Boxrec: How to tell me your rankings are for sale, without telling me your rankings are for sale

[Finkel]

Last week Ngannou (0-1-0) was ranked by your "algorithm" around #500
Today he is ranked #25.

How much are you being paid for that one?

[margaret thatcher]

check out the ratings thread for discussion on ngannou and the formula changes that led to his increase

viewtopic.php?t=249171&start=400

this was the most recent change i think

My last approach to handle the Fury - Ngannou ranking problem resulted in building a pronouced ranking ladder system.

Now I found a more sophisticated approach by keeping a standard number of 3 prior bouts for all boxers with at least 6 career bouts. And working with a reduced number o prior bouts only for boxers with less career bouts. 0.5 prior bouts for a boxer with 1 career bout, 1 prior bout for a boxer with 2 career bouts ... 3 prior bouts for boxers with 6 or more career bouts.

This even improves the resulting winner prediction ratio from 85.67 to 85.84 percent!

And it doesn't really touch the ratings of boxers with 6 and more career bouts. Only boxers with less career bouts are more tied now to the rankings of their opponents than to virtual prior opponents. Especially boxers after their first bout are rated closer to their opponent.

Ngannou would be rated at #28 after his SD loss to Fury.

I will propose that solution for the live system now.

[Finkel]

check out the ratings thread for discussion on ngannou and the formula changes that led to his increase

viewtopic.php?t=249171&start=400

this was the most recent change i think

My last approach to handle the Fury - Ngannou ranking problem resulted in building a pronouced ranking ladder system.

Now I found a more sophisticated approach by keeping a standard number of 3 prior bouts for all boxers with at least 6 career bouts. And working with a reduced number o prior bouts only for boxers with less career bouts. 0.5 prior bouts for a boxer with 1 career bout, 1 prior bout for a boxer with 2 career bouts ... 3 prior bouts for boxers with 6 or more career bouts.

This even improves the resulting winner prediction ratio from 85.67 to 85.84 percent!

And it doesn't really touch the ratings of boxers with 6 and more career bouts. Only boxers with less career bouts are more tied now to the rankings of their opponents than to virtual prior opponents. Especially boxers after their first bout are rated closer to their opponent.

Ngannou would be rated at #28 after his SD loss to Fury.

I will propose that solution for the live system now.

That reads like one man (who keeps) fiddling the rankings to get the desired outcome (of someone with a brown envelope)

Their whole rankings have been a mess since at least the major overhaul in 2020. Each time they fiddle with it they get worse and worse. But hey, it keeps Brick Top happy.
Dubois at #4
Chisora previously at #6 pre Fury fight
Now, Ngannou at #25 post Fury fight


As I suggest in my OP, it looks like they have completely sold out. We already have rankings for cash with the orgs (Brcik Top, WBO, cough cough). So what's the point in having boxrec?

[joshj909]

check out the ratings thread for discussion on ngannou and the formula changes that led to his increase

viewtopic.php?t=249171&start=400

this was the most recent change i think

My last approach to handle the Fury - Ngannou ranking problem resulted in building a pronouced ranking ladder system.

Now I found a more sophisticated approach by keeping a standard number of 3 prior bouts for all boxers with at least 6 career bouts. And working with a reduced number o prior bouts only for boxers with less career bouts. 0.5 prior bouts for a boxer with 1 career bout, 1 prior bout for a boxer with 2 career bouts ... 3 prior bouts for boxers with 6 or more career bouts.

This even improves the resulting winner prediction ratio from 85.67 to 85.84 percent!

And it doesn't really touch the ratings of boxers with 6 and more career bouts. Only boxers with less career bouts are more tied now to the rankings of their opponents than to virtual prior opponents. Especially boxers after their first bout are rated closer to their opponent.

Ngannou would be rated at #28 after his SD loss to Fury.

I will propose that solution for the live system now.

That reads like one man (who keeps) fiddling the rankings to get the desired outcome (of someone with a brown envelope)

Their whole rankings have been a mess since at least the major overhaul in 2020. Each time they fiddle with it they get worse and worse. But hey, it keeps Brick Top happy.
Dubois at #4
Chisora previously at #6 pre Fury fight
Now, Ngannou at #25 post Fury fight


As I suggest in my OP, it looks like they have completely sold out. We already have rankings for cash with the orgs (Brcik Top, WBO, cough cough). So what's the point in having boxrec?

Correct. He did it with Fury when he wanted him ranked first but the previous ratings didn't have him there.

