Ratings - please read before commenting - Archived

[DazBoxingFan]

Computerrank why did you suspend me? I just disagreed (as have many) to your new ranking system. Not listening to your users on here is a disgrace. You have no reasoning behind your ranking systems, the best ranking system you’ve ever had was the one before this, in fact you had few complaints and then boredom hit and you totally changed it. Like I said, your pound 4 pound ranking is hilarious. Rankings shouldn’t be based on trying to improve prediction ratings, sounds nonsense. That’s the beautiful thing about boxing anything can happen and it’s unpredictable but there should be a ranking system based on better opposition. From previous posts on here - anyone who disagrees with you you either ignore or suspend. It’s like a dictatorship. I appreciate the work you do but that’s not the way to go about it

[margaret thatcher]

Computerrank why did you suspend me? I just disagreed (as have many) to your new ranking system. Not listening to your users on here is a disgrace. You have no reasoning behind your ranking systems, the best ranking system you’ve ever had was the one before this, in fact you had few complaints and then boredom hit and you totally changed it. Like I said, your pound 4 pound ranking is hilarious. Rankings shouldn’t be based on trying to improve prediction ratings, sounds nonsense. That’s the beautiful thing about boxing anything can happen and it’s unpredictable but there should be a ranking system based on better opposition. From previous posts on here - anyone who disagrees with you you either ignore or suspend. It’s like a dictatorship. I appreciate the work you do but that’s not the way to go about it

How do you know he suspended you?

[computerrank]

Computerrank why did you suspend me? I just disagreed (as have many) to your new ranking system. Not listening to your users on here is a disgrace. You have no reasoning behind your ranking systems, the best ranking system you’ve ever had was the one before this, in fact you had few complaints and then boredom hit and you totally changed it. Like I said, your pound 4 pound ranking is hilarious. Rankings shouldn’t be based on trying to improve prediction ratings, sounds nonsense. That’s the beautiful thing about boxing anything can happen and it’s unpredictable but there should be a ranking system based on better opposition. From previous posts on here - anyone who disagrees with you you either ignore or suspend. It’s like a dictatorship. I appreciate the work you do but that’s not the way to go about it

I cannot suspend someone.
Nobody responded on my last post regarding current p4p list. viewtopic.php?p=5350559#p5350559

[DazBoxingFan]

My mistake, I thought you did. Either way I’m still unsure as to why I was suspended. The ranking is still poor, I mean... josh Warrington at number 9.. really? I just think there are many strange changes in the rankings. Even looking at the previous post mentioning Alex Dilmaghani slipping from 4 to 7, no way should be below Bowen or sharp. I do see how you are trying to improve and perfect as much as possible and you contribute a lot to this site. But I just think your previous ranking system was much better as many have said, just my opinion. And when you ignore many peoples responses or agreement it comes across condescending as if you have superior knowledge. I think this current system is over complicated but no problem in the light of current situations it’s hardly important.

[marcianofan]

... something that keeps being avoided is the question of how it takes into account A-side advantage.

This is just one guy's opinion, but I don't think that's something that can ever be reasonably applied to a computer model. Too many variables involved in what actually causes A-side advantage and how much it affects or doesn't affect individual fights. If the idea is to predict, then really any actual effect being the A-side might cause would be baked into the rating, and would help the prediction.

Plus, bad officiating comes in all shapes and sizes, and sometimes when you least expect it. Unless you're going to go in and plug your own opinions on individual scorecards in to the model, I think any attempt to control for that would just make things more confused, and more controversial.

Personally, that's why a "what-if" mode that allows you to replace fight results with your own opinion of how the fight should have turned out (or to run an alternate history scenario) is on my wishlist. This functionality clearly exists behind the scenes, since computerrank was able to give us exact rankings on if Fury-Wilder I had been a NWS win for Fury instead of a Draw. Allowing us to mess with it from our individual logins would be rad. And I don't use that word lightly.

