You know I support what you guys are trying to do over there, but explain this to me.
Spencer Fisher was #10 or #11 at 170. Then after an update (over a period of time where he hadn't fought) he completely disappears from the top 20. How does something like this occur? It really hurts the legitimacy of your rankings.
TC--did he change fighting weight?
Well he's moving down to 155 to fight Stout at UFC 58, but that is well after the update was made. Plus, if he was top 10 at 170, you'd figure he'd be top 10 at 155, and he's nowhere to be found there, either.
If you see a listing of 20 names, you are looking at the COLLECTIVE rankings (combining statistical rating, and fan votes)
In Jan he was the number 1 rated 170 lber, statistically. Then, he may have had a good number of votes to combine and get him in at 9 (other more popular fighters tend to get LOADS of votes to place high in the COLLECTIVE).
Then if the fickle voters stop voting for Spencer in Feb, while the others continue to get loads of votes, he drops off.
I personally only regard the stat rankings (the fan votes are bunk). So, mmaranks legitimacy is not bad, it is the perspective of the fans (like Sherk being voted 6th for 170 lbs in Jan, but being 338 overall statistically)
They really need to hone their statistical method and work towards dropping the fan voting portion entirely. Every time I mention improvements to the statistical rankings, though, I am met with a stone wall. I don't know if they're too lazy to make any significant change or if they really think the method they have right now is fine, but it's obvious to me that it needs work.
im number 2 thousand something hoo-ray
Actually, the stat system is pretty solid...
It is based on the last 6 fights or the last 3 years of fights (not going beyond 6 fights)
So old fights are not quantified for current rankings
Additionally, if you beat a fighter, you are linked to him for the next 6 fights. Thus, if he starts beating tough guys (with good scores) the fighter you beat will have his stat score increase, and thus yours will increase... If he goes on a losing streak and his score drops, your drops to reflect that beating him may not be all that special.
So, you are linked to the last 6 guys you fought, and each of them is linked to their last 6, and so on. So one loss somewhere could have an impact on the scores of alot of people (albeit only slightly for fighters further removed from you).
There really needs to be some human element in this system, I think. I have suggested writers.
Let's say someone like Vanderlai injures his hand, and is never the same. Thus, he goes on to lose his next 4 fights.
It shouldn't impact Arona's victory--Vanderlai was certainly great at the time of Arona's victory.
I don't disagree for a rankings... But I think it is OK for the statistical 'rating' system that scores based on true actual results.
But I have always endorsed expert opinion for RANKINGS
" Actually, the stat system is pretty solid..."
Not really. If you just look at the stat rankings many of them are totally absurd. Jardine is the #4 LHW whereas Silva isn't even in the top 10. Rory Markham is #1 at 185, not counting Almeida who is retired. Goulet is #3 at 170 despite just getting KTFO by Ludwig who barely ever fights. The system is not the worst thing I've ever seen, but it has SERIOUS flaws.
"It is based on the last 6 fights or the last 3 years of fights (not going beyond 6 fights)"
This is a serious problem. As I've pointed out before, it penalizes guys for staying busy. I have no problem with the idea of a shelf life on fights but picking an arbitrary quantity to consider is ridiculous.
"Additionally, if you beat a fighter, you are linked to him for the next 6 fights."
This is another problem. You beat the fighter you beat. You didn't beat the fighter your opponent would go on to become 2 years later.
" There really needs to be some human element in this system, I think. I have suggested writers."
I think the mixing is a bad idea. A system should be entirely human or entirely computerized. Here, the fan voting just seems to be a way to try and counteract the ridiculous rankings that their statistical system will occaisonally produce. I am 100% against voting systems. If you're not going to use a statistical method, you should just use your head and figure out what makes sense in an intuitive fashion. Having a bunch of guys with totally different opinions vote can cause really wacky results as is seen with mmaweekly's rankings.
--- "If you just look at the stat rankings many of them are totally absurd. Jardine is the #4 LHW whereas Silva isn't even in the top 10. Rory Markham is #1 at 185, not counting Almeida who is retired. Goulet is #3 at 170 despite just getting KTFO by Ludwig who barely ever fights. The system is not the worst thing I've ever seen, but it has SERIOUS flaws."
You are ranking fighters based on you opinion of their abilities (as seen in competition) to determine where they should fit "right now" in the weight class hierarchy. OK... no problem there.
The mmaranks stat system does not do that. It is telling us something different entirely. It is not measuring who should be "ranked" higher now and for future consideration. It measures who, in fact, has better real, measurable, quantifiable accomplishments as a fighter in the recent PAST. It measures these past performances and quantifies them with a calculation. This is quanitative v. qualitative
The site does go on to simply apply these stats numerically to form an ordered list of rankings... You don't have to like that, just realize that it is clearly different than the qualitative rankings.
The best way to use this IMO is like a QB rating v. football writers saying that so-and-so is the best QB. The writers use a qualitative appraisal of the QBs, in which I am sure they included QB rating for consideration
Now, I agree that Silva is a top ten guy. He should be "ranked" as such. However, for close calls on whether he should be 3rd, 4th, or 5th, or whatever... we can look at the "rating" from the stat system and use that to help define the rankings by our opinion.
Honestly, it would be wrong to tweek the stat system to intentionally coincide with our 'opinions' of who should be ranked higher/lower.
