Ratings - please read before commenting - Archived

[JCS]

The winner of Shannon Briggs vs Brian Minto should be the number one fighter! Undoubtably the best, super duper that it can possibly be achieved.

JCS, re-read the stuff about Liverpool and Chelsea. That is the essence of the problem. The predictive rankings were a form guide. So in form boxers got rated higher. With the above example of totally crap results or outliers as we called them.

As far as I'm concerned, publish the damm predictive as well, the univers e is big enough for both of them. Just dont make them the real rankings next to a boxers name, it's embarrassing.

conan

Maybe a more refined metric to use as analysis would solve problems such as Briggs and Minto... but that's the beauty of that system, its flaws can be accepted as long as people understand the basis of that system. Since there's no Minto Briggs fight on the horizon.. its really a moot point. They could draw each other.

Should Roy Jones really be ahead of Zsolt Erdei?

Does Luis Collazo really not deserve a Top 40 ranking?

Harrison above JMM?

Definitely unacceptable considering a TRADITIONAL ranking..

Well there you go, hoist on your own petard. You're arguing things that are totally borderline i.e. Jones ahead of Erdei (based entirely on results its totally justifiable) and I'm arguing things like Shannon friggen Briggs and minto as top 12 and Stipe drews at #2 at light heavy that are just red hot stinking turds in the world of rankings.

If that's all you got to compare, then you got no argument.

conan

What's absurd in a "traditional system" isn't exactly comparable to what's absurd in a "predictive system"... at least in my opinion. If I get a traditional system thrown at me, I expect it to make perfect sense. The Ring, BoxingTalk, FightNews.. are all pretty much better.

[conan_the_cribber]

Maybe a more refined metric to use as analysis would solve problems such as Briggs and Minto... but that's the beauty of that system, its flaws can be accepted as long as people understand the basis of that system. Since there's no Minto Briggs fight on the horizon.. its really a moot point. They could draw each other.

Should Roy Jones really be ahead of Zsolt Erdei?

Does Luis Collazo really not deserve a Top 40 ranking?

Harrison above JMM?

Definitely unacceptable considering a TRADITIONAL ranking..

Well there you go, hoist on your own petard. You're arguing things that are totally borderline i.e. Jones ahead of Erdei (based entirely on results its totally justifiable) and I'm arguing things like Shannon friggen Briggs and minto as top 12 and Stipe drews at #2 at light heavy that are just red hot stinking turds in the world of rankings.

If that's all you got to compare, then you got no argument.

conan

What's absurd in a "traditional system" isn't exactly comparable to what's absurd in a "predictive system"... at least in my opinion. If I get a traditional system thrown at me, I expect it to make perfect sense. The Ring, BoxingTalk, FightNews.. are all pretty much better.

You are the only one calling it a traditional system. There is nothing 'traditional' about using a computer and solely the results (no interpretation) to rank fighters.

Your silly perfect sense expectation can never be fulfilled, because the Ring is done by people who, for instance, can a) look at Tarvers performance against Hopkins and think, that sucked so bad that he cant be number one, or b) RJJ's performance against Johnson and think, man is he shot. For the computer, its just a loss UD over 12 against a p4p top 5 fighter or a KO 8 loss against a p4p top 20 fighter.

What these rankings are, is an objective career performance rating i.e. what they achieved over their career. It's excellent at sorting out who's done and achieved the most.

As you never ever conceed or reflect on any point that I ever make, just rehash your own arguments, let me mention one thing again. You have the nerve to use a split-hair decicion like Jones before Erdei as a critism, yet your preferred alternative has Chad Dawon and Stipe Drews before both of these fighters. Go figure which is more reasonable.



conan

[JCS]

Well there you go, hoist on your own petard. You're arguing things that are totally borderline i.e. Jones ahead of Erdei (based entirely on results its totally justifiable) and I'm arguing things like Shannon friggen Briggs and minto as top 12 and Stipe drews at #2 at light heavy that are just red hot stinking turds in the world of rankings.