Aside from that, I think you have to stick with the official rulings, since they are the only objective way to measure results, unless of course crowd scores were factored in. But its not like a fighter's popularity or stature can't influence that heavily, as well. And it would also allow random members to influence the rankings to an undue degree, presumably, if we score obscure fights differently than the original judges, since probably not a lot of averaging is going to happen with other votes.

[computerrank]

Looking at the bout rankings in the record, you will see the undisturbed WHR ratings. In case of an upset, the losers WHR rating may be higher than the winner's.
Looking at the division ratings, the current ratings will reflect the rule winner above loser.

When the time comes to publish both rankings wouldn’t it be reasonable to leave the WHR without the winner rule given that the old restored one will retain it?

If it stays as it is there will be nowhere to see the full WHR rankings without the rule.

I understand the site owners wanting the rule when there is only one set published.

This an option for the time coming. But no schedule in place.

[computerrank]

...
Personally, that's why a "what-if" mode that allows you to replace fight results with your own opinion of how the fight should have turned out (or to run an alternate history scenario) is on my wishlist. This functionality clearly exists behind the scenes, since computerrank was able to give us exact rankings on if Fury-Wilder I had been a NWS win for Fury instead of a Draw. Allowing us to mess with it from our individual logins would be rad. And I don't use that word lightly.
..

Such a feature is in now with the current ratings. Boxrec will use it to update the ratings instantly with each new or changed bout result.
This feature also allows a what-if function to study the effect of changed bout results or of an additional bout instantly.
But finally it is a question of resources and impact. So Boxrec will decide that at the time given.

[Finkel]

... something that keeps being avoided is the question of how it takes into account A-side advantage.

This is just one guy's opinion, but I don't think that's something that can ever be reasonably applied to a computer model. Too many variables involved in what actually causes A-side advantage and how much it affects or doesn't affect individual fights. If the idea is to predict, then really any actual effect being the A-side might cause would be baked into the rating, and would help the prediction.

Plus, bad officiating comes in all shapes and sizes, and sometimes when you least expect it. Unless you're going to go in and plug your own opinions on individual scorecards in to the model, I think any attempt to control for that would just make things more confused, and more controversial.

Personally, that's why a "what-if" mode that allows you to replace fight results with your own opinion of how the fight should have turned out (or to run an alternate history scenario) is on my wishlist. This functionality clearly exists behind the scenes, since computerrank was able to give us exact rankings on if Fury-Wilder I had been a NWS win for Fury instead of a Draw. Allowing us to mess with it from our individual logins would be rad. And I don't use that word lightly.

Aside from that, I think you have to stick with the official rulings, since they are the only objective way to measure results, unless of course crowd scores were factored in. But its not like a fighter's popularity or stature can't influence that heavily, as well. And it would also allow random members to influence the rankings to an undue degree, presumably, if we score obscure fights differently than the original judges, since probably not a lot of averaging is going to happen with other votes.

Thanks for the considered response. But that is actually my point. The A-side advantage cannot be calculated by the model, but it is a well known factor in boxing. The A-side or defending champion will often get preferential treatment by judges, and well known to have a big impact on the outcome of some very high profile fights. This shouldn't come as a shock to anyone on this forum.

I read the research paper ComputerRank provided. The WHR was tested and modelled using the game Go. That is a game strictly confined by its rules. It is not a game where the outcome is subjective. Sure there will be some variable factors at play besides skill such as whether a player can handle the pressure at the top level but they would naturally be including within the rating.

But if there is no way to account for A-side advantage in the prediction model, then it ignores a big part of what is often at play in professional boxing. And means any prediction is going to be flawed.

For example in the case of a champion whose results were heavily reliant on being the A-side, if he no longer holds a belt, then the system needs to be able to account for the disappearance of that A-side advantage. If WHR cannot model for this, then it is going to inflate the rankings of fighters who have just lost their belt.

