The idea is simple. Evaluate another fortress player across several areas on a 1-10 scale. 10 being the best, 1 being the worst. Collect as many evaluations as possible and average them. See who comes out on top in each area.
Edit: Change in procedure. Rather than attaching your name and IP to each rating, each rater will now get a code, registered to your name, Global ID, email address and IP address. These four criteria for identification should make falsifying ratings almost impossible, and give us a lot of different ways to confirm true identifies should we need to. It also gets rid of rater names attached to their ratings. I was going to strip that anyway when publishing results, but this allows me to publish mixed intermittent results allowing each of you to identify your ratings but your ratings alone. (Ratings that have already been submitted are fine and valid. No need to resubmit)
(To clarify, evaluate as many players as you know, not just one)
Oh boy, this is going to be interesting! I’ll come back and add some later. Gotta think about this all serious-like…
aw man something like that was going to be part of my plan for pru’s return
(To clarify, evaluate as many players as you know, not just one)
yes. I wanted to challenge everyone on Arma forums to beat me to get some attention before I start recruiting
well, I’ll do it anyhow
I should mention your IP is attached to the rating as well, I’ve already seen some attempts at distortion.
Damn, this is hard. I should have decided before hand who I wanted to submit and picked scores relative to each other. I think my submissions are all over the place. And it’s hard to keep track of who you vote for so I’m going to stop now. I learned my lesson for next time.
I was actually thinking about this and how it could be ‘easier’ to do the rating. Instead of rating per person, you could do it so you just pick the better person from between two people (randomly picked from a certain pool) for one quality. Then you do this a large number of times and with a lot of different people and calculate (similar to dlh’s armarating?) a rating out of it.
For example:
Best defender:
Pokemaster vs Shoebat
(possible options:)Pokemaster/Shoebat/Don’t know
I like Lacka’s idea, I think for now many people won’t do this because they’d like to have player lists or they generally do not want to spend so much time on rating others (not speaking for myself, I am too interested in the results to miss this chance).
the problem with Lacka’s idea is that it presents the opportunity for paradox
Let’s say we are trying to rank 3 players defending skill from best to worst. Poke, Shoe, and Concord
This doesn’t make any sense. And maybe a rater won’t make a loop this simple. But if he’s rating 50 people, there’s a possibility for inconsistency.
If the person just assigns each a value 1-10, we get more information from fewer ratings and no possibility for paradox.
I wouldn’t say it’s a paradox as much as it just shows that the teams are probably equal. The outcome of each choice should not be taken as some kind of truth: just because Lack picks Shoe over Concord doesn’t mean Shoe is actually better than Concord.
I guess it would be better if you imagine the results of the pick as the winner of a matchup, if you look at it like that it’s perfectly fine/normal. For example: it’s hardly a paradox if you let three teams play eachother and team1 beats team2, team2 beats team3 and team1. It just means they are more or less equally matched.
When you let them play for more games, and against more teams, it should become obvious which team is actually the better one.
I’ll be posting the results sorted by submitter (stripped of the submitter’s name)
I’m going to wait a bit on this, so people aren’t rating based on what other people say.
(I’ll try to set up some automated receipt too, and I’ll email your bunch, sine)
If anyone wants to rerate, just submit again, and the old will be discounted.
Edit: Change in procedure. Rather than attaching your name and IP to each rating, each rater will now get a code, registered to your name, Global ID, email address and IP address. These four criteria for identification should make falsifying ratings almost impossible, and give us a lot of different ways to confirm true identifies should we need to. It also gets rid of rater names attached to their ratings. I was going to strip that anyway when publishing results, but this allows me to publish mixed intermittent results allowing each of you to identify your ratings but your ratings alone. (Ratings that have already been submitted are fine and valid. No need to resubmit)