First of all, sorry for my grammar and my general English. A bit hard to phrase this question properly, but I'll try.
I have created a thread in a sub sports forum asking people to rate this years football players from 1-5 (1 being the very best). I wanna make a stat on how well people did in their rating of players when comparing their rating to the outcome. So if one person (let's call him Frank) gave this as their answer:
1. [player name A]
2. [player name B]
3. [player name C]
4. [player name D]
5. [player name E]
But when adding up all other answers, it turns out that this is the general opinion of this sub forum regarding the players ranking:
1. [player name B]
2. [player name E]
3. [player name C]
4. [player name D]
5. [player name A]
This will be the key/solution/answer.
So Frank - the first list - got correct on 'C' and 'D', and 'B' was just one spot away from being correct.
How can I properly rank all the users answers? If I'd use only correct and wrong answers I would use percentage. That means Frank would have an accuracy of 2/5 (2 being correct out of 5). Thereby 40% accuracy on his answer. But when I wanna give credit for /almost/ being correct, it gets tricky and my simple head can't find a solution.
Any ideas? Rank answers with points? A formula of some sort?