If that's all you got to compare, then you got no argument.

conan

What's absurd in a "traditional system" isn't exactly comparable to what's absurd in a "predictive system"... at least in my opinion. If I get a traditional system thrown at me, I expect it to make perfect sense. The Ring, BoxingTalk, FightNews.. are all pretty much better.

You are the only one calling it a traditional system. There is nothing 'traditional' about using a computer and solely the results (no interpretation) to rank fighters.

Your silly perfect sense expectation can never be fulfilled, because the Ring is done by people who, for instance, can a) look at Tarvers performance against Hopkins and think, that sucked so bad that he cant be number one, or b) RJJ's performance against Johnson and think, man is he shot. For the computer, its just a loss UD over 12 against a p4p top 5 fighter or a KO 8 loss against a p4p top 20 fighter.

What these rankings are, is an objective career performance rating i.e. what they achieved over their career. It's excellent at sorting out who's done and achieved the most.

As you never ever conceed or reflect on any point that I ever make, just rehash your own arguments, let me mention one thing again. You have the nerve to use a split-hair decicion like Jones before Erdei as a critism, yet your preferred alternative has Chad Dawon and Stipe Drews before both of these fighters. Go figure which is more reasonable.



conan

Stop focusing on semantics. Its a computerized system based on traditional principles, made to mock that of a traditional system such as "The Ring". Your point about the subtleties of realizing a shot fighter, etc, is absolutely correct and I believe that wholeheartedly highlights the problem of such a system where "winner beats loser = winner automatically above winner" and so on and so forth. Traditional principles and computerized system are like oil and water.

I think we'd be better off focusing on Top 25s for each division and nominating a group of members from the site to vote and set them on a weekly or bi-weekly basis.

At this point, I'd concede with some of the other posters who say we should just ditch the computerized system altogether and do as I said. Nominate the most educated posters on the site as a BoxRec governing body so to speak, who will be responsible of maintaining a Top 20, 30, 40, whatever for each division.

[conan_the_cribber]

Hi Cobwebcat,

you're right, I am being naughty. I am wasting my time talking to JCS, when I could be doing more purposeful things in my life. There was a time when I would take the trouble to break down sentence for sentence what JCS was saying, but that is indeed a waste of time, when he never follows a single line of argument.

Five posts ago, he's saying that we should have the predictive ratings again because it was the best and it was unique, then last post he's saying that we should elect an all-knowing team to do exactly what Ring and others do. Gee that first opinion lasted a long time. Why didn't it last, because his orginal idea was shot down with the Stipes/Dawson argument and he cant comment on it.

Creating performance based rankings is a worthwhile exercise, because you have all the data there and it makes the whole world independent of the WBC, WBA, WBO ratings, which are flawed by their ommissions. You cannot match people rankings, but you have many advantages too, a) there are regenerated daily, so they are always up to date b) they extend right through the division, so it is possible to compare a great range of fighters c) they are not biased because of race, nationality or access of a person to fights from that region. Honestly, who's the best between Matt Skelton, Cengis Koc and Tony Thompson? Three races, three different places, three careers at different points. No person can say something meaningful about that. At least boxrec can give you an opinion. And JCS you want to say that's not worthwhile? Or cause it is.

As far as I'm concerned, continue with the predictive rankings or in-form rankings, because they are interesting, even if they are odd. Just dont make them the default rankings, for the oft-repeated reasons.

And, as far as I'm concerned, get 5,10 or 20 regular posters, who a) see enough fights b) are prepared to spend a lot of time each month rating all 17 divisions, c) are not biased d) are smart and eloquent and polite enough to back up their decisions when queried and e) get some ratings process sorted out and off you go. Another worthwhile exercise. No problemo. I bet you a million dollars though, that if you find these 5 people, then they wont rank Stipe Freaken Drews at number 2 at Light Heavy.

