Arrow tier...
Also I did something awhile back on UMVC3. I had folks vote on the 12 newcomers to the game in a survey and honest to God use statistics called box and whisker plots to determine predicted tiers for us to review after the game came out to see how well we read things (which is MUCH harder in this case with no meter or damage info to go by).
Here is how it appears.
The way it works is you fill out the a numbered tier list survey. Just voting 1 to whatever in rank.
I take the MEDIAN and that is the middle of the box and whisker plot. The central number. Now here is where it gets really calculated though. I put all the voted ranks in order and take the median of the numbers on either side of the true median. You don't take the average here because outliers that are just like "Hsien Ko is number 1!!!" happen and throw off results. Just the median .
So what you get is a middle ground, and a box surrounding it that bridges several number ranks. The smaller the box the more generally sure folks are on the tier of a character. The wider the box the more uncertain folks appear to be on the character. If the median leans further to one side of the box than the other then its likely this character will move up a tier as more information comes about. Outliers are the dashes extending off the ends of the plot.
Here is how the Box and Whisker plots for Marvel turned out before launch.
Even before launch folks were predicting Vergil pretty heavily to be in a tier around number 1 in those 12 characters. Likely he would have ended up just as high had the ranking gone all the way from 1-50 and included the whole cast and not just the 12 newcomers. His median rank was around #2 but his voted tier ranking said he'd be around 4th to 1st in rank though.
Firebrand's median was so high that had more tech been discovered they would have increased his tier and narrowed his box. However, his matchups and damage were bad so his median began to shift down in his tier box and eventually his upper quartile median dropped too. Frank West's box later shrank and his median and upper and lower quartiles all shifted higher on the list as folks figured out how easy he was to level.
You get the idea of how these work now.
Something like this is actually pretty decent and If I could get a coder to generate me a site based on taking surveyed numbers and pumping them through to generate a graphic like this every month I could honest to god have a fairly accurate tier list generator for any fighting game that could evolve every month. Tiers aren't number rank. They are generally based on how good someone is with a character. Knowing in general that a character tends to hang around a certain rank number gives you an idea of how strongly they are built. Plus the outliers let you know that there are still players who think they can beast with them even if they aren't the norm.
If you guys wanna do something like this and then retake the survey after launch the comparative results might be interesting. The results on the pre launch survey would likely be all over the fucking place and on the post launch survey they would likely be laughable altogether, but they are great for looking back upon. Doing a survey like this yearly to tri monthly is acutally something I think communities might benefit from.
