guess it makes for fun arguments, similar to who is the greatest baseball player of all time?

Just my opinion.

Yeah....good luck scrapping the point system.

A numerical wine score is a measure of how much the reviewer likes the wine. In addition to the prose, I enjoy the quantification.

quote:

Originally posted by Board-O:
A numerical wine score is a measure of how much the reviewer likes the wine. In addition to the prose, I enjoy the quantification.

And there in lies another major problem with the 100 point system. It is meant to measure the inherent qualities of the wine, not how much it is liked or disliked, which is another reason the 100 point system is based on a bed of quicksand.

Another problem is a taster like Tanzer who really doesn't believe in 97-100 pt wines. Or at least I've never heard of a review from him in that range. If his idea of a near perfect wine is 95-96 pts., this is different than when Parker says 95-96. So, even though some subscribe to the 100 pt. system, some don't fully believe in it.

I don't mind the score so much. Sometimes the notes can as mysterious as the score. I am not a very skilled taster, I can pick up a predominat flavor and maybe 2-3 more, but I have never noticed the laundry list of very distinct flavors in a wine that are sometimes listed.

The system you are proposing is not that different than the five star system that Decanter (a U.K. publication I beleive) uses.

One thing you may be overlooking, it is a matter of taste. Critics taste vary so you will never have the reproducibility you are hoping for.

Indeed wine is a matter of taste, and all ratings must therefore be subjective regarless of whether one uses stars, numbers, or an excellent-good-fair-poor system. What bothers me specifically about the point system, however, is that no one who uses them can define the criteria for their scores. Even Mr. Parker, who, for instance, will give anywhere from 1-15 points for bouquet can not say what determines a 9 point bouquet from an 8 point bouquet from a 10 point bouquet. Herein lies the problem. We have applied what is essentially a scientific/mathematical evaluation system to a craft which is essentially an art. Nobody rates paintings, movies or music from 0-100, becuase everyone would say it is silly. But for some reason, no one questions it with wine, at least in this country.

Yes, Spo, Decanter does indeed use 5 stars, and they define exactly what a star means. Most Europeans laugh at our fixation on numbers. The Italians use Gambero Rosso's Tre Bicchieri system (one glass to three glasses). Oz Clarke and Hugh Johnson avoid numerical scores.

Wine Spectator says that 95-100, in their scale, means a classic, and 90-94 means outstanding. They should just stick to the descriptive terms unless they can define what differentiates a 93 from a 94. Their scale only defines what differentiates a 95 (a classic) from a 94( outstanding). By extension, then a single point difference between a 94 and 95 is indicative of a huge difference in quality, whereas the same one point difference between a 93 and 94 indicates minimal difference in quality. Numerical scales should not be used unless each point represents the same incremental difference in quality.

Wine Joe, you suggest that we replace the point wine rating scale with a general category scale. You suggest, "Either you think the wine was near perfect, or you thought it was excellent, or you thought it was pretty good, average, or poor."

This works and WS and Parker basically have this category gradation connected to their spreads of points. For example, WS calls all wines that score 95-100 "classic" while calling wines that score 90-94 "outstanding." A wine that scores 80-84 according to WS is "Good." Again, this works; however, general category rating systems can't differentiate between similar yet individual wines.

Let's consider an illustration of my above point. Say you taste 100 wines and call 17 of them, "Excellent." Now say I ask you if you equally liked all 17 of your "excellent" wines. You would almost certainly respond with something like, "no I like these few the most, then these and finally these were excellent but not as excellent as the others." You see? A point scale may cause debate, but it is needed to distinguish between individual but like wines. Furthermore, the debate that a point scale causes isn't always negative but instead often forces us to think more discriminatingly about why we like one wine just a little more than another. This is a good thing that only adds to the enjoyment of wine.

Still, in my opinion what helps us transcend beyond the general categories and point scales is the critic's verbal description of the wine. Granted, these descriptions can sound similar from one category to the next, but the more the description actually conveys individual difference moving up the scale the better the critic that gave it.

Thanks
Bavinck

Bavnick,

If all of the excellent wines in your example can be subdivided with meaningful differences in quality, then it implies you need another category. My point is that if you can not put in words the exact criteria for different rankings, then the scale fails to meet the critieria that most academic disciplines use when deciding whether to adopt a rating scale.

For a fun description of rating scales, look at the rating scale used on www.VinoCritic.com . You may agree or disagree with how they scored the wines, but at least the criteria for arriving at how many stars each wine gets are spelled out clearly (and humorously).