Honestly Cobwebcat, I wouldn't yell, if the other side would listen or at least stick to one line of argument or argue a point out at a time.

conan

[Freiheit]

At this point, I'd concede with some of the other posters who say we should just ditch the computerized system altogether and do as I said. Nominate the most educated posters on the site as a BoxRec governing body so to speak, who will be responsible of maintaining a Top 20, 30, 40, whatever for each division.

I think that is a very bad idea. Its not the top 20-50 that is used in the boxing world. THe sanctionbodies are using ie top 100 to 300 to see if a fighter is qualified for a shot of one of their titles.

[conan_the_cribber]

Of course its a bad idea.

JCS. Come on now, Conan is right here. There's no need to throw the baby out with the bathwater.

If I could just make you both see you are both right but for different sections (Current and all-time)

Hi,

just to clarify something Cobwebcat, I'm not interested in the all-time, but from the little of what I've seen, for whatever reason the predictive algorithm produces better results for the all-time rankings.

In my opinion the p4p is better with the career performance ratings though.

conan

[JCS]

Hi Cobwebcat,

you're right, I am being naughty. I am wasting my time talking to JCS, when I could be doing more purposeful things in my life. There was a time when I would take the trouble to break down sentence for sentence what JCS was saying, but that is indeed a waste of time, when he never follows a single line of argument.

Five posts ago, he's saying that we should have the predictive ratings again because it was the best and it was unique, then last post he's saying that we should elect an all-knowing team to do exactly what Ring and others do. Gee that first opinion lasted a long time. Why didn't it last, because his orginal idea was shot down with the Stipes/Dawson argument and he cant comment on it.

Creating performance based rankings is a worthwhile exercise, because you have all the data there and it makes the whole world independent of the WBC, WBA, WBO ratings, which are flawed by their ommissions. You cannot match people rankings, but you have many advantages too, a) there are regenerated daily, so they are always up to date b) they extend right through the division, so it is possible to compare a great range of fighters c) they are not biased because of race, nationality or access of a person to fights from that region. Honestly, who's the best between Matt Skelton, Cengis Koc and Tony Thompson? Three races, three different places, three careers at different points. No person can say something meaningful about that. At least boxrec can give you an opinion. And JCS you want to say that's not worthwhile? Or cause it is.

As far as I'm concerned, continue with the predictive rankings or in-form rankings, because they are interesting, even if they are odd. Just dont make them the default rankings, for the oft-repeated reasons.

And, as far as I'm concerned, get 5,10 or 20 regular posters, who a) see enough fights b) are prepared to spend a lot of time each month rating all 17 divisions, c) are not biased d) are smart and eloquent and polite enough to back up their decisions when queried and e) get some ratings process sorted out and off you go. Another worthwhile exercise. No problemo. I bet you a million dollars though, that if you find these 5 people, then they wont rank Stipe Freaken Drews at number 2 at Light Heavy.

Honestly Cobwebcat, I wouldn't yell, if the other side would listen or at least stick to one line of argument or argue a point out at a time.

conan

You're too busy trying to come off witty and be a smartass anymore.. that's what's really disappointing. I'm not really trying to argue with you. I realize you came in only a little while ago and made some valuable points, but.. what gives man?

Do I think Stipe Drews should be ranked 2nd?? No I dont. Is that the answer you were looking for? That should've been a pretty obvious answer knowing how vast your intellect is. When considering a traditional ranking system, or one that is based on traditional princicples, do I consider Stipe Drews 2nd more unrealistic than Jones being above Erdei - sure I do... but these were experiences found on a performance based predictive system.. not one w/ traditional principles.

[generic screen name]

I like complaining about the old system more.

[emile]

Martin - welcome back, if you are back.

I have noticed a few strange aberrations in the rankings that maybe you can explain to me. I thought this system was using the rule that a fighter who beat another fighter became ranked above that fighter.

Satoshi Kogumazaka at #3 at Minimumweight. Last results a loss to #9 Takayama and a draw against #14 Misawa. Can't see how he'd be at 3 and above Takayama, who has not fought since the win.

