A perfect 5.0 in an “expert” rating means maxing out all 5 categories. But most people are not giving maximum scores in even one category easily, let alone multiple or all of them. That way, ratings not going up all the way to 5.0 are in a big part due to the rating system used. I don’t think users don’t enjoy their beer if they don’t use the high end rating scale, it’s rather the system which inherently pulls user into not using it.
A short combinatorial comparison:
If I roll a 50-sided die and divide the result by 10, my chance of getting a 4.6+ is 10%
With two dice of 25 sides, the chances for a 4.6+ are already at 2.4%
If I roll one 20-sided die, two 10-sided and two 5-sided dice, reaching 4.6+ has a chance of 0.252%.
Reaching a number close to the maximum becomes much harder the more dice/categories you add.
I’m just pointing out that the rating system shapes the way ratings are distributed. The dice thing is a model how it does that. With the consequence that the low amount of high end ratings is most likely not due to snobby users which are impossible to please.
Exactly. And because of this, you get something that approaches a bell-curve for the distribution of the total even if your distribution of the scores was uniform (and Iznogud’s claim seems to be scores should be uniformly distributed). In the limitting case, you get exactly a normal bell curve, this is the strong law of large numbers. Which is why any normally-distributed example works as an example to highlight the fallacy.
Sorry, but I am not suggesting what you’re saying. I don’t expect anyone to have 13% of rates in that range.
I’m simply suggesting you guys are not using 4.5-5 range and your ratings support that.
I think I have to agree with that.
I think that using the 5 attributes score is reducing the higher/lower score range because 1 or 2 “bad/average” attribute scores reduce the overall score of the beer…even if the Overall attribute Score is high.
That doesn’t mean people are snobbing beers with lower scores than those on UT. On UT people scores are just pure Overall score.
I myself is a perfect example of this trend…I got almost nothing in the 0-5-1.5 and 4.5 to 5.0 ranges…
~15 years ago, I decided on what the 0-5 scale meant to me. As I started on RB, I even wrote down a list of what the scores 1-10/1-5 meant to me, and I’ve been brutally consistent. If my review seems to be mostly those keywords, it gets that score. One of the things I designed in from the start was that middling beers would get middling scores, namely 2.7-2.8. So my keywords for total score would be along the lines of: 2.6=dull/poor, 2.8=vaguely adequate, 3.0=adequate 3.2=decent, 3.4=good, 3.6=very good, 3.8=excellent, 4.0=wow (or other lack of words). Some people seem to be about 0.5 higher for each of those classes, that’s their choice, whatever works for them.
My words-numbers mapping may not have changed, but some parts of my tastes have. I used to bury lagers as I never found anything interesting in them, but I’m now learning to appreciate them, and am giving them slightly higher scores. So the first few thousand were a bit unlucky, but I can’t rerate them all.
Yes and no and yes and no…
What makes UT reliable is… It is reliable. Not in a score or review wise, ohh satan in Trump pants no (thou I and you even know guises and grls who write serious reviews in UT), but in search speed and accuracy, places and breweries taking care of their own abominations etc. To snob consumer, as we are, super duper handy. Common sense helps to swim around anomalies of UT thou.
In comparison RB has gone from bad to worse for me with every update. I have tried to use RB on my trips, but if search takes 20 secs vs 1 there is smth done wrong. Constant disconnection when Wi-Fi is low etc think of all the common problems you hit when travelling on budget and RB fails for me. I hope it’s gonna change, but I can’t see it. Ericks app had the best feature ever, offline rating, the most coolest idea ever, and most short sightedly it was killed by… add your own words here… Stupid act after stupid act after… Ohh fck I turned into rantPhil I better stop here
