Ratings - please read before commenting - Archived

computerrank · Post by **computerrank** » 18 May 2006, 12:49

Lennox wrote:
computerrank wrote:
Lennox wrote: Crap ratings and exactly how not to prodice a computerized rating system.
http://www.spam.com
Beat them ...
I think you have beaten yourselves.

exactly ...

JCS · Post by **JCS** » 18 May 2006, 12:49

Lennox wrote:
computerrank wrote:
Lennox wrote: Crap ratings and exactly how not to prodice a computerized rating system.
http://www.spam.com
Beat them ...
I think you have beaten yourselves.

Nothing is more fun!

Lennox · Post by **Lennox** » 18 May 2006, 13:14

Well I am still not convinced your new ratings will produce a better prediction rate, looking at heavyweight, you have Matt Skelton higher than Danny Williams, Danny just won and there is a rematch soon that seems strange your method says Skelton would be favoured, also Tye Fields at 9 beats Hashim Rahman, come on I know in boxing anything can happen but Rahman would be a 10-1 on favourite IMO. the new WBO champ low 20s??? I think JohnShep will scrub these ratings pronto.

JCS · Post by **JCS** » 18 May 2006, 13:18

Lennox wrote:Well I am still not convinced your new ratings will produce a better prediction rate, looking at heavyweight, you have Matt Skelton higher than Danny Williams, Danny just won and there is a rematch soon that seems strange your method says Skelton would be favoured, also Tye Fields at 9 beats Hashim Rahman, come on I know in boxing anything can happen but Rahman would be a 10-1 on favourite IMO. the new WBO champ low 20s??? I think JohnShep will scrub these ratings pronto.

Scrub these ratings?? He has final say.

Grey · Post by **Grey** » 18 May 2006, 13:36

Personally, I think the original boxrec rating system was much better.

NickR · Post by **NickR** » 18 May 2006, 13:45

Agreed

JCS · Post by **JCS** » 18 May 2006, 13:46

These new ratings aim for objectivity though.. really all you can expect out of a computer. So, you must realize their basis is completely different from the previous systems.

If you want subjective ratings... go read Dan Rafael's ratings, or one of the ABC organizations.

Lennox · Post by **Lennox** » 18 May 2006, 13:58

JCS83MD wrote:These new ratings aim for objectivity though.. really all you can expect out of a computer. So, you must realize their basis is completely different from the previous systems.

If you want subjective ratings... go read Dan Rafael's ratings, or one of the ABC organizations.

Why are you convinced that the IBO ratings are not objective, the only thing I noticed was that Kirk Johnson was still still ahead of Mora, if it was a bit bent I would have thought they would have their champions higher, the only one that looks high is Joval. Their website claim to be the only sanction body using computer ratings.

JCS · Post by **JCS** » 18 May 2006, 14:19

Lennox wrote:
JCS83MD wrote:These new ratings aim for objectivity though.. really all you can expect out of a computer. So, you must realize their basis is completely different from the previous systems.

If you want subjective ratings... go read Dan Rafael's ratings, or one of the ABC organizations.
Why are you convinced that the IBO ratings are not objective, the only thing I noticed was that Kirk Johnson was still still ahead of Mora, if it was a bit bent I would have thought they would have their champions higher, the only one that looks high is Joval. Their website claim to be the only sanction body using computer ratings.

They are not objective. Trust me.. they wouldn't have a leg to stand on because they look like the system here which is not suitable for a sanctioning body.

Again, BoxRec is not trying to be a sanctioning body.

mattyp151 · Post by **mattyp151** » 18 May 2006, 15:10

This system is terrible...the other had it's flaws every now and again, but there are glaring problems now...mainly Calzaghe, Cotto, Hatton, and Abraham 2-5, and Manny Pac sitting at 20...ridiculous.

Grey · Post by **Grey** » 18 May 2006, 16:06

JCS83MD wrote:These new ratings aim for objectivity though.. really all you can expect out of a computer. So, you must realize their basis is completely different from the previous systems.

If you want subjective ratings... go read Dan Rafael's ratings, or one of the ABC organizations.

I didn't say I wanted subjective ratings.

The old boxrec rating system was objective as well. It simply had a better formula than this new system.

JCS · Post by **JCS** » 18 May 2006, 16:07

GreyMinus wrote:
JCS83MD wrote:These new ratings aim for objectivity though.. really all you can expect out of a computer. So, you must realize their basis is completely different from the previous systems.

If you want subjective ratings... go read Dan Rafael's ratings, or one of the ABC organizations.
I didn't say I wanted subjective ratings.

The old boxrec rating system was objective as well. It simply had a better formula than this new system.

It was objective, but fighters were capped for purpose of "looks". This system is purely objective with statistics far better than the previous system.

BoxBuzz · Post by **BoxBuzz** » 18 May 2006, 16:08

so who makes the calls? Are you one of the decision makers? Or is it more of an objective software based formula than can no longer be tweeked?

JCS · Post by **JCS** » 18 May 2006, 16:10

BoxBuzz wrote:so who makes the calls? Are you one of the decision makers? Or is it more of an objective software based formula than can no longer be tweeked?