Shane Swartz at #241 at Cruiserweight. Last result a TD win over #19 Dale Brown. Is a technical decision treated differently?

[JCS]

Martin - welcome back, if you are back.

I have noticed a few strange aberrations in the rankings that maybe you can explain to me. I thought this system was using the rule that a fighter who beat another fighter became ranked above that fighter.

Satoshi Kogumazaka at #3 at Minimumweight. Last results a loss to #9 Takayama and a draw against #14 Misawa. Can't see how he'd be at 3 and above Takayama, who has not fought since the win.

Shane Swartz at #241 at Cruiserweight. Last result a TD win over #19 Dale Brown. Is a technical decision treated differently?

TDs aren't being counted.

[emile]

Martin - welcome back, if you are back.

I have noticed a few strange aberrations in the rankings that maybe you can explain to me. I thought this system was using the rule that a fighter who beat another fighter became ranked above that fighter.

Satoshi Kogumazaka at #3 at Minimumweight. Last results a loss to #9 Takayama and a draw against #14 Misawa. Can't see how he'd be at 3 and above Takayama, who has not fought since the win.

Shane Swartz at #241 at Cruiserweight. Last result a TD win over #19 Dale Brown. Is a technical decision treated differently?

TDs aren't being counted.

Thanks. That seems illogical to me - was that a function of the predictive system?

Any ideas about the first case?

[JCS]

Martin - welcome back, if you are back.

I have noticed a few strange aberrations in the rankings that maybe you can explain to me. I thought this system was using the rule that a fighter who beat another fighter became ranked above that fighter.

Satoshi Kogumazaka at #3 at Minimumweight. Last results a loss to #9 Takayama and a draw against #14 Misawa. Can't see how he'd be at 3 and above Takayama, who has not fought since the win.

Shane Swartz at #241 at Cruiserweight. Last result a TD win over #19 Dale Brown. Is a technical decision treated differently?

TDs aren't being counted.

Thanks. That seems illogical to me - was that a function of the predictive system?

Any ideas about the first case?

TDs were not a function because factoring them in any form actually lowered prediction rates. It does make sense.. the fights are abridged, usually without the losing fighter's being at fault. If we had every scorecard for every TD, perhaps this would have helped..

In this system, I think TDs SHOULD be factored in.. no question about it.

Satoshi lost that fight by TD.. which isn't counted. He was most likely ranked above Takayama at that point.. Draws only change the ratings a small amount.

[emile]

TDs aren't being counted.

Thanks. That seems illogical to me - was that a function of the predictive system?

Any ideas about the first case?

TDs were not a function because factoring them in any form actually lowered prediction rates. It does make sense.. the fights are abridged, usually without the losing fighter's being at fault. If we had every scorecard for every TD, perhaps this would have helped..

In this system, I think TDs SHOULD be factored in.. no question about it.

Satoshi lost that fight by TD.. which isn't counted. He was most likely ranked above Takayama at that point.. Draws only change the ratings a small amount.

Ah thanks - slipped by me that Kogumazaka's loss was a TD as well.

[conan_the_cribber]

Hi Martin or JCS,

a small suggestion. It would be nice, if there was a small "Ratings last updated dd/mm/yy" note somewhere on the ratings page or at the top of each list, just to see how up to date they are.

cheers

conan

[conan_the_cribber]

TDs aren't being counted.

Thanks. That seems illogical to me - was that a function of the predictive system?

Any ideas about the first case?

TDs were not a function because factoring them in any form actually lowered prediction rates. It does make sense.. the fights are abridged, usually without the losing fighter's being at fault. If we had every scorecard for every TD, perhaps this would have helped..

In this system, I think TDs SHOULD be factored in.. no question about it.

Satoshi lost that fight by TD.. which isn't counted. He was most likely ranked above Takayama at that point.. Draws only change the ratings a small amount.

Good detective work. I agree with you JCS, TDs should be counted. A related question though, are technical draws excluded? They are definitely a meaningless result.

conan

[JCS]

C-Rank

What is the deal w/ TDs? Will they ever be included in the ratings?? I'd consider that a pretty big flaw considering the current rating system.

