Because it was on their front page I looked over sherdog's mma rankings posted today. I don't normally give a crap about that sort of thing but found their rankings pretty absurd. Randomly popping around the forums I've seen people accuse the site's staff of being anti-UFC and always figured that was pretty silly and probably sour grapes.
However, their criteria for dropping someone from rankings is inactivity in a weightclass for a 12 month period. How then does Gomi warrant No1 when his only fight in the weightclass was an overturned loss? Wtf? He certainly doesn't warrant No1 for the previous 12 month period and shouldn't even be on the list given their own criteria. Similar, despite agreeing that Fedor is probably the best, how do you rank number one with your two fights being against a freak and a middleweight.
I've wondered if we could do a better ranking system here and ensure somehow that it was truly objective.
Might be worth looking into. Not sure how to do it though.
You could just do it the same way everyone else does.
Originally Posted by Phrost
Get two or three guys who watch fights together, talk about fighters together, probably train together, and pretty much agree about everything type up a list, and post it for the world to see as definitive and unbiased unbiased.
Then, everyone gets to argue about it.
What's so hard about this?
I think a good system could be put in place. Initially work it out statistically, then take individual differences into account and mitigating circumstances. Then do some stats on the people in the top ten. I.E would you place forrest above shogun?. Would be quite fun working out.
Or we could just vote on it.
This smells too much like the BCS.
Originally Posted by Jadonblade
A good idea, being taken too far.
The reason I made the sarcastic post before is that ultimately, thats what any workable system done by unpaid people will come down to.
As far as the recommendation quoted above, statistical analysis works pretty good in baseball, where everything is easily quantified and there are so many games played each season, anomolies iron themselves out via a large sample size. In Pro Football, its fairly useful, but inevitably incomplete.
Rankings - either via formula or human opinion - work well enough for determining large bracket tournament seedings - like in basketball, golf, and tennis. Enough spots exist that a really good competitor won't get passed over, and there is enough room for the scrappers to claw their way in and have a good run.
Just over half the time, rankings barely work well enough in lieu of a playoff off system in college football. And thats with really good regression generated formulas blended with the opinion of hundreds of professionals.
I just don't think they work well enough for this to justify their use. Or in this case, their utter lack of use.
What good is any independent ranking system no fight org pays attention to? Even if it is good to be used, injuries, rest time, the ability to sell a PPV and personal feeling are as involved.
All another ranking system will give people something else to argue about. There is plenty enough material to argue about fights that are actually going to happening (or already happened).
Ranking system is ok...!! I think it is valid that some people's opinions matter more for the direction of Bullshido. You could have bullshido account level determine the "points" like:
5: Instructor, Fisted, etc
3: Regular Bullshido BBC members
1: Regular members (however if account creation date < date() - 10 days or postcount < 10 then they are 0)
Then ensure only one vote per account per grouping is used a month -- actually one vote FOR a fighter and one vote AGAINST a fighter.
This ensures someone doesn't register a bunch of accounts to force their stupid-ass hero Michael Bisping to the top of a list.
A descending tree of surveys could be taken on the bullshido.net members. The initial survey consisting of a simple ranking of who each person thought fit where in the different weight classes. The next survey focusing on the nuances of each fighter in one weight class, to square that weight class away. The following survey would be the same type but in the next weight class. Repeat until all weight classes are fulfilled. Additionally, I think there should be some criteria for who should vote: ie. you must have been a member of bullshido.net for this long and have made a minimum of this many posts, say 50.