I don't think there really are anymore decisions except what the "prediction rate" is for fights after 1/1/00. JohnShep apparently made the decision to convert to totally objective ratings since thats all we can hope for from a computer.

Grey · Post by **Grey** » 18 May 2006, 16:10

Either way, I can't see how one could defend Alexander Dimitrenko being ranked above James Toney using any kind of formula.

BoxBuzz · Post by **BoxBuzz** » 18 May 2006, 16:12

the "blindness" of such objectivity is sort of a painful nuisance. All these ratings can do is give you something to launch a meaningful subjective discussion from.

But in that regard it is a great foundation. The scenario Grey talks about almost looks like a glitch...is that a possibility?

JCS · Post by **JCS** » 18 May 2006, 16:23

BoxBuzz wrote:the "blindness" of such objectivity is sort of a painful nuisance. All these ratings can do is give you something to launch a meaningful subjective discussion from.

But in that regard it is a great foundation. The scenario Grey talks about almost looks like a glitch...is that a possibility?

A glitch?? Probably not... A circumstance not considered or dealt with properly.. Of course it could be.

How good is Toney at Heavy though?

Grey · Post by **Grey** » 18 May 2006, 16:24

JCS83MD wrote:
BoxBuzz wrote:the "blindness" of such objectivity is sort of a painful nuisance. All these ratings can do is give you something to launch a meaningful subjective discussion from.

But in that regard it is a great foundation. The scenario Grey talks about almost looks like a glitch...is that a possibility?
A glitch?? Probably not... A circumstance not considered or dealt with properly.. Of course it could be.

How good is Toney at Heavy though?

He's accomplished more than ALEXANDER DIMITRENKO.

JCS · Post by **JCS** » 18 May 2006, 16:26

GreyMinus wrote:
JCS83MD wrote:
BoxBuzz wrote:the "blindness" of such objectivity is sort of a painful nuisance. All these ratings can do is give you something to launch a meaningful subjective discussion from.

But in that regard it is a great foundation. The scenario Grey talks about almost looks like a glitch...is that a possibility?
A glitch?? Probably not... A circumstance not considered or dealt with properly.. Of course it could be.

How good is Toney at Heavy though?
He's accomplished more than ALEXANDER DIMITRENKO.

Sure he has, but who's better at HW.. right this second and can you say for sure? Remember, Dimitrenko has fought most of his fights on an extremely biased home turf thus elevating his rating (probably)... The rating system can't really take that into account.

BoxBuzz · Post by **BoxBuzz** » 18 May 2006, 16:29

Never mind

computerrank · Post by **computerrank** » 18 May 2006, 16:37

GreyMinus wrote:Either way, I can't see how one could defend Alexander Dimitrenko being ranked above James Toney using any kind of formula.

... just because the new ratings based on the new rules, which now rate Dimitrenko's results higher than Toney's, have more winning boxers correctly rated higher before the bouts than losing boxers - much more than any rules used before ...

The ratings simply yield the highest overall probability for the truth of the statement, a boxer rated higher will defeat a boxer rated lower.

A rating which show best conformance with this reality should be regarded the best rating, what else.

It seems, that common rankings and their rules keep old merits too long too high, longer as some indicators say, which BoxRec rules now use in order to better predict the real outcomes.

jujigatame · Post by **jujigatame** » 18 May 2006, 16:40

A rating which show best conformance with this reality should be regarded the best rating, what else.

Why?

Seriously, there is a serious logical breach that you are ignoring. You claim the best rankings are ones that "predict the real outcomes" but why is that such a given? It sounds like you're creating a betting system rather than a logical set of boxing rankings.

JCS · Post by **JCS** » 18 May 2006, 16:50

jujigatame wrote:
A rating which show best conformance with this reality should be regarded the best rating, what else.
Why?

Seriously, there is a serious logical breach that you are ignoring. You claim the best rankings are ones that "predict the real outcomes" but why is that such a given? It sounds like you're creating a betting system rather than a logical set of boxing rankings.

But what he's trying to say is we're aiming for a system that speaks the truth in most circumstances for who will win what bout, etc. If subjectivity is desired, its obvious the system here never was the answer. I am with the audience in saying that wow, these ratings look really fucked up.. but this time, there is noone to argue with.. all we can hope for is a system with a better prediction rate in the future. And who knows, that may look even more strange!!

JCS · Post by **JCS** » 18 May 2006, 16:51

I think there is some room for improvement here.. I think a few guys are moving up too fast because of a ton of activity... but maybe this is correct?

computerrank · Post by **computerrank** » 18 May 2006, 16:53

jujigatame wrote:
A rating which show best conformance with this reality should be regarded the best rating, what else.
Why?

Seriously, there is a serious logical breach that you are ignoring. You claim the best rankings are ones that "predict the real outcomes" but why is that such a given? It sounds like you're creating a betting system rather than a logical set of boxing rankings.

... your logic means rating some boxers higher based on "achievements", but "knowing" they would be defeated by some lower rated ones ...