[tiny_acres]

No way can Nganu be rated that high objectively. This is a joke
Dubois at 4th is as much of a joke

[funso banjo baby]

It's a shame boxrec rolled out all the other combat sports. This site should have stayed am and pro boxing only

[margaret thatcher]

i disagree, it's interesting to see someone's whole combat history, and can help you get a better idea of their experience. a guy turning pro with a bunch of kickboxing fights is different than a guy turning pro with no background for example.

[gilgamesh]

To me it's not crazy to rank a guy in the Top 30 when he holds the Heavyweight Champion of the World to a Split Decision and knocks him down.

[Jeff_lacy_ko]

The top 10 in a division is pretty easy for just a fan to do. Making algorithms that spits out nutty results is funny though - ITS OBJECTIVE says they guy fiddling with the numbers to produce the results

[jwfg]

To me it's not crazy to rank a guy in the Top 30 when he holds the Heavyweight Champion of the World to a Split Decision and knocks him down.

This.

[joshj909]

To me it's not crazy to rank a guy in the Top 30 when he holds the Heavyweight Champion of the World to a Split Decision and knocks him down.

The problem is, although Ngannou won in a lot of people's eyes, he lost on paper. I had him losing but Boxrec should be based on results rather than an individual's subjective opinion.

[margaret thatcher]

well tbf his ranking now is still based on the result, it's just interpreted differently. although also tbf it seems the recent change in ranking formula was based on the subjective view that his previous ranking was too low....

it's a very weird situation. it should matter that he has next to no boxing results, but it should also matter that he dropped and ran very close vs the arguable world #1 at the time. very few others have a comparable effort, and those guys in the #500 range that fran was previously at are universes away from that. i think you could feasibly put fran at many different spots in the top 25-100 range

[Finkel]

well tbf his ranking now is still based on the result, it's just interpreted differently. although also tbf it seems the recent change in ranking formula was based on the subjective view that his previous ranking was too low....

it's a very weird situation. it should matter that he has next to no boxing results, but it should also matter that he dropped and ran very close vs the arguable world #1 at the time. very few others have a comparable effort, and those guys in the #500 range that fran was previously at are universes away from that. i think you could feasibly put fran at many different spots in the top 25-100 range

Well, you are kind of arguing for objective rating to include subjectivity there. But, I agree it is certainly "interpreted differently", It must be the same "different" interpretation that has a fighter at #4 because he scored a non-KD (officially a foul) and got stopped off a jab against the Boxrec champion.

For me the underlying problem is what Boxrec presents itself as.
We have our own subjective community ranking system. Ngannou made it into the top 15 with less than a quarter of the voters. Fine, I don't agree with it, but it is the system that is in place, and I'm not the dictator of our community rankings. But to my knowledge, BoxRec isn't meant to have a dictator with his finger on the balance either.
This is the promise made by BoxRec on their website: ©BoxRec is the official record keeper for 410 sports authorities worldwide, it is not under direct control of any single authority. That is what most fans think of when they hear the Brand Boxrec: an organization whose main goal is to maintain boxer's resumes.
Given that is the case, what assumption are the general public making when they read their rankings? That they are an objective (or even subjective) ranking based purely on resume? Or they are another ranking system that includes factors such as "the eye test", "in ring performance", "popularity" - basically The Ring.

I argued this with ComputerRank a few years back. He doesn't see the problem with it (so be it). However, I think it is incongruent with the very thing BoxRec is advertising itself as being, to have this kind of Rankings for Sale system. (I mean look at the state of the current top 10, then consider who those rankings benefit - I can break it down, but it should be bloody obvious what is going on).

Personally, I think they have completely sold out.

[margaret thatcher]

well tbf his ranking now is still based on the result, it's just interpreted differently. although also tbf it seems the recent change in ranking formula was based on the subjective view that his previous ranking was too low....

it's a very weird situation. it should matter that he has next to no boxing results, but it should also matter that he dropped and ran very close vs the arguable world #1 at the time. very few others have a comparable effort, and those guys in the #500 range that fran was previously at are universes away from that. i think you could feasibly put fran at many different spots in the top 25-100 range

Well, you are kind of arguing for objective rating to include subjectivity there. But, I agree it is certainly "interpreted differently", It must be the same "different" interpretation that has a fighter at #4 because he scored a non-KD (officially a foul) and got stopped off a jab against the Boxrec champion.

