ENnies V - and beyond...

[Old Fezziwig]

It would seem more logical to vote simply for your first and second choices out of each list. That should be enough to allow a fair determination of a winner, and keep "manipulation" to a minimum.

This seems like a really good option, simple and elegant, but I'm sure that someone more knowledgeable than I am about this could probably tell me why this won't work as well as it would seem to work. I can't, right now, see a problem with it.

Best,
Nick

 

[nikolai]

Countering Vote Stacking

Note that you can still manipulate this system, by voting 10 for you favorite, 1 for your anti-favorite, and 5 or 6 for everything else. Note that if you want to manipulate the system, it will always be in your best interest to score *all* choices, even if you haven't even heard of them.

One, old and simple, way to assign an point value to something is to take the Median, instead of the Mean, of the votes. This isn't a new idea, but isn't implemented very often on online rating systems, and counters attempts to push the number up or down through tactical voting. This would solve most the "influence" arguments.

I'm not suggesting that taking the highest Median is the best voting system. I think there's still thought needed on (1) what criteria is wanted for the Ennies voting system, (2) and which is the best system to meet it. But if you want to juryrig the current system to counter vote-stacking, this is likely the best way to do it.

On the other hand, this would be a good way of changing the rating system for the Enworld reviews section, and would have an effect of the positioning of the Book of the Righteous - for example.

 

[Umbran]

This seems like a really good option, simple and elegant, but I'm sure that someone more knowledgeable than I am about this could probably tell me why this won't work as well as it would seem to work. I can't, right now, see a problem with it.

Some problems cannot be seen until you know how the vote is processed. You vote for a 1st and 2nd favorite. Fine. Now how do you select a winner from the stack of ballots? What do you do if a voter doesn't have a second favorite?

 

[Conaill]

It would seem more logical to vote simply for your first and second choices out of each list. That should be enough to allow a fair determination of a winner, and keep "manipulation" to a minimum.

That would work. One issue with this approach is that you don't distinguish between choices you're unfamiliar with and those you *know* to be worse than your top two. We could add an additional "don't know" option, but at that point it already becomes so complex we might as well let the voters rank all the ones they're familiar with.

 

[Crothian]

Because otherwise I don't see the point of a scoring system that ranks from 1-10, best to worst. It would seem more logical to vote simply for your first and second choices out of each list. That should be enough to allow a fair determination of a winner, and keep "manipulation" to a minimum.

Because then market pentration is what matters and not what you thought was best. We did this orginally and it was easy to see that the books that were more well known were voted for more often not because they were best but because more people were familiar with them and since they were not familiar with the others, they didn't vote for them. It really had little to do with what book was best because few people actually have all 5 books nominated in a category to form an opinion on all of them.

 

[Old Fezziwig]

Some problems cannot be seen until you know how the vote is processed. You vote for a 1st and 2nd favorite. Fine. Now how do you select a winner from the stack of ballots? What do you do if a voter doesn't have a second favorite?

I believe this is what they use for most of the major awards in baseball, and points are assigned based on what people vote (in a 2 place system, probably 2 points for a 1st option, 1 for a 2nd). You add up the points and then whoever had more points wins. This is just a way to do it, the weighting could be done differently.

One issue with this approach is that you don't distinguish between choices you're unfamiliar with and those you *know* to be worse than your top two. We could add an additional "don't know" option, but at that point it already becomes so complex we might as well let the voters rank all the ones they're familiar with.

Fair enough. That's something that hadn't been considered. But, at the same time, is it really important to differentiate between the things you wouldn't have chosen and things you haven't heard of?

Best,
Nick

 

[Old Fezziwig]

Because then market pentration is what matters and not what you thought was best. We did this orginally and it was easy to see that the books that were more well known were voted for more often not because they were best but because more people were familiar with them and since they were not familiar with the others, they didn't vote for them. It really had little to do with what book was best because few people actually have all 5 books nominated in a category to form an opinion on all of them.

I guess what this avoids, though, is people voting against products -- if you can't score everything, then you can't be giving 4/5 products a 1 and then your favorite a 10. I'm absolutely sympathetic to wanting to avoid the results being determined by market penetration, but this seems like it might reduce some manipulation.

Best,
Nick

 

[Conaill]

Little tidbit for any other voting theory nerds out there...

You can estimate the amount of flexibility you give the voters by calculating how many bits of information their vote contains. The more bits of information, the more detailed they can express their opinion, but typically also the more opportunity there is for "strategic" voting.

First of all, the *outcome* of the vote will be only one #1 and one #2 winner. With 5 candidates, there are 20 possible combinations of outcomes (ignoring ties), so that's only 4.3 bits of information.

Here's how many bits of information each voter can input into the system, if we do NOT allow a "don't know" option:

1-10 score: 16.6 bits
1-5 score: 11.6 bits
1-5 ranking: 6.9 bits
yes/no score: 5 bits
pick a #1 and #2: 4.3 bits
pick a #1 only: 2.3 bits

Here's how many bits of information per voter with a "don't know" option for each candidate (which, depending on the method, could just mean leaving a candidate blank):

1-10 score: 17.3 bits
1-5 score: 12.9 bits
1-5 ranking: 8.3 bits
yes/no score: 7.9 bits
pick a #1 and #2: 7.3 bits
pick a #1 only: 6.3 bits


Note that in the current implementation, the output only contains 4.3 bits of information, but we allow each voter to input 17.3 bits of information into the system (more than 4 times more than the information content of the output)! Hence my "overkill" claims earlier. In comparison, a partial ranking would only take 8.3 bits, and approval voting (yes/no vote for however many candidates you want) would only take 5 bits.

Only allowing the voters to pick one #1 and #2 is not that much better compared to a partial ranking or approval voting, and the latter have the advantage that your opinion is still taken into account even if your #1 or #2 are not in the top 2 overall.

 

[Nisarg]

Manipulating the vote is playing with the numbers of the voting system with the express purpose of trying to get your favored product to win (or your unfavored one to lose), without regard to the actual merits of the product.

Well most times the products you favour or disfavour are for a reason. There is a reason for giving a product a low vote, and the ideology of the company that produces that product could certainly be a part of it.

In order to skip the standard White Wolf example, let's look at Avalanche.. would it be fair to give their product a low vote because of a propensity for putting scantily clad women on their covers? Because of what you perceive to be a sexist attitude?

What are the "rules" that define how its right or wrong to dislike a product?

I put the blame on the choice of voting systems.. if the people running the awards did not want people to vote politically against companies they dislike, they should not use that system of voting, period. Otherwise, its implicit in the voting system that any way you can legitimately vote is an appropriate way to vote.

Nisarg

 

[Nisarg]

Because then market pentration is what matters and not what you thought was best. We did this orginally and it was easy to see that the books that were more well known were voted for more often not because they were best but because more people were familiar with them and since they were not familiar with the others, they didn't vote for them. It really had little to do with what book was best because few people actually have all 5 books nominated in a category to form an opinion on all of them.

What's wrong with that? And, incidentally, how does the current voting system fix that? Or are you saying that the current system only takes count of the overall SCORE and not the number of people who voted for it?

If so, I think that's a pretty flawed way of doing things. Are you sayinga product that gets only one vote but its a 10 beats a product that got 1000 votes of all 9s?

In any case, I think numbers of votes is important, and if that reflects market presence, no problem! Getting your product into the market is one of the factors that should influence the definition of a "good product"... and the will of the people should be heard. That is, if we want these awards to be an accurate reflection of what the gaming public enjoys playing, and not an origins-style farce or an "indy" award.

Nisarg