Too give you an example,

Imagine there is an examiner testing writing proficiency. Unfortunately said examiner is deaf. During the written exam a fire alarm goes off. Wailing away for 30 minutes. The examiner notes the student appears agitated, but just assumes he hasn't prepared well enough for the test. At the end when the results came in, it turned out the student didn't perform as well as the other students in a different location.

WHR is the deaf examiner.
A-side advantage is the screaming alarm.




Edit:. Anyway, I suspect I'm beating a dead horse on this thread, so I'll just stop

Cheers

[DazBoxingFan]

This is just one guy's opinion, but I don't think that's something that can ever be reasonably applied to a computer model. Too many variables involved in what actually causes A-side advantage and how much it affects or doesn't affect individual fights. If the idea is to predict, then really any actual effect being the A-side might cause would be baked into the rating, and would help the prediction.

Plus, bad officiating comes in all shapes and sizes, and sometimes when you least expect it. Unless you're going to go in and plug your own opinions on individual scorecards in to the model, I think any attempt to control for that would just make things more confused, and more controversial.

Personally, that's why a "what-if" mode that allows you to replace fight results with your own opinion of how the fight should have turned out (or to run an alternate history scenario) is on my wishlist. This functionality clearly exists behind the scenes, since computerrank was able to give us exact rankings on if Fury-Wilder I had been a NWS win for Fury instead of a Draw. Allowing us to mess with it from our individual logins would be rad. And I don't use that word lightly.

Aside from that, I think you have to stick with the official rulings, since they are the only objective way to measure results, unless of course crowd scores were factored in. But its not like a fighter's popularity or stature can't influence that heavily, as well. And it would also allow random members to influence the rankings to an undue degree, presumably, if we score obscure fights differently than the original judges, since probably not a lot of averaging is going to happen with other votes.

Thanks for the considered response. But that is actually my point. The A-side advantage cannot be calculated by the model, but it is a well known factor in boxing. The A-side or defending champion will often get preferential treatment by judges, and well known to have a big impact on the outcome of some very high profile fights. This shouldn't come as a shock to anyone on this forum.

I read the research paper ComputerRank provided. The WHR was tested and modelled using the game Go. That is a game strictly confined by its rules. It is not a game where the outcome is subjective. Sure there will be some variable factors at play besides skill such as whether a player can handle the pressure at the top level but they would naturally be including within the rating.

But if there is no way to account for A-side advantage in the prediction model, then it ignores a big part of what is often at play in professional boxing. And means any prediction is going to be flawed.

For example in the case of a champion whose results were heavily reliant on being the A-side, if he no longer holds a belt, then the system needs to be able to account for the disappearance of that A-side advantage. If WHR cannot model for this, then it is going to inflate the rankings of fighters who have just lost their belt.

Too give you an example,

Imagine there is an examiner testing writing proficiency. Unfortunately said examiner is deaf. During the written exam a fire alarm goes off. Wailing away for 30 minutes. The examiner notes the student appears agitated, but just assumes he hasn't prepared well enough for the test. At the end when the results came in, it turned out the student didn't perform as well as the other students in a different location.

WHR is the deaf examiner.
A-side advantage is the screaming alarm.




Edit:. Anyway, I suspect I'm beating a dead horse on this thread, so I'll just stop

Cheers

To be honest it appears that way with everyone who disagrees, falls on deaf ears

[computerrank]

Thanks for the considered response. But that is actually my point. The A-side advantage cannot be calculated by the model, but it is a well known factor in boxing. The A-side or defending champion will often get preferential treatment by judges, and well known to have a big impact on the outcome of some very high profile fights. This shouldn't come as a shock to anyone on this forum.

I read the research paper ComputerRank provided. The WHR was tested and modelled using the game Go. That is a game strictly confined by its rules. It is not a game where the outcome is subjective. Sure there will be some variable factors at play besides skill such as whether a player can handle the pressure at the top level but they would naturally be including within the rating.