For me the underlying problem is what Boxrec presents itself as.
We have our own subjective community ranking system. Ngannou made it into the top 15 with less than a quarter of the voters. Fine, I don't agree with it, but it is the system that is in place, and I'm not the dictator of our community rankings. But to my knowledge, BoxRec isn't meant to have a dictator with his finger on the balance either.
This is the promise made by BoxRec on their website: ©BoxRec is the official record keeper for 410 sports authorities worldwide, it is not under direct control of any single authority. That is what most fans think of when they hear the Brand Boxrec: an organization whose main goal is to maintain boxer's resumes.
Given that is the case, what assumption are the general public making when they read their rankings? That they are an objective (or even subjective) ranking based purely on resume? Or they are another ranking system that includes factors such as "the eye test", "in ring performance", "popularity" - basically The Ring.

I argued this with ComputerRank a few years back. He doesn't see the problem with it (so be it). However, I think it is incongruent with the very thing BoxRec is advertising itself as being, to have this kind of Rankings for Sale system. (I mean look at the state of the current top 10, then consider who those rankings benefit - I can break it down, but it should be bloody obvious what is going on).

Personally, I think they have completely sold out.

there is no way to completely exclude subjectivity. certain values must be decided on an input into the system to make it run - that means subjectivity.

likewise, subjectivity is involved in our assessments of whether a computerized, so-called objective system is doing a good job and reflects reality.

[Finkel]

well tbf his ranking now is still based on the result, it's just interpreted differently. although also tbf it seems the recent change in ranking formula was based on the subjective view that his previous ranking was too low....

it's a very weird situation. it should matter that he has next to no boxing results, but it should also matter that he dropped and ran very close vs the arguable world #1 at the time. very few others have a comparable effort, and those guys in the #500 range that fran was previously at are universes away from that. i think you could feasibly put fran at many different spots in the top 25-100 range

Well, you are kind of arguing for objective rating to include subjectivity there. But, I agree it is certainly "interpreted differently", It must be the same "different" interpretation that has a fighter at #4 because he scored a non-KD (officially a foul) and got stopped off a jab against the Boxrec champion.

For me the underlying problem is what Boxrec presents itself as.
We have our own subjective community ranking system. Ngannou made it into the top 15 with less than a quarter of the voters. Fine, I don't agree with it, but it is the system that is in place, and I'm not the dictator of our community rankings. But to my knowledge, BoxRec isn't meant to have a dictator with his finger on the balance either.
This is the promise made by BoxRec on their website: ©BoxRec is the official record keeper for 410 sports authorities worldwide, it is not under direct control of any single authority. That is what most fans think of when they hear the Brand Boxrec: an organization whose main goal is to maintain boxer's resumes.
Given that is the case, what assumption are the general public making when they read their rankings? That they are an objective (or even subjective) ranking based purely on resume? Or they are another ranking system that includes factors such as "the eye test", "in ring performance", "popularity" - basically The Ring.

I argued this with ComputerRank a few years back. He doesn't see the problem with it (so be it). However, I think it is incongruent with the very thing BoxRec is advertising itself as being, to have this kind of Rankings for Sale system. (I mean look at the state of the current top 10, then consider who those rankings benefit - I can break it down, but it should be bloody obvious what is going on).

Personally, I think they have completely sold out.

there is no way to completely exclude subjectivity. certain values must be decided on an input into the system to make it run

ahem, true, but your whole argument was based on subjective assessment, I was trying to be fair to you by not pointing it all out.

"it's a very weird situation. it should matter that he has next to no boxing results, but it should also matter that he dropped and ran very close vs the arguable world #1 at the time. very few others have a comparable effort, and those guys in the #500 range that fran was previously at are universes away from that. i think you could feasibly put fran at many different spots in the top 25-100 range"

At the end of the day, what the new ranking algorithm has effectively done is legitimise poor opponent selection by the top guys after the fact. Do we really need that as fans when the WBC already exists? Does anyone really want BoxRec to be doing that?

[margaret thatcher]

that's my just take on where i think ngannou should or at least can feasibly be rated. i can then compare this to where boxrec has him to get a sense of how similar our rankings are, but it's not me trying to build the inputs for a computerized system. although that said, 'close fight' and world #1 are things that can be made concrete - .e.g a split dec could count as a close fight in someone'es ranking system. a system could have had fury with more points than anyone else. a fighter may not have a win, draw, or any close loss vs someone with anywhere near that many points, etc.

you simply cant get rid of all subjective elements. your own criticisms of the system are riddled with subjectivity too and seem heavily based on your own personal disagreement with them

[Finkel]

that's my just take on where i think ngannou should or at least can feasibly be rated. i can then compare this to where boxrec has him to get a sense of how similar our rankings are, but it's not me trying to build the inputs for a computerized system. although that said, 'close fight' and world #1 are things that can be made concrete - .e.g a split dec could count as a close fight in someone'es ranking system. a system could have had fury with more points than anyone else. a fighter may not have a win, draw, or any close loss vs someone with anywhere near that many points, etc.

you simply cant get rid of all subjective elements. your own criticisms of the system are riddled with subjectivity too and seem heavily based on your own personal disagreement with them

Please make the argument. I welcome trying to get to the truth of the matter. Explain where what I have said is subjective, why you take issue with it. And I will either accept you are correct or make my case more.