[computerrank]

Ratings refined, keeping the basics:

- TD (win/loss, not a draw) is rewarded now - as tradition rewuires - as a close decision. This makes the main impact.

- Decisions on points are rewarded in relation to scorecards, if available. So rewards may be near to nothing and up to full value.
- Rewards for decisions on points are in relation to rounds boxed - 12 rounds required for maximum reward. 12 round for full reward - 4 rounds only a third.

- Ratings now corrected for mean population of weight divisions.
- This has an impact ot p4p and all times.
- All Time Ratings now only consider the best 5 annual ratings in the weight division, the boxers is assigned to by the editors. So Sugar Ray Robinson only for middllweight, James Toney only for heavyweight ...

[computerrank]

... as he has only 1 annual rating in welterweight 2005, he is just #205 til now in welterweight.

In lower divisions his ranking would be much higher, but the editors have him at welterweight now - which is quite correct.

[conan_the_cribber]

Ratings refined, keeping the basics:

- Decisions on points are rewarded in relation to scorecards, if available. So rewards may be near to nothing and up to full value.
- Rewards for decisions on points are in relation to rounds boxed - 12 rounds required for maximum reward. 12 round for full reward - 4 rounds only a third. ...

Hi Martin,

can you post some more on the close points stuff.

I dont see much point reducing it for 4 rounders, as generally, they are not battling for many points anyhow. Conversely, it might have the advantage of highlighting 12 round fights, which are generally the crossroad fights.

thx

conan

conan

[computerrank]

Ratings refined, keeping the basics:

- Decisions on points are rewarded in relation to scorecards, if available. So rewards may be near to nothing and up to full value.
- Rewards for decisions on points are in relation to rounds boxed - 12 rounds required for maximum reward. 12 round for full reward - 4 rounds only a third. ...

Hi Martin,

can you post some more on the close points stuff.

I dont see much point reducing it for 4 rounders, as generally, they are not battling for many points anyhow. Conversely, it might have the advantage of highlighting 12 round fights, which are generally the crossroad fights.

thx

conan

conan

Please look at the ratings description first ...

Points are simply limited by proportion to rounds boxed (max 12 rounds) and mean score difference per judge (max 6 points).

Examples:

Full score for UD is 34.5% of opponents points = full reward.

- This is for a 12 rounder. For a 6 rounder only half the value can be achieved at maximum.

- This is for 6 points mean score differnece per judge. If the difference is only 4 points, only 2/3 is achieved.

So there is a considerable effect of low rounders and close decisions - see Mora at middleweight and a lot of Japanese boxers.

[computerrank]

A correction release is launched.

Changes:

1. Bug fix for decisions on points. Boxers could lose points in spite of a clear decision on points.

2. Dynamic equalization for P4P and All Time Ratings:

Ratings are equalized relating to rating level of average of #9,#10,#11 per year and division. The equalization is normalized to 500 points, but only down to a minimum of 150 net points.

[jujigatame]

First off, let me say I really like the new rankings, they are working out very well. I do have a question, though. How the hell did Derrick Gainer get his ranking? He washes out at 126, has 1 fight against a journeyman at 135 and is all of a sudden top 10? Something seems amiss there.

[computerrank]

He won until 2003 and was in the tops.

He only lost to John and Marquez - this cannot be turned against him as it
was quality opposition.

So:

- he was active inside the last 18 months - no penalty
- he boxed quality opposition inside the last 18 months - no penalty
- he had a SD against weak opposition last time - there was some penalty of 58 points for this
- he was top before

[jujigatame]

Maybe the penalty for inactivity should kick in after 12 or 15 months? He had 2 big layoffs (for 17 and 16 months, respectively) that it seems incorrect to entirely ignore.

[JCS]

Martin, I thought you acted on my recommendation to have the inactivity penalties hit at 15 months? It improved rates when I tested.. but that could be different now I suppose.