But if there is no way to account for A-side advantage in the prediction model, then it ignores a big part of what is often at play in professional boxing. And means any prediction is going to be flawed.

For example in the case of a champion whose results were heavily reliant on being the A-side, if he no longer holds a belt, then the system needs to be able to account for the disappearance of that A-side advantage. If WHR cannot model for this, then it is going to inflate the rankings of fighters who have just lost their belt.

Too give you an example,

Imagine there is an examiner testing writing proficiency. Unfortunately said examiner is deaf. During the written exam a fire alarm goes off. Wailing away for 30 minutes. The examiner notes the student appears agitated, but just assumes he hasn't prepared well enough for the test. At the end when the results came in, it turned out the student didn't perform as well as the other students in a different location.

WHR is the deaf examiner.
A-side advantage is the screaming alarm.




Edit:. Anyway, I suspect I'm beating a dead horse on this thread, so I'll just stop

Cheers

To be honest it appears that way with everyone who disagrees, falls on deaf ears

WHR ratings can be adapted to a home advantage. Even the old rating could be adapted to such. And I already did this about 10 years ago. E.g. it was connected with the boxer promoter relation. So a boxer linked to a promoter by more than a minimum number of bouts, statistically significant had an advantage, when this promoter made the event.

But although this advantage was significant, It would have been unfair to a lot of boxers, to apply this advantage in the ratings, because the advantage was assumed to be very different for different boxers and promoters.

And the same would be with other criteria, such as to be a title holder.

So to make it clear, the reason for using best winner prediction as criteria for the ratings quality is not to win bets. Best winner prediction is just an objective criteria for having the boxers rated best before the bout.

[fraac]

Hi. These are probably explained somewhere but I can't find it:

1. Why are ratings on a boxer's page now different from ratings on the rankings page? What's the relationship?

2. Are you now including amateur records in ratings? I see a lot of top pros rating between 5 and 10 on their debut.

Cheers.

[Rocky-]

Japanese tend to have been the most overrated by BoxRec imo, at least before. Something about the ratings led to them getting too many points for their domestic Japanese level fights. You'd get Japanese guys up at 160 and 168 getting it too, I think the guy Murata debuted against was ranked top 20 at the time. at least under the old rankings, but in reality he was nowhere near that

I agree to an extent. There are so few Japanese middleweights, they mostly fight for the Japanese or OPBF belts against fellow countrymen, up-and-coming Aussies, and Thai guys with heavily padded records, so good luck accurately gauging their ability. Top Nippon 160 guys of all time would be the triumvirate of Ryota Murata, Shinji Takehara & Nobuhiro Ishida… that's how thin their ranks are.

Easiest way to rapidly shoot up the Boxrec rankings right now at 160 is to take out Shinobu Charlie Hosokawa (3 stars), Kazuto Takesako (4 stars), Yuki Nonaka (42 year old dude worth 3 stars). The version of Kassim Ouma that fought GGG would wipe these guys out however Boxrec rates him as merely a 2 star fighter.

[Rad#33]

hello, this is a little bit off the subject of "ratings" but didn't there used to be a feature that allowed you to plug in two fighters and and get a percentage prediction, because I just now looked for that and cant seem to find it, that was a pretty cool feature to mess around with

[computerrank]

Hi. These are probably explained somewhere but I can't find it:

1. Why are ratings on a boxer's page now different from ratings on the rankings page? What's the relationship?

2. Are you now including amateur records in ratings? I see a lot of top pros rating between 5 and 10 on their debut.

Cheers.

ad 1:
The ratings on the ratings page are converted to the boxer's weight division, Lower divisions are displayed with more points, The factor is in relation to the third power of the division reference weight ratio. E.g. Heavyweight has factor 1; Cruiserweight has (240/200)*(240/200)*(240/200) = 1.2*1.2*1.2 = 1.73; Light Heavyweight has (240/175)*(240/175)*(240/175)= 1.37*1.37*1.37 = 2.57 etc.
In an upcoming version ratings on the boxer's page will be displayed converted to the bout division.

ad2:
Bouts from other sports are not included in the ratings of a boxer, A boxer has his own rating for each sport. Top pros may have a higher start rating due to better success (results and opponents) in the start phase of their career, or having already shown top success.
As the the ratings are connected back and forth in history now, the start rating may move with later success.