What I expect from an algorithm is different, I won't agree with everything (they are completely out of whack with our own rankings btw - which is absolutely fine) but the process is just as important as the outcomes for the algorithm. What do the rankings represent? The PBO had a very solid ranking system based on an algorithm. The current BoxRec top 10 is an abomination (subjective I know).

[margaret thatcher]

lol is that a serious question? you make all sorts of subjective assessments and assumptions. for one that the rankings are for sale

loads of others in different ranking threads about fighters ability level and the worth of their resumes. it's peoples subejective assessments that lead them to complain about boxrec rankings pretty much - there's a problem because they dont match

there's really no way to be completely objective either tbf

[margaret thatcher]

well there you go, some big assumptions!

[Finkel]

lol is that a serious question? you make all sorts of subjective assessments and assumptions. for one that the rankings are for sale

loads of others in different ranking threads about fighters ability level and the worth of their resumes

there's really no way to be completely objective either tbf

Yeah, I was tempted to reword that at the end, but I thought you would be able to fill in the blanks. Especially considering I said it was my opinion.

Maybe better to say, the outcomes are so in tune with the best possible outcomes for the promotional companies connected to the Saudi cards that the rankings might as well be for sale. Maybe that is just a coincidence that this keeps happening after adjustments to the algorithm.

[Finkel]

well there you go, some big assumptions!

Big assumptions about what? That ranking Dubois at #4 is terrible.
I think if an algorithm is spitting out Dubois #4 and Chisora #10, this is no longer just subjectively bad. But sure, I guess you can argue there is no such thing as true objectivity. But I'm not sure how that moves this discussion forward.

ComputerRank is telling us how it works. He even posted a link to the original chess paper he took the new algorithm from. We also know he is fiddling with it, to get more desirable outcomes. At best we can say it is just a coincidence that those desirable outcomes would be very desirable for the Saudi promotions and their respective promoters.

But you are correct, my whole argument is that BoxRec should rank resumes. And you are correct that is my subjective opinion. And in fact, as a private enterprise BoxRec can do whatever they like. But it is also true that boxing fans exist who believe that BoxRec is purely ranking resumes, and so are very confused when they look at the rankings and see things like Dubois at #4.

Now a counter argument would be "well it is ranking resumes", but I would like to see a breakdown of how that is being done to get the top 10 we have now (with a straight face).

And maybe they could follow that up with how everything the WBC does with their rankings is above board too.

[Finkel]

The top 10 in a division is pretty easy for just a fan to do. Making algorithms that spits out nutty results is funny though - ITS OBJECTIVE says they guy fiddling with the numbers to produce the results

The algorithm giveth, and the algorithm taketh away. All hail the algorithm.

[gilgamesh]

To me it's not crazy to rank a guy in the Top 30 when he holds the Heavyweight Champion of the World to a Split Decision and knocks him down.

The problem is, although Ngannou won in a lot of people's eyes, he lost on paper. I had him losing but Boxrec should be based on results rather than an individual's subjective opinion.

It is taking the result into account. The result was a highly competitive fight with a Split Decision that caused a lot of debate.

All losses are not created equal.

[joshj909]

To me it's not crazy to rank a guy in the Top 30 when he holds the Heavyweight Champion of the World to a Split Decision and knocks him down.

The problem is, although Ngannou won in a lot of people's eyes, he lost on paper. I had him losing but Boxrec should be based on results rather than an individual's subjective opinion.

It is taking the result into account. The result was a highly competitive fight with a Split Decision that caused a lot of debate.

All losses are not created equal.

That subjective opinion cannot be factored into an objective decision. I thought Ngannou won, however, you can only have rankings based on wins and losses and the method of victory/loss. Anything else is subjective and this changing of the algorithm once again to support Fury and his opponents is getting a bit ridiculous.

[margaret thatcher]

that the result was a split decision, and not a ud, or a stoppage, etc is objective. even the forum game jcs used to run would differentiate between methods of victory when it came to awarding points, with anything split being classed as 'close'. id have to think any good ranking system would go beyond black and white wins and losses

i think fran's current ranking is better than him being in the 500s. that said if you read the ranking thread the situation was one where the rankings were changed to produce a specific outcome (ngganou getting a boost),, and tbf i do see where that its on shaky grounds to make system wide changes in order to change one boxer's ranking