Ratings - please read before commenting - Archived

[JCS]

Wangek over Estrada is another weird one... it isn't like he's had some superb wins recently..

[computerrank]

Wangek over Estrada is another weird one... it isn't like he's had some superb wins recently..

You are right. A bug with Julian and Gregorian days. 13 days difference. So the restriction got lost 13 days too early. I will correct that. But Estrada's win will get void after a year anyway.

[DazBoxingFan]

Wangek over Estrada is another weird one... it isn't like he's had some superb wins recently..

I will look into that.

There are so many weird ones to be honest, hence why the previous ranking system was far better

[JCS]

Wangek over Estrada is another weird one... it isn't like he's had some superb wins recently..

You are right. A bug with Julian and Gregorian days. 13 days difference. So the restriction got lost 13 days too early. I will correct that. But Estrada's win will get void after a year anyway.

Only a year?

[computerrank]

You are right. A bug with Julian and Gregorian days. 13 days difference. So the restriction got lost 13 days too early. I will correct that. But Estrada's win will get void after a year anyway.

Only a year?

Why more? That override rule compensates in an upset situation. Such an upset demands for confirmation, not for long ongoing uncertainty.

[JCS]

Only a year?

Why more? That override rule compensates in an upset situation. Such an upset demands for confirmation, not for long ongoing uncertainty.

In MMA... a year has always felt a little short for this type of thing... I ended up using 540 days for a similar tactic.

[Rad#33]

Hello is there a way to look back at the ratings of any weight class for a particular week and/or month in the past, if not, then perhaps that is something that could be in the works going forward? it sure would be cool to be able to look at a fighters rating and their opponents rating compared to others in their weight class thank you

[marcianofan]

Why more? That override rule compensates in an upset situation. Such an upset demands for confirmation, not for long ongoing uncertainty.

In MMA... a year has always felt a little short for this type of thing... I ended up using 540 days for a similar tactic.

I don't know why it ought to be a time-related thing. I would suggest it ought to be based entirely on what the fighters do thereafter, when it comes to how long the winner stays on top. Let's say both fighters fight stay-busy opponents 6 months later and then don't fight again for the remainder of that year- the guy that won the fight ought still to be ranked ahead when the year expires. The upset victim lost. If he turns around and beats the #1 fighter 3 months later and the winner doesn't do anything worthwhile in that time, I think the loser of the upset could reasonably regain a higher rank despite the head-to-head loss. If neither do anything to distinguish themselves in any given time period, the winner should be on top indefinitely.

[computerrank]

In MMA... a year has always felt a little short for this type of thing... I ended up using 540 days for a similar tactic.

I don't know why it ought to be a time-related thing. I would suggest it ought to be based entirely on what the fighters do thereafter, when it comes to how long the winner stays on top. Let's say both fighters fight stay-busy opponents 6 months later and then don't fight again for the remainder of that year- the guy that won the fight ought still to be ranked ahead when the year expires. The upset victim lost. If he turns around and beats the #1 fighter 3 months later and the winner doesn't do anything worthwhile in that time, I think the loser of the upset could reasonably regain a higher rank despite the head-to-head loss. If neither do anything to distinguish themselves in any given time period, the winner should be on top indefinitely.

I think, there should be a maximum time for that. Upsetters tend to hide behind that upset otherwise. Maybe 18 months would be better.

[Manrae]

Have the annual ratings been updated? Or are we waiting on another rating database update?

[computerrank]

Have the annual ratings been updated? Or are we waiting on another rating database update?

I am busy with preparating additional features to the ratings.

[DazBoxingFan]

Have the annual ratings been updated? Or are we waiting on another rating database update?

I am busy with preparating additional features to the ratings.

What features and how will this effect the rankings? Hopefully better like the previous rankings and not for worse like the current ones

[computerrank]

I am busy with preparating additional features to the ratings.