[computerrank]

Well looking there, it seems that Fury and Povetkin both got substantial improvements in this system, so it follows that Wilder and Hunter would since they drew with them. Not sure it's some systematically pro USA thing

Japanese tend to have been the most overrated by BoxRec imo, at least before. Something about the ratings led to them getting too many points for their domestic Japanese level fights. You'd get Japanese guys up at 160 and 168 getting it too, I think the guy Murata debuted against was ranked top 20 at the time. at least under the old rankings, but in reality he was nowhere near that

I looked into the new and into the old ratings:
- in the old ratings Japan has 134 boxers within the current top 1000
- in the new ratings Japan has 89 boxers within the current top 1000

old/new
Murata #8/#8
Takesako #13/#19
Nonaka #30/#61
Hosakawa #45/#65

[DazBoxingFan]

@computerrank the current ranking system is awful as many have stated previously. The old one needs to come back but yet seemingly everyone who totally disagrees with you goes “missing” and blocked

[margaret thatcher]

Well looking there, it seems that Fury and Povetkin both got substantial improvements in this system, so it follows that Wilder and Hunter would since they drew with them. Not sure it's some systematically pro USA thing

Japanese tend to have been the most overrated by BoxRec imo, at least before. Something about the ratings led to them getting too many points for their domestic Japanese level fights. You'd get Japanese guys up at 160 and 168 getting it too, I think the guy Murata debuted against was ranked top 20 at the time. at least under the old rankings, but in reality he was nowhere near that

I looked into the new and into the old ratings:
- in the old ratings Japan has 134 boxers within the current top 1000
- in the new ratings Japan has 89 boxers within the current top 1000

old/new
Murata #8/#8
Takesako #13/#19
Nonaka #30/#61
Hosakawa #45/#65

[marcianofan]

@computerrank the current ranking system is awful as many have stated previously. The old one needs to come back but yet seemingly everyone who totally disagrees with you goes “missing” and blocked

For what it's worth, I've been about as critical as anybody, and I haven't been sanctioned or gone missing, and I even usually get a response. I still think it's crazy and counterproductive to have this system as the only public-facing ratings available to the users as is currently the case, but I can't say my opinion, at least, has been treated any worse than being substantively dismissed and occasionally ignored. That's not great, but I at least can't lend any testimony in support the censorship narrative.

[marcianofan]

For example in the case of a champion whose results were heavily reliant on being the A-side, if he no longer holds a belt, then the system needs to be able to account for the disappearance of that A-side advantage. If WHR cannot model for this, then it is going to inflate the rankings of fighters who have just lost their belt.

Too give you an example,

Imagine there is an examiner testing writing proficiency. Unfortunately said examiner is deaf. During the written exam a fire alarm goes off. Wailing away for 30 minutes. The examiner notes the student appears agitated, but just assumes he hasn't prepared well enough for the test. At the end when the results came in, it turned out the student didn't perform as well as the other students in a different location.

WHR is the deaf examiner.
A-side advantage is the screaming alarm.