--- "This is a serious problem. As I've pointed out before, it penalizes guys for staying busy. I have no problem with the idea of a shelf life on fights but picking an arbitrary quantity to consider is ridiculous."
6 fights in three years is pretty standard for guys at the higher levels. This is really not a problem.
--- "I think the mixing is a bad idea."
I agree... I feel that a rating (like a QB rating in football) can be a tool to help the "experts" evaluate fighters and concoct a "ranking". But ranking should be based on expert opinion and qualitative evaluation. The quantification process is something separate.
--- "A system should be entirely human or entirely computerized."
There can be both in existance that serve two different functions. That is why I dont like the collective or the fan vote.
--- "Here, the fan voting just seems to be a way to try and counteract the ridiculous rankings that their statistical system will occaisonally produce."
fan voting is senseless... while I feel that the stat system is solid for its purpose (it should not be considered for true ranking for MMA... IMO). I like it as a tool to give an additional credential to a fighter - like 8-man tournament champ, KOTC Champ, NAGA World Champ, and a rating of XXXX... and as an additional reference for consideration on how rankings could be influenced
Sorry for being long-winded, but I take the matter seriously
When your statistical analysis come up with 337 people that are ranked higher than Sean Sherk at 170lb, you should start questioning your formula.
"The mmaranks stat system does not do that. It is telling us something different entirely. It is not measuring who should be "ranked" higher now and for future consideration. It measures who, in fact, has better real, measurable, quantifiable accomplishments as a fighter in the recent PAST. It measures these past performances and quantifies them with a calculation. This is quanitative v. qualitative"
So you're telling me that Vanderlei Silva has less quantifiable accomplishments in the recent past than Jardine? Please tell me you're joking. Vanderlei beat Rampage twice (both times Rampage was a top 5 fighter), Nakamura, Kondo, and Arona. Jardine has beaten nobody anywhere near the top 10. There are MANY examples like this that prove the system needs major improvements.
"Now, I agree that Silva is a top ten guy. He should be "ranked" as such. However, for close calls on whether he should be 3rd, 4th, or 5th, or whatever... we can look at the "rating" from the stat system and use that to help define the rankings by our opinion.
Honestly, it would be wrong to tweek the stat system to intentionally coincide with our 'opinions' of who should be ranked higher/lower. "
This is absurd. You're using the system to justify its own existence. How are you supposed to improve the stat system if not by intuition and general MMA knowledge. By your standard, the system never needs tweaking because somehow it's not right to critique it based on our opinions of who should be ranked where. Well let me tell you, when Rory Markham is rated above Hendo and Franklin, THE SYSTEM NEEDS FUCKING TWEAKING.
"6 fights in three years is pretty standard for guys at the higher levels. This is really not a problem."
You're missing the point. Think of it this way. Two guys both get a big win that pushes them way up the charts. The first guy sits on his win for the next 6 months and does nothing. The second guy fights 6 smaller shows against mediocre opponents over the next 6 months just to stay busy. The guy who stayed busy is PENALIZED for doing so, his big win is wiped right off the charts by your system, while the guy who sat on his win and did nothing gets to keep the points he earned for it. Do you not see the problem here?
Please, go to BoxRec.com for a computerized ranking that works properly and is constantly being tweaked and improved upon.
Yves is overall rated at 25 - Kawajiri at 40 (that does not consider weight class, so they are actually much closer).
The Yves/Kawajiri match was more than 6 fights ago for both (and was nearly 3 years ago) so it is not being factored into their current stat rating for either guy
Both guys are 5-1 in their last 6 fights
Kawajiri (these are listed as overall, but do filter by wt class)
Win Mindaugas Laurinaitis - rated 1636
Win Vitor Ribeiro - rated 16
Win Jani Lax - rated 197
Win Kim In Seok - rated 1637
Win Luis Firmino - rated 210
Lose Takanori Gomi - 1
avg for opponents defeated - 739.2
Edwards
Win Hermes Franca - rated 276
Win Josh Joseph Thomson - rated 181
Win Naoyuki Kotani - 273
Win Hermes Franca - 276
Win Dokonjonosuke Mishima - 176
Lose Joachim Hansen - 33
avg for opponents defeated - 236.4
So essentially, Yves has been facing and defeating a higher grade of competition "recently". Laurinitus and Kim In Seok don't help Kawajiri in he ratings department.
This system encourages facing tough opposition.
Silva is rated 92... Jardine is 106
LOL @ Vera being ranked higher than Overeem.
Well Jardine is #4 if you look at the stat ranking at mmaranks.com, whereas Silva isn't listed in the top 10.
--- "Laurinitus would be Kawajiris 7th fight and should be left out. His 6th fight is Shaolin. He just "won" over Hansen who beat Yves before that"
The current ratings at mmaranks.com are through Jan
--- "And one of Yves big wins over Franca is not really as important anymore since he lost twice after losing to Yves"
That is accounted for Franca, because his score is now lower than it was when Yves defeated him back then. This current score is what affect Yves current score.
Sophos ama
As for Sherk... in the last 6, he defeated Strebendt... then who???
This is not a rankings based on who you think can beat who in a future matchup.
This is a rating based on who DID beat who in recent performances
Tomato... I see... Well, the Jardine thing must be because Jardine had a fight in Feb and they were so close before hand. The short list of 10 on that page is up throuigh Feb. The longer list on the other page is up through Jan.