What features and how will this effect the rankings? Hopefully better like the previous rankings and not for worse like the current ones

- no change for the ratings algorithm itself
- include separated amateur boxing career ratings
- what if match-up with rating values achievable after a bout

[DazBoxingFan]

What features and how will this effect the rankings? Hopefully better like the previous rankings and not for worse like the current ones

- no change for the ratings algorithm itself
- include separated amateur boxing career ratings
- what if match-up with rating values achievable after a bout

So as per usual you ignore everyone else’s criticism of the current ranking algorithm

[Manrae]

Have the annual ratings been updated? Or are we waiting on another rating database update?

I am busy with preparating additional features to the ratings.

cool

[DazBoxingFan]

So as per usual you ignore everyone else’s criticism of the current ranking algorithm

You know that’s unfair. If you want the system to reflect your personal preferences the only way to do that is to do your own ratings and no objective system is based on popularity.

Whatever anyone thinks of Computerrank he’s always open to debate and answers question which he isn’t under any obligation to do.

Look throughout this thread and everyone that’s disagreed strongly have either been ignored or “disappeared”/given up. He’s had countless criticism and doesn’t bend to any rules. Also from previous posts he doesn’t even have boxing experience or statistical experience so I’m unsure how this makes him qualified. I appreciate him taking time to contribute to this site however I don’t understand why he never takes criticism on board. His previous ranking system was much better and I don’t have any personal reasons I just sick of seeing fighters mis ranked so badly. But is what it is, computerrank will just continue to do his own thing

[JCS]

You know that’s unfair. If you want the system to reflect your personal preferences the only way to do that is to do your own ratings and no objective system is based on popularity.

Whatever anyone thinks of Computerrank he’s always open to debate and answers question which he isn’t under any obligation to do.

Look throughout this thread and everyone that’s disagreed strongly have either been ignored or “disappeared”/given up. He’s had countless criticism and doesn’t bend to any rules. Also from previous posts he doesn’t even have boxing experience or statistical experience so I’m unsure how this makes him qualified. I appreciate him taking time to contribute to this site however I don’t understand why he never takes criticism on board. His previous ranking system was much better and I don’t have any personal reasons I just sick of seeing fighters mis ranked so badly. But is what it is, computerrank will just continue to do his own thing

The tale of two extremes. Mr. Cobweb kissing butt and praying to the statistics.... and you're just being a tool.

I knew most would hate this new system and called it successfully. After all, Martin and I learned this ~15 years ago when we chased only prediction rate. He did the ratings back then... whereas I foolishly convinced him that prediction rate was the sole key to victory and assisted in the endeavor, only to have everyone hate it... and we eventually reverted back to more of a hybrid system that balanced prediction rate and traditional sensibility. Personally, I learned my lesson from that point on. Tradition and common sense must be observed as primary attributes, while prediction rate helps guide change.

[mike1989]

It turns out that more boxers in heavy weight categories have a higher rating, although they may have worse skills. The new ratings are the sibiosis of indicators from boxing bodybuilding and powerlifting))): who is bigger and stronger is better))).
In my opinion, the basis for the rating should be the dominance of the boxer. And weight categories should be compared in strength. And this indicator should be variable since in different years the strength of the weight categories in comparison is different.

The p4p ratings are quite artificial. So the absolute values could be converted, giving the lighter divisions a higher rating, without changing the ranking within the divisions.

Now it’s heavyweights that lightweights carry out the same number of fights, because this confirmation is not objective ???

[computerrank]

Look throughout this thread and everyone that’s disagreed strongly have either been ignored or “disappeared”/given up. He’s had countless criticism and doesn’t bend to any rules. Also from previous posts he doesn’t even have boxing experience or statistical experience so I’m unsure how this makes him qualified. I appreciate him taking time to contribute to this site however I don’t understand why he never takes criticism on board. His previous ranking system was much better and I don’t have any personal reasons I just sick of seeing fighters mis ranked so badly. But is what it is, computerrank will just continue to do his own thing

The tale of two extremes. Mr. Cobweb kissing butt and praying to the statistics.... and you're just being a tool.