Edit:. Anyway, I suspect I'm beating a dead horse on this thread, so I'll just stop

Cheers

Yeah I understand your point. But I have to agree with computerrank on this one. It's a real thing, but I think it's too uneven to be fairly adjusted for. Take Felix Sturm, for example. He gets 3 or 4 lucky or questionable decisions almost in a row at home, and then under the same circumstances, he fights 12 rounds with Daniel Geale, almost all of which could have legitimately gone either way, and comes out with a loss 116-112 on two cards. You can't map that. And how do you decide who is the A-side between a bigger star without a belt vs. the guy with the belt? And how do you quantify who is the bigger star. To me, there are just too many variables that are particular to each fighter, each judging team, each referee, each promoter, etc, and if the judges do their jobs right and there turns out to be no advantage, you've given the advantage to the other guy by mistake. It strikes me as a "two wrongs don't make a right" situation.

[Finkel]

For example in the case of a champion whose results were heavily reliant on being the A-side, if he no longer holds a belt, then the system needs to be able to account for the disappearance of that A-side advantage. If WHR cannot model for this, then it is going to inflate the rankings of fighters who have just lost their belt.

Too give you an example,

Imagine there is an examiner testing writing proficiency. Unfortunately said examiner is deaf. During the written exam a fire alarm goes off. Wailing away for 30 minutes. The examiner notes the student appears agitated, but just assumes he hasn't prepared well enough for the test. At the end when the results came in, it turned out the student didn't perform as well as the other students in a different location.

WHR is the deaf examiner.
A-side advantage is the screaming alarm.




Edit:. Anyway, I suspect I'm beating a dead horse on this thread, so I'll just stop

Cheers

Yeah I understand your point. But I have to agree with computerrank on this one. It's a real thing, but I think it's too uneven to be fairly adjusted for. Take Felix Sturm, for example. He gets 3 or 4 lucky or questionable decisions almost in a row at home, and then under the same circumstances, he fights 12 rounds with Daniel Geale, almost all of which could have legitimately gone either way, and comes out with a loss 116-112 on two cards. You can't map that. And how do you decide who is the A-side between a bigger star without a belt vs. the guy with the belt? And how do you quantify who is the bigger star. To me, there are just too many variables that are particular to each fighter, each judging team, each referee, each promoter, etc, and if the judges do their jobs right and there turns out to be no advantage, you've given the advantage to the other guy by mistake. It strikes me as a "two wrongs don't make a right" situation.

Fair enough. But it does seem that we agree that this is an influential factor that perhaps cannot be modelled.

So whilst, some people might hold the position that it can't be helped, therefore the rankings are good enough under the circumstances in regards based on what can and cannot be accounted for within the WHR algorithm. I hold the opinion that it just means this is not an appropriate algorithm to base your rankings on if you are forced to ignore a potential huge external factor.

Perhaps we could say there would be a smoothing effect across a fighters career, similar to the mantra in soccer that the good decisions and bad decisions even out across a season. Ignoring for a moment that many a fan of an unfashionable team would strongly oppose this soccer axiom. Soccer is a sport where teams play each other at least twice every season in a competitive league. Whereas professional boxing at the highest level can be anywhere upward of 6 months between fights.

Basically, my position is the fairest way to model an impartial ranking system is to base it on who beat whom and at what time (i.e. relative to each opponents at time ranking)

As such, I always thought of rankings based on who wins hypothetical head-to-heads as being the domain of organizations such as Ring magazine. In contrast I always assumed Boxrec was all about rankings based on strength of resume.

For me, that was what was great about boxrec. It previously appeared to be something closer to unbiased rankings based on resume. Maybe that was a long-term misunderstanding on my part.

But these current rankings just seem like a pointless exercise if they are trying to calculate head-to-head probability whilst being unable to account for the outside factors under discussion.

[marcianofan]

Fair enough. But it does seem that we agree that this is an influential factor that perhaps cannot be modelled.

So whilst, some people might hold the position that it can't be helped, therefore the rankings are good enough under the circumstances in regards based on what can and cannot be accounted for within the WHR algorithm. I hold the opinion that it just means this is not an appropriate algorithm to base your rankings on if you are forced to ignore a potential huge external factor.

