There's a lot of controversy on Board Game Geek, ignoring the bulk of it, I want to focus on one that comes up time and again. The game ranking system.
Ranking Games to avoid hyper numbers
The Main issue, in board game rankings, complained about the most, is the assignment of a value system. The 'arbitrary 1-10 vote' you assign to a game. Anyone can vote for any game, and as a result, there is a very easily abusable way social media can push a game into the top 100, when it hasn't even been played, or people voting down a game to offset this.
A system that could replace this, has to allow for people to vote, has to include some bias for new accounts vs long term accounts, but also, needs to allow for the varied ways people vote.
If a game has just arrived at your house, and you play it, and love it, surely you're going to vote it into your top 10, but as time goes by, and you get newer games, should you need to go back and vote that game down, so you can allow for your next great loved game to take that top position?
Games that last the test of time, need to be able to rise above. Games which are a fad, games which are fun today yet boring tomorrow, who is going to go back and edit all their previous posts & scores and change to reflect this?
Sites haven't yet incentivized people to do this, and they could, but they don't feel the need for it, and it'd be a lot of work for very little gain.
But, having thought this over for many years now, this one clicks for me:
The game is not stored as a number, instead its stored as a list. your ranking, of your games, from least to most, in three very relevant situations.
Computers, which can crunch numbers quickly can assign a weighting to your list from 0 to 1, and use this weighting to 'list' all games in the system.
This will create a top 10, or top 100, without needing to score them as 10s. The top 10, will reflect the games that of all games voted for, so far, these games have the most people, saying it is their best game.
There will be no need to chastise the hundreds of people saying its a '10' because they haven't, they've just said, of all the games 'they' have played, its in first place.
Anyone who votes a game down to 'tank' its ratings, are doing nothing more than voting all the games above it, as being better. In their opinion.
Participants, who have only ever voted for 10 or 20 games, wont be able to influence the system to say their no.1 is the best game ever (or in last, as the worst), because, its only the best game, for them, out of 20 games they've played. (see below for the math)
Furthermore, instead of asking "is this 'the' best game", your asking, Is this game, the better choice, based on your previous rankings, of other games you've experienced. and its asked 3 times, in three ways, that often, don't match. new players, games for your friends, or games for you.
Furthermore, instead of asking "is this 'the' best game", your asking, Is this game, the better choice, based on your previous rankings, of other games you've experienced. and its asked 3 times, in three ways, that often, don't match. new players, games for your friends, or games for you.
And this is what we want, we don't want games with 1 vote, jumping into top place, an we don't want some kind of 'catch system' to say, Oh, we need to have 300 votes before we publicize the number, because, once social media finds out, they'll get 301 people to vote it a 10, and then the opponents of this, will vote 1s to cancel out the ratio, we don't even need the arbitrary 1,000 in there, just give all uses a base 0.5 middle and scale accordingly, sure the math will end up with values between 0.4 and 0.7, but that's also arbitrary, because, the users will only see the 0.7 as 'first' and the 0.4 as 'last'
And if anyone requires to see a 'value' score, then math can easily spread a 0.4 - 0.7, into a
And if anyone requires to see a 'value' score, then math can easily spread a 0.4 - 0.7, into a
No comments:
Post a Comment