Current ratebeer.com Ontario top 10

[JerCraigs]

Consistent with what? My ratings are (mostly) consistent compared to my other ratings. I can explain to you what a "Jeremy Score" means, btu that does not help you understand someone elses rating on Ratebeer even if we are ostensibly scoring using the same system. Ratebeer scores are consistent... compared to other Ratebeer scores.

Saying something is an 80% beer doesn't actually *mean* anything without a context. E.g. a comparison to a different beer, or a detailed explanation.

You and I both got 90% in our class - Yay! But that doesn't mean much if you were taking Particle Physics and I was taking Remedial Basket Weaving. Comparing RB scores to Beer Advocate is like celcius and fahrenheit, its just not the same scale.

[cfrancis]

The thing I find funny about ratebeer is is fear of people rating anything above 4. When in reality that's an 80% beer. There are a select few beers that get above 4 and god forbid 4.5 or a 5. It seems everyone just automatically lops off .5 on their rating for "respect".

I don't think people "fear" anything. If you look at the rating scale Aroma x/10, Appearance x/5, Flavour x/10, Palate x/5, Overall x/20.

An average beer would score (for me anyways) A6/A3/F6/P3/O10-12. I am personally stingy on the appearance and palate scores, so 3 is probably what I give out the most. Most decent microbrews would start at 7s for aroma and flavour, and work up from there depending on how impressed I am. I give out very few 10s in these categories. Where other people fall with their rating patterns will vary wildly, but if you look at the scoring system it is not at all surprising that there are very few individual 5.0s. Even less surprising that there are no scores of 5.0 since that would require hundreds of people to agree that the same beer is a 5.0

I don't think you can compare that to sitting down and giving something a single score out of 100 (which is basically what the Overall score is). Or if you do, then the answer to the question ""What would a beer have to do in order to get higher than that?" will vary. For me its often a case of aroma could be less mild, palate could be more full bodied, flavour could have more X/better balance etc.

I have rated 10 beers from Ontario 4+, none higher than 4.2. They were all pretty awesome beers, but all had areas where they could have been just a bit more... something.

More than 3.5 for me is a solid beer. More than 4 is pretty awesome. More than 4.5 is superlatively good. It is all subjective, its all relative.

I hear what you are saying and when you break it down it makes sense but taking a Westy 12 which is the highest rated beer is a weighted average of 4.48. I understand that with that many ratings it's not going to come in a a 5, I just thought it's weighted avg. would be higher.

[JerCraigs]

Somewhat related:
http://www.ratebeer.com/forums/beer-wit ... 158539.htm

[Derek]

Somewhat related:
http://www.ratebeer.com/forums/beer-wit ... 158539.htm

"5
AROMA 10/10 APPEARANCE 5/5 TASTE 10/10 PALATE 5/5 OVERALL 20/20
TallyMan99 (2) - USA - SEP 29, 2001 does not count [click to see why this rating of Olde English 800 does not count]
This shit rules! I can’t feel my legs."

[Belgian]

Consistent with what? My ratings are (mostly) consistent compared to my other ratings....

Well I'm not math guy but I'll have a crack at this.

What I mean is numerically everyone can have a different 'middle' point for what they think of as an average beer. You personally could say that is 2.9 or 3.5, and two people who disagree on this would not be wrong if either way agrees with their OWN scale. You could also say an average 'really good' beer is around 3.9 OR around 4.5, depending how you personally use the scale.

Then compiling different users' RB ratings for an overall score and percentiles becomes inconsistent.

MORE inconsistent still that some people are rating more logarithmic than linear. Maybe they do this to use more of the center range of the scale to show meaningful ranking distinctions, instead of clustering a million beers around the same score. So with some users' arbitrary log scoring AND the more linear scoring of others we are combining different raters' data to show a result that is a strange mixture of both - WITH the (aforementioned) different values already assigned to different points in the scale that is already throwing things off.

So unless everyone uses ratebeer consistently, the percentiles and rankings are inconsistent. MAYBE people are 'trained' by imitation of others but that's about it for consistency - and perhaps it's best to accept rating beer is more fun that science. Otherwise, "believing is seeing."

[JerCraigs]

So unless everyone uses ratebeer consistently, the percentiles and rankings are inconsistent.

Are you saying the users need to be consistent with themselves or with each other? The whole point of a weighted average is that the greater the number of people who rate a beer, the more "accurate" the score is.

[Belgian]

So unless everyone uses ratebeer consistently, the percentiles and rankings are inconsistent.

Are you saying the users need to be consistent with themselves or with each other? The whole point of a weighted average is that the greater the number of people who rate a beer, the more "accurate" the score is.

I see your point. So in other words, if everyone is inconsistent with others, it matters less because a large amount of 'weighted' data will statistically 'average out' - at least for the purposes of ratebeer scores!

For the purposes of comparing ratebeer scores to your own scores, I would still expect it to be off? Again, I think we still do tend to imitate the rating patterns of others so maybe this effect is a kind of 'feedback loop' for some ways we rate and even for some specific beers.

