I look for IPA's I haven't tried yet, then IPA's I have tried and love, then move on to other styles starting from least expensive up.

quote: Originally posted by Cottrell I buy beers I like, and beers I've heard good things about. ... I have trouble understand [sic] the reviewers who pick up stuff ... just for the sake of boosting their number or reviews.

I like this thread, too. Southern seems to have a knack. Now, crack open a quality beverage, unbunch yer undies, and ponder these ponderables: --If I only ever buy what I already know I like, how will I discover something new that I like even better? --If I only try something new that someone else has already praised, will my own praise be valid? --If I gagged on Sam Adams Triple Bock in 1996, should I save my money and never try it again? Tastes change--thank God. I'm willing to try stuff for the adventure of it. My definition of favorite style has changed over time. I like that BeerPal encourages exploration and development. I'm probably not going to fall in love with that Red Stripe chilling in the Foam Fridge, but what the heck, I'll give it a whirl. And you know what, even if I don't care for it, I am going to enjoy adding Jamaica to my passport, mon!

quote: Originally posted by FoamDome Now, crack open a quality beverage, unbunch yer undies, and ponder these ponderables...

I couldn't agree with FoamDome more. People who only drink what they think is good stuff have very boring reviews. What's the validity of having 90% of your reviews between 3.5 and 4.5? I once played with the stats on the site to put everyone's median review at 3.0, which is the middle rating at BeerPal. I even spread them out so that their highest and lowest ratings went from 1.0 to 5.0, in cases where they never did rate that high or low. This way a person's reviews truly did follow the scale of 1.0 = worst, 5.0 = best, and 3.0 the median. I spent at least 10 hours slaving over equations. Once every rating was normalized like this, I computed the beer ratings again and whoah! It was really strange. So strange that I knew too many people would be pissed off if it replaced the existing rating system. THe problem was that people who gave all of one style something like 4.5, 4.4, 4.5, 4.3 lost most of their high value because they hardly differentiated between good and bad! Anyway, I got frustrated with the number games and put it off to the side. Now I'm trying to remember if I had a point. Don't rate predictably?

I look for anything new that is an import and then for anything from a "domestic" micro. I don't think I'd go for an untried swill just for the sake of a review; I'd rather go back to one I've had previously and liked.

Flash, that kind of a system is somewhat interesting, but it just doesn't work to normalize over that singe individual user. Some have been much more selective in their drinking than others, so their median SHOULD be higher, even if they go broad and try lots of different styles of beers. For example, I went through all the styles and tried to pick higher rated selections within each style. That means my mean should be shifter higher. I have more recently killed beers within a style that are just not up to par (the most recent being that nasty Witbier from Spanish Peaks.... just a terrible example for the style, IMO). A normalizing system should normalize for each beer relative to how EVERYONE else ranked that beer.... Then come back and indicate how "elevated" or "deflated" your rankings are. I'd love to see how overhyped or underhyped my ratings are. Ok, that's enough tech talk. Bringing out the engineer.

quote: Originally posted by eaglefan538 Flash, that kind of a system is somewhat interesting, but it just doesn't work to normalize over that singe individual user. Some have been much more selective in their drinking than others, so their median SHOULD be higher, even if they go broad and try lots of different styles of beers. For example, I went through all the styles and tried to pick higher rated selections within each style. That means my mean should be shifter higher. I have more recently killed beers within a style that are just not up to par (the most recent being that nasty Witbier from Spanish Peaks.... just a terrible example for the style, IMO). A normalizing system should normalize for each beer relative to how EVERYONE else ranked that beer.... Then come back and indicate how "elevated" or "deflated" your rankings are. I'd love to see how overhyped or underhyped my ratings are. Ok, that's enough tech talk. Bringing out the engineer.

But that is not valid. You cannot assume that it's not the actual median if you do not give proof (by rating) a lower beer. In reality, their reviews could be accurate if they did start drinking some worse beers, but we don't know how true that is. Another problem is some people rate a beer in comparison to others of the same style, and others do not. My normalization was being made for each style. That way if a person was already rating based on style, the normalization was minimal. In the case of others, the normalization fixed corrected it. The crappy part was that many people didn't have enough data for each style. I think I was using 5 ratings as a minimum. Look at certain styles, such as abbey quads. I don't think there is one beer rated below a 3. Obviously, one of them has to be the worst (a 1.0) and one is the best (a 5.0). The beer scoring equation already takes into account everyone's ratings, anyway. If I normalized it more, then every beer would be a perfect 3.0. The purpose of what I was doing was to normalize each person's biases to each style. Also, the standard deviation is listed for each beer, and that will tell you how inconsistent all the reviews are. The lower the number, the better. If you want to see how inflated or deflated your reviews are compared to everyone else, I could easily do that. What I'd do is take the deviation of your rating from the mean rating of each beer and then find the mean. Sort of like a standard deviation, but with a positive or negative value.

