Hope the problem can't be fixed

[sixtywatz]

Sites down for repairs due to the ratings being completely lost and all rankings as n/a

Imagine if it had to start over. Would we have the same top 10?

[Brandon]

Would actually pretty interesting, maybe they'll figure out how to combine expansions and limited editions into the base game for ratings as well. 

[Ron Parker]

(And... someone at BGG noticed I was there and shut down the beta site too. Heh.)

On Wednesday, January 9, 2013 11:55:29 AM UTC-5, Ron Parker wrote:

The ratings aren't really lost. All of the individual ratings are still there; it's just the cumulative totals that have been wiped out.

(How do I know this? I'm looking at Die Macher on the beta site right now, and it says rating N/A ranking N/A, but it shows the correct graph of individual ratings, it says that 3687 users have rated the game, and it shows all of their individual ratings.)

[leroy43]

If it were up to me, and the ratings were redone from scratch, I'd institute a system where each user would have to choose games simply by preference over other games, a lengthier process whereby users would have to make binary choices of x over y until their games were ordered. 

[Thanos]

It would be an interesting experiment, certainly.

I know many games that have gotten great (or poor) ratings right after they came out, but would definitely not get the same ratings a year+ later.

[Daniel Vore]

I noticed this right before it went down. lets hope it gets fixed soon. I've got nothing to do at work today.

[abrannan]

I'd rather the composite rating were adjusted for how far over or under a user's average rating they were (if you want to get fancy, maybe by how many standard deviations).  I tend to start my rating for a game at 7, and adjust up or down from there.  I suspect there's probably a user who does the opposite (Starts at 3).  With the more computationally intensive system, you'd theoretically get a more accurate spread of ratings.  It would also slow down shill voters, as anyone who only rated one game a ten would be giving it an average rating, and even those who rate one ten and a few ones and zeros would have a wide standard deviation so the impact would be negligible.

[Ron Parker]

It doesn't exist. Never did.

On Wednesday, January 9, 2013 12:10:35 PM UTC-5, Andrew Corpuz wrote:

Where is this so called beta site you speak of? :P

On Wednesday, January 9, 2013 8:55:29 AM UTC-8, Ron Parker wrote:

The ratings aren't really lost. All of the individual ratings are still there; it's just the cumulative totals that have been wiped out.

(How do I know this? I'm looking at Die Macher on the beta site right now, and it says rating N/A ranking N/A, but it shows the correct graph of individual ratings, it says that 3687 users have rated the game, and it shows all of their individual ratings.)

On Wednesday, January 9, 2013 11:47:19 AM UTC-5, sixtywatz wrote:

[ddyer]

The most important flaw in the rankings is that it includes the opinions of dead people.
(or rather, people who haven't logged in for years).  There ought to be an expiration
date for an opinion.

[NuMystic]

I don't see how an opinion is any less valid because someone isn't active on the site. So you think Amazon reviews should only include those that have bought something from Amazon in the last 6 months, and if not the review should be purged? 

Heck, just look at Out For Blood who actually is deceased. His contributions to the community were immeasurable and his opinions SHOULD be immortalized. And if that's so, why should anyone else's not be equally so.

That said, you did bring up an interesting point. Dead opinions rather than "dead" people.

Many do not update their reviews as often as they shift and change. I think it would be a fantastic feature if BGG prompted you to "confirm" your rating of each game once a year. I guarantee that there would be a LOT of shuffling of scores and rankings then. 

[ralpher]

The rating system is flawed. Fewer than 10% of games have an average rating of 1-5 and most games are clustered between 6-9. The rating system should be divided in half, from a 1-5, 1 being bad, 2 being fair, 3 being good, 4 being great and 5 being excellent. Much easier to understand and for people to rate consistently.

[leroy43]

So that most ratings can be clustered in the 2-4 range instead of the 6-9 range? :)

[ddyer]

My underlying point is that the system matters depending on what you are trying to measure.  As a thought experiment, imagine if BGG had been running for 500 years, with cumulative rankings for all that time.  Go and Chess would be #1 and #2 just because over time, lots of people played them and thought they were great.  Maybe that wouldn't be a bad idea, but it would make it impossible for a newcomer like (enter your favorite game here) to even get on the charts.

[NuMystic]

You're forgetting that games aren't ranked on BGG merely by average rating, but BGG Rank which is a more complex algo. Newcomers constantly crack the top of the charts. So your point is what exactly?

[NuMystic]

Case and point, 3 of the current top 10 games are from just last 18 months… Eclipse, Mage Knight, and Android: Netrunner