Blind tasters rating a score without discussing it tends to produce more scattered results, and stray more from entrenched popular opinion (unlike, say, the effusive Westy 12 ratings that are mostly up in the mid-fours.)

[cannondale]

The weighting method employed by the folks at ratebeer (specifically a bayesian model) is not intended to, nor does it address the types of inconsistencies that Belgian seems to be concerned about. It is an arbitrary method that normalizes ratings for beers with relatively few votes toward the mean rating of all beers. This is intended to, among other things, confer greater 'reliability' on the ratings for beers with relatively many votes. As the number of votes for a beer increases, its' weighted average moves toward its' straight arithmetic mean rating.

As for Belgians' troubling inconsistencies, they tend to be ironed out by 'standard' arithmetic averaging. This typically becomes more effective as the size of a given dataset increases. In fact, this is clearly the impetus for the averaging of virtually any set of numerical qualitative measures in the realm of aesthetics in the first place!

[Lackey]

5
AROMA 10/10 APPEARANCE 5/5 TASTE 10/10 PALATE 5/5 OVERALL 20/20
copenhagen69 (1) - Winnipeg, CANADA - MAR 10, 2004 does not count
OE is the shiznit, nuttin fukin beats it, Shanner the G rulz u don;t lik it sux my balls

[JerCraigs]

5
AROMA 10/10 APPEARANCE 5/5 TASTE 10/10 PALATE 5/5 OVERALL 20/20
copenhagen69 (1) - Winnipeg, CANADA - MAR 10, 2004 does not count
OE is the shiznit, nuttin fukin beats it, Shanner the G rulz u don;t lik it sux my balls

Which doesn't count for any of the scoring. Most of the real numbnuts can't be bothered to enter enough ratings to be counted. Similarly, apparently the View Askew forums had like a 90% drop in dumbasses just by charging a nominal $1 membership fee.

[SteelbackGuy]

The thing I find funny about ratebeer is is fear of people rating anything above 4. When in reality that's an 80% beer.

How did you come to the conclusion that people fear rating a beer above a 4?
I use ratebeer.com regularly, and I rate beer as I see fit, using their scale and using it according to my tastes. But not once, have I feared rating a beer above 4. If it makes the cut, it makes the cut. My average rating is about 3.3 or 3.4, cause most beers I have are, IMO, fairly average, with few really good ones, and even fewer great ones (the ones at 4.0 and higher).

It seems everyone just automatically lops off .5 on their rating for "respect".

There is only one website that requires the "respect" you talk about, and it certainly isn't ratebeer.com.

[sprague11]

Glad to see that at least SOMEONE on here respects beer.

[cfrancis]

The thing I find funny about ratebeer is is fear of people rating anything above 4. When in reality that's an 80% beer.

How did you come to the conclusion that people fear rating a beer above a 4?
I use ratebeer.com regularly, and I rate beer as I see fit, using their scale and using it according to my tastes. But not once, have I feared rating a beer above 4. If it makes the cut, it makes the cut. My average rating is about 3.3 or 3.4, cause most beers I have are, IMO, fairly average, with few really good ones, and even fewer great ones (the ones at 4.0 and higher).

It seems everyone just automatically lops off .5 on their rating for "respect".

There is only one website that requires the "respect" you talk about, and it certainly isn't ratebeer.com.

I'm obviously generalizing. But with the more popular beers if you start looking down the ratings. There are a number of them that shy to the south side of a 4 but the narrative is full of "the best beer ever" It's just something that I've noticed in my two years on there. And I found myself doing it to "be in the pack" with my ratings as my palate developed. Now that I feel more confident rating beers, I'm looking at the general ratings and seeing a trend.

It by no means includes everyone, because you can always get the guy going "I can't feel my legs, this shit is awesome" or the raters that rate what they want and not be afraid to potentially stand out from the mob.

[northyorksammy]

The median beer on BA is a 3.55, a top-200 beer clocks in at 4.22
For RB its 3.2 and 3.87
(this was as of a few months ago)
there is a .35 difference and it has to do with the way the questions are asked.
When I view the larger samples at RB, I automatically add .35 to know what its like in the BA world, it works virtually every time for beers with a minimum sample size, i.e. 10 reviews beyond the home city or country

Similarly, any sample's answer scores are a reflection of the way questions are asked.

[matt7215]

current as of dec 11th 2010:

1 - Denisons Weissbier
2 - Black Oak Ten Bitter Years
3 - Scotch Irish John By Imperial Stout
4 - Flying Monkeys Smash Bomb Atomic Strong IPA
5 - Wellington Imperial Russian Stout
6 - Durham Hop Head
7 - Denisons Dunkel
8 - Black Oak Nutcracker
9 - Mill Street Tankhouse Ale
10 - Grand River Russian Gun Imperial Stout

about a year later we have:

1 - Black Oak 10 Bitter Years
2 - Denisons Weiss
3 - Flying Monkeys Smash Bomb
4 - Wellington IRS
5 - Muskoka Mad Tom
6 - Durham Hop Head
7 - Speahead Hawiian Style Pale Ale
8 - Denisons Dunkle
9 - Flying Monkeys Netherworld
10 - Amsterdam Tempest