I think I saw someone using the grading system (A+ to F) on this site and Cellartracker. I like that idea, it can be tied to both percentage score as well as having some element of personal bias. The fact that a number is not clearly assigned leaves room for interpretation and puts more emphasis on the qualitative aspect of the grade. If the point system was ever scrapped, this would be my first pick of the replacement system. The 3-glass system used by Gambero Rosso is analogous to this, so I would have it as a close second.

Wine Joe, I think you make an excellent point when you say, "if all of the excellent wines in your example can be subdivided with meaningful differences in quality, then it implies you need another category." This is how I see the 100 point scale, just a large list of categories and not a scale of measurable equal increments. Thus, my criticism of your earlier comments is really just that we need more than the few categories that you suggest. We need more so that we can better distinguish between like but different wines. Perhaps what we need is sub-categories within your categories. Intestinally, if we have five categories each with five sub-categories, then we have a 25 point "category" scale which isn't much different that the 50-100 point scales used by WS and Parker.

Frankly Joe, I don't think WS or Parker in anyway claim that their 100 (really just 50) point scales are exacting equal increment scales. Certainly they just intend their scales as categories of liking one wine a little more than another? Perhaps you have misunderstood the 100 scales in that you see it as other than it is intended or even can logically be. You are right that we can't measure wine in equal increments of quality. That as you argue is impossible, but I really don't think that is what WS or Parker claim to do. I for one don't see or use the 100 point scale in that way but only as an expanded version of your proposed category scale.

PS--Thanks for the interesting link to the VinoCritic. I am always looking for sites like that and hadn't seen this one. This one has potential, apparently starting with a critic's judgment on a wine and then inviting user votes/comments as well. Readers can even suggest wines for review. I hope this one takes off!

Thanks
Bavinck

Yeah, i like the vinocritic wine of the week. It's sort of like the PBS version of a wine review, where you get expanded coverage of one topic (i.e. a whole page about one winery, history of the vineyard, etc). Most wine reviews now more analogous to a CNN quikie version of the news. Not that it's bad, but sometimes it's nice to get the in depth look. Too bad they don't keep the wines of the week from last year posted though.

Wine Joe has made some good points, none of which I can disagree with but one I can expand upon.

Another ridiculous aspect of the numerical scorecrops up when you compare say a very drinkable Jacobs Creek (or even a slightly higher quality Barossa Shiraz) that is designed for immediate consumption and will never get any better, against a First Growth Bordeaux that is 20 years off its best and in a thumping great big hole. Yet on many occasions, they will both of the same score.

The you seriously made to tell me that the Jacobs Creek, as enjoyable as it may been now, is the same quality as that First Growth Bordeaux - because that is exactly what those scores (when taken in isolation, which happens fairly frequently) are indicating.

Many people do not like the system, but right now it is better than no system at all. I agree with Torb. Lets say you take a $20 Shiraz that gets a 90 point rating. Then you take a 2003 Bordeaux that is $90 that gets the same rating. You are supposed to assume that both these wines are of the same quality. If consumed now, they would both be the same quality. But no mentions of it going up 5 points in 10 years or so. I think there needs to be more classifications. Such as potential and other things.

That's why many rate young wines with a range of points.

The scores are not "arbitrary". They are "subjective".
If you take them for that, they provide some useful info.
It's like grading an essay exam, not a math exam.

When I was in school, essays were graded A,B,C,D, and F, but were not given numbers. Perhaps other schools were different. Maybe the English teachers back then were having the same discussion we are now!

I respectfully disagree. It's not the intrinsic nature of the scores that matter, but what they say to you as a buyer and consumer.

Points scores are like any other type of reference, in that one must educate oneself on the various sources of the scores, over a period of time, to make them useful. I know that (with some exceptions), my tastes are similar to both Steiman and Parker when it comes to my favourite dry red wines, the Aussies. I've been drinking them for 10 years, and reading reviews and scores from Parker and Steiman for as long, so when one of them (ideally both) gives a high score to a wine, and the flavour descriptors match what I like (big, chewy and fruity, generally), then I want it and I almost always like it. So their point scores are useful, to me at least.

On the other hand, I ignore scores for styles of wines that are not really to my taste, though I do use those scores to buy the odd bottle, to continue to try new things, if nothing else. On what else can I base my choices, when I'm buying wine with which I am unfamiliar? Two wines of the same price, similar TNs, but one scores 2 points more than the other - no points for guessing which one I'm buying.

I give point scores to wines for which I post TNs here, as do many others, and I find those useful too, especially when it's a score given by someone whose taste matches my own. If Seaquam likes a big Aussie red, or one of my European portloving buddies likes a port, I'm pretty sure that I'm going to like it too, because we've been here a while and I've read many of their TNs on wines that I've also tasted, so I've got something by which to calibrate their scores.