Perhaps we could say there would be a smoothing effect across a fighters career, similar to the mantra in soccer that the good decisions and bad decisions even out across a season. Ignoring for a moment that many a fan of an unfashionable team would strongly oppose this soccer axiom. Soccer is a sport where teams play each other at least twice every season in a competitive league. Whereas professional boxing at the highest level can be anywhere upward of 6 months between fights.

Basically, my position is the fairest way to model an impartial ranking system is to base it on who beat whom and at what time (i.e. relative to each opponents at time ranking)

As such, I always thought of rankings based on who wins hypothetical head-to-heads as being the domain of organizations such as Ring magazine. In contrast I always assumed Boxrec was all about rankings based on strength of resume.

For me, that was what was great about boxrec. It previously appeared to be something closer to unbiased rankings based on resume. Maybe that was a long-term misunderstanding on my part.

But these current rankings just seem like a pointless exercise if they are trying to calculate head-to-head probability whilst being unable to account for the outside factors under discussion.

I guess the way I'd look at the predictability model in that kind of situation is "if both of these guys perform typically and get the typical treatment from officials that they normally get (eg any unfair judging because of A-side effects or just incompetence or style preferences), here's who's predicted to win."

So if he gets treated worse or better, the prediction might have a higher error rate. But I don't know that it's much different than a guy having an especially good night or a new trainer or a freak injury. All of that kinda gets baked in or needs to be understood when processing the prediction. Or in other words, it's striving to be a good predictor overall- not a perfect one for specific fighters under their exact individual circumstances.

As you can tell from my previous posts, I strongly prefer the resume-only approach, too. I think Boxrec has been striving for predictability all along, but I think this is the first time those of us who don't care as much about that have noticed it, if only because this is the first time they think they've got something more predictable than what you and I would largely see as a pretty straightforward resume analysis. In other words, I think our goals and the site's goals largely coincided until now, and have diverged (at least until the toggle is added).

[DazBoxingFan]

Why is everyone talking about predictions? Rankings aren’t there for predictions. The rankings are terrible and everyone has complained about them. You’re trying to turn it into some statistical prediction model which is silly

[marcianofan]

Why is everyone talking about predictions? Rankings aren’t there for predictions. The rankings are terrible and everyone has complained about them. You’re trying to turn it into some statistical prediction model which is silly

In fairness, the discussion you are reacting to is a discussion of what is actually being done among two people who, like yourself, don't much like it. I don't have a problem with the site having prediction-driven rankings, really. I just absolutely hate the fact that there's no alternative if that's not what you care about (as I don't, at least for my main purposes). Although to be honest, I'd happily stomach the new rankings and just make common-sense adjustments on my own where needed if only it had a similar scale to the previous model such that one could make an apples to apples comparison (of course the ability to look at previous rankings by date would help with that, as well). The inconsistency and constant state of flux is really what throws a monkey wrench in my cogwork. I've got a top-20 rankings site for each division between heavyweight and 140, and I was using boxrec as a basis for evaluating any given candidate's wins against non-ranked but still significant fighters (ie not top 20, but still top 50). I fell behind by years because of a demanding day-job, and was hoping to reboot it just by re-scaling the old numbers to match the last system and going from there, but then those numbers collapsed again halfway through that process, and now the apples-to-apples is no longer even close to consistent even adjusting for scale, since the whole basis is different.

This all makes me wish the Wayback Machine didn't run into so many redirects while trying to archive the site, because I would all but kill for the ability to look through archived fighter pages and rankings lists consistently. I've had decent success reconstructing 2015 estimated rankings on a fighter-by-fighter basis, but around June of that year, the page URLs changed and about 90% of the pages end up redirecting to the "please login" when crawled.

As much as I don't like the theory behind the new rankings, I'd be willing and able to deal with them if it weren't for the fact that they effectively erased all the ratings boxrec had ever done before, at least for the time being- ratings upon which I had based a great deal of work.

[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.

[computerrank]

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.