I knew most would hate this new system and called it successfully. After all, Martin and I learned this ~15 years ago when we chased only prediction rate. He did the ratings back then... whereas I foolishly convinced him that prediction rate was the sole key to victory and assisted in the endeavor, only to have everyone hate it... and we eventually reverted back to more of a hybrid system that balanced prediction rate and traditional sensibility. Personally, I learned my lesson from that point on. Tradition and common sense must be observed as primary attributes, while prediction rate helps guide change.

from my experience doing ratings is balancing within:
- learning from the past by achieving best winner prediction for past bouts, with only using pre-bout results - this gives an objective measure
- learning from the past by achieving best correlation between after-bout ratings and results - this gives another objective measure
- keeping things as simple as possible and not add endlless refinements - this lets me always look for simpler and better solutions
- obeying traditional rules like the winner is ranked above the loser after the bout

The innovation with WHR ratings is:
- all Boxrec ratings before rewarded a bout only on base of information before the bout. Even for past bouts. So information from later bouts of the opponents was simply neglected. This systematically ingnored valuable information, every human rater would intuitively use.
- now this information can flow back and forth between all bouts and ratings connected
- in spite of that the basic WHR ratings algorithm is very simple - it uses a single parameter, how fast a rating can change over time.

[JCS]

The tale of two extremes. Mr. Cobweb kissing butt and praying to the statistics.... and you're just being a tool.

I knew most would hate this new system and called it successfully. After all, Martin and I learned this ~15 years ago when we chased only prediction rate. He did the ratings back then... whereas I foolishly convinced him that prediction rate was the sole key to victory and assisted in the endeavor, only to have everyone hate it... and we eventually reverted back to more of a hybrid system that balanced prediction rate and traditional sensibility. Personally, I learned my lesson from that point on. Tradition and common sense must be observed as primary attributes, while prediction rate helps guide change.

Quite interested in which bit is kissing butt. Do you not agree that Computerrank is reasonable and answers questions? Quite surprised at that.

As for calling it correctly that the overwhelming majority wouldn’t like the change well I think we pretty much all said that so I don’t think you need award yourself a trophy just yet.

I've been discussing this topic with Martin for over 15 years now. Where did I say he wasn't reasonable? If he wasn't, I would've stopped a long time before now. Afraid, you're getting me mixed up w/ the guy who keeps talking trying to rev him up.

[DazBoxingFan]

I’m
Not revving him up and I hardly see a guy that doesn’t take everyone’s opinion on the new ranking system being worse than the one before and still keeping it the same whilst pretty much ignoring them as “reasonable”

[JCS]

I’m
Not revving him up and I hardly see a guy that doesn’t take everyone’s opinion on the new ranking system being worse than the one before and still keeping it the same whilst pretty much ignoring them as “reasonable”

If you try to please everyone, you'll please no one. Everybody has their own opinions...and I imagine ours aren't worth as much as the site owner, ya know?

[JCS]

I've been discussing this topic with Martin for over 15 years now. Where did I say he wasn't reasonable? If he wasn't, I would've stopped a long time before now. Afraid, you're getting me mixed up w/ the guy who keeps talking trying to rev him up.

I’m just trying to understand your bottom kissing statement. If we both agree CR is reasonable then either you are mistaken or we are both having a snog of each cheek are we not?

Your cheerleading for the newest changes seems to be a bit much. Maybe I should've used a different term to explain it.

[Manrae]

Hey, I have a question. Might sound broad but, basically... are the ratings of today's boxers inflated because we don't yet have the data of what they might become in the future?

I just wonder why today's boxers' ratings are so high compared to almost all other eras. Especially when top fighters rarely face each other these days...

[computerrank]

Hey, I have a question. Might sound broad but, basically... are the ratings of today's boxers inflated because we don't yet have the data of what they might become in the future?

I just wonder why today's boxers' ratings are so high compared to almost all other eras. Especially when top fighters rarely face each other these days...

On Nov 2, 2019 I posted a test regarding inflation in this thread:
viewtopic.php?p=5264409#p5264409

I switched Wladimir Klitschko's career into the time period of Joe Louis and Sam Langford - and found, Klitschko with his results would have been superior even at those times.

Another result was, top current boxers' on top of their career currently may decrease in future while descending in performance. So I reduce their current career top rating based all time ratings. For all boxers with career top rating before 2005 there is no reduction.