The bottom line is that the point system adds context to one's wine buying/drinking choices. How can that be a bad thing?

quote:

It seems to me, the only reason to have numerical scores is to sell wine. It's good marketing. And I guess it makes for fun arguments, similar to who is the greatest baseball player of all time?

I agree with you. Now... do you have a solution/alternative in mind? Or are you saying that since subjective, personal judgements are at play, reviews/scoring should be ommitted altogether?

I think I would like to see what some are already doing become the standard: assign and show smaller scores to each element in a wine, add them up, and report the total score along with the component scores. This might look something like:

Color: 19/20
Bouquet: 23/25
Palate: 24/30
Balance: 22/25
--------------
Total: 88/100

Now if 2 reviewers diverge, we can drill down and see, perhaps, the one or two areas where they differed in scoring.

Eswrite, most tasting scores are build up like that.
I think the taste/overall impressions get a higher part of the total and appearance gets a lot less and nose gets someething in between.

Anyway, if you don't publish the details of your scoring card, you still only get the overall picture.

Imho, a score is just an easy way to express appreciation.
And it's just as easy a way to communicate.

If you're not into experimenting to much, you can rely on a 88+ wine. And a dull, average wine will probably never score over 90...

That's it. The rest of the numbers is shere boredom.

quote:

Originally posted by Rik:
Eswrite, most tasting scores are build up like that.

Yes, but I was suggesting that all reviewers use the same standard breakdown, which at the moment, I don't think is the case. Moreover, I was suggesting the component scores should also appear. This wouldn't be too cumbersome. For my example, something like: 19+23+24+22=88, with the order of the components telling us what they are (i.e., Color always first, etc.). Again, I have seen this done in some places, but not in all, and there seems to be a lack of a standard list of components and assigned maximum scores.

Notwithstanding the many valid points about the shortcomings inherent in a points system, there is nothing whatsoever wrong with the critics offering a points system as a shorthand for their subjective evaluation of wine quality.

The problem is that some portion of the readers use the points system the wrong way -- take it as gospel, use it to show off their collection, close their minds to wines that didn't get high scores, take it personally if some wine they like gets a low score, etc. -- just as some people use Zagat restaurant ratings, golf course ratings, movie reviews, NFL draft player ratings, etc. in the wrong ways.

Most people don't have time to read every review and couldn't differentiate between what the critic is trying to say even if they did. I think the numerical score provides a useful, though of course not dispositive, shorthand to help people with limited time and energy identify wines to try which are more likely to be good wines. There is nothing wrong with that.

I also think the scoring system becomes more useful once you learn what critics tend to like and where your tastes converge and diverge from them. Then you can really put the data point that is the score in its proper perspective for assisting your buying decisions.

The point system tells the relative value a particular reviewer places on a bottle of wine. As long as I understand that, I don't get into trouble. I've learned, for example, that a Robert Parker 93 point Spanish wine, is probably simple but diverting garnacha that I would give a B to. It's very different from a James Suckling 93 point Brunello, which would blow my socks off if I could afford it.

Good point Spenser.

Known reviewer + good written description + point rating = useful information. That's all. It isn't written in stone, but the combination of the 3 has worked for me nicely. Remove any one of those components, and you lose much more than a third of the value of the commentary.

PH

Any scoring system worth its weight has to be reproducible. In other words, if the same critic is given the same bottle to taste twice, the score should be identical (bottle variation will not be a factor since the wame bottle is used for both tastes). It should not matter how many wines are tasted that day, or what order they are sampled. If you give it a 92 on the first test, then a 93 on the next test, it means your system of tasting and scoring can not differentiate a one point difference, and therefore wines should not be scored with one point increments. My guess is that if you thought a wine was a "4-star" versus a "3-star", your score would be much more consistent and reproducible than 1-100 scales, although still not perfect.

Here is my challenge to all of you who stand behind the scoring system:


I would challenge any wine critic, professional or otherwise to take the following challenge. Let me line up 30 high quality California Cabernets, all scoring well on the Wine Spectator scale (or Parker's scale). I will give them to the taster blindly and in random order. 1 hour later, I will mix the order and have them all re-tasted and re-scored. If the critic gets the same exact score on 75% of the wines or better, I will buy him/her a bottle of all five first growth Bordeaux's from Vintage 2000. If you fail to nail at least 60% of them, you buy me the bottles. Between 60 and 75 %, we'll call it a wash.

If you can't reproduce your own score fromn the same bottle on the same day, I would say that your scoring system is not reproducible, and should be scrapped in favor of the most reproducible system out there.

Any takers on my challenge? I will pay for the 30 bottles myself. If you lose, we split the cost of the bottles.