I hear ya Flash. If the 1-10 scale is relative to only that one style, then my ratings would change. I use them as absolute scales of what I want in a beer (average across all styles or not). The hardest part for me is making the rating subjective when I don't care for a particular style in the first place. For, example, I would never drink an entire six pack of SN Celebration Ale.... but it was a darned creative and well done beer. I don't know that I want my ratings normalized WITHIN style, cuz all styles aren't the same. A best swill isn't worth comparing to a mediocre quad. If we could sort within style (hint hint), rather than drilling to the top of the list of one style, then that would meet my needs. The deviations I was referring to was certainly not aimed at getting the same thing you were trying to get at. But the "deviation" calculation you mentioned is exactly what I was meaning (however you have to square it, sum, and take the sqrt, cuz a +/- mean won't work mathematically). Regarding Red Stripe, Foam, I'm on the same page. I'm actually liking some of the foreign swills I've been drinking during summer activities (or lawn mowing, as you mentioned). I think there's a spot for just about all beers, despite the fact that there are far superior beers out there.

quote: Originally posted by FoamDome <blockquote id="quote"><table width=90% cellpadding=10><tr><td bgcolor=FFFFCC><font size="1" face="Verdana, Arial, Helvetica" id="quote">quote: Originally posted by Cottrell I buy beers I like, and beers I've heard good things about. ... I have trouble understand [sic] the reviewers who pick up stuff ... just for the sake of boosting their number or reviews. </font id="quote"></td></tr></table></blockquote id="quote"> I like this thread, too. Southern seems to have a knack. Now, crack open a quality beverage, unbunch yer undies, and ponder these ponderables: --If I only ever buy what I already know I like, how will I discover something new that I like even better? --If I only try something new that someone else has already praised, will my own praise be valid? --If I gagged on Sam Adams Triple Bock in 1996, should I save my money and never try it again? Tastes change--thank God. I'm willing to try stuff for the adventure of it. My definition of favorite style has changed over time. I like that BeerPal encourages exploration and development. I'm probably not going to fall in love with that Red Stripe chilling in the Foam Fridge, but what the heck, I'll give it a whirl. And you know what, even if I don't care for it, I am going to enjoy adding Jamaica to my passport, mon!

I totally agree with Foam's comments... and thank you for the kind words! I am enjoying beers all the time that I used to hate, forcing me to revise my reviews. I also agree that if you don't try anything new you won't find what could possibly be your next fav. But I think I'm going to take a conservative approach. If it's from a brewer that I've never liked, or a style that I have never liked, or something that simply looks like something I won't enjoy, I will not get it. Where I am right now (300 something reviews) I feel there are so many choices out there to take sort of an educated guess on a new beer. I think cost is a factor too. I'll take the gamble if it's cheap, but maybe not if it's gonna cost me! As far as Flash's comments, I am very interested in his findings, and have gave that some thought too. Sounds like a whole hell of a lot of work though. I appreciate the interest in this thread. Seeing how we all try new beers all the time, I have been curious for a while what everyone's "strategy" was.

A suggestion: have a tasting party for 10. Get 10 clear 4 oz glasses. Everyone brings 40 oz of beer. The bottles are masked, so you don't know what your're sampling. You are not influenced by words like "swill." No marketing ploy can sway you, since you cannot even read the label. Now, every taster plays judge, rates each beer on its own merit, votes for his or her favorite, and tries to pick the one he brought. Do this to get away from preconceptions associated with marketing. Do it to really test your tasting skills. Mostly, do it because it is FUN.

quote: Originally posted by FoamDome A suggestion: have a tasting party for 10. Get 10 clear 4 oz glasses. Everyone brings 40 oz of beer. The bottles are masked, so you don't know what your're sampling. You are not influenced by words like "swill." No marketing ploy can sway you, since you cannot even read the label. Now, every taster plays judge, rates each beer on its own merit, votes for his or her favorite, and tries to pick the one he brought. Do this to get away from preconceptions associated with marketing. Do it to really test your tasting skills. Mostly, do it because it is FUN.

Yeah, that does sound like a whole lot of fun!