Ratings - please read before commenting

Rating Queries
Locked
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

1628268 wrote: 19 Sep 2024, 10:53 Hello,

I hope you’re doing well.

I’m writing to ask for some clarification regarding recent changes to Victor Nagbe’s ranking on Boxrec.com.

After the Nikita Tszyu vs Koen Mazoudier fight on 28/08/2024, Victor was ranked #5 in the Super Welterweight division, with Koen dropping to #8. However, about a week later, without any new fights or events, Victor’s rank dropped from #5 to #6, and Dan Hill moved up to #5. This is surprising given that Victor had, and still has, a superior record (8 Wins, 3 KOs, 1 Loss, 0 KOs) compared to Dan Hill (6 Wins, 3 KOs, 1 Loss, 1 KO).

Interestingly, after Dan Hill’s fight on 14/09/2024, the rankings automatically returned Victor to #5, but a few days later, his rank dropped again without any fights taking place. This has raised some questions, particularly around the possibility of human intervention in the rankings.

Could you please help us understand how these ranking changes occurred and whether there were any factors or interventions involved? We are trying to get a clearer picture of how the system works and what might have influenced these adjustments.

Please note that we have read the information about how the Boxrec algorithm works. But any information you can provide regarding this case would be very helpful.
Hello,
BoxRec ratings only depend on official results and the BoxRec ratings of the opponents at time of their bout.
That means BoxRec ratings are completly and automatically calculated by computer and are not influenced by any rankings in ABC organizations or titles.
The changes you mention are the pure result of these calculations.
Greetings
Martin Reichert
Andrew Brown
Minimumweight
Posts: 6
Joined: 02 Jun 2024, 06:05

Re: Ratings - please read before commenting

Post by Andrew Brown »

Dear Martin,

Thank you for your response. It gives us peace of mind to know that there are no human influences in the Boxrec ratings, other than the official bout results. This consistency and automation ensure that the system applies the same logic across the board, avoiding any potential bias or subjective interpretation, which is always a concern when human intervention is involved.

I appreciate you taking the time to clarify this, and it’s reassuring to understand how the rankings are calculated automatically.

Thank you once again for your time.
JCS
Heavyweight
Heavyweight
Posts: 6174
Joined: 17 Dec 2004, 13:27

Re: Ratings - please read before commenting

Post by JCS »

Hate to bring up the Ngannou thing again since he did get crushed by Joshua...

But having him outside of the Top 700 seems a bit cruel, no? Am I crazy for thinking he belongs in the Top 50 for nearly drawing with Fury?
margaret thatcher
Super Bantamweight
Posts: 38085
Joined: 22 Jul 2019, 15:43

Re: Ratings - please read before commenting

Post by margaret thatcher »

1628268 wrote: 20 Sep 2024, 22:42 Okay so you are not from Boxrec!

You are just swinging statements and throwing irrelevant names and scores without actually explaining yourself or anything.

I have provided evidence and records.
But you wish to ignore that.

Then you go into Boxrec manipulation, which no one mentioned other than you.

I have asked legitimate questions and if you don't have any answers for me which you clearly don't, please mind your business and let someone who knows what they're talking about handle this.

Thank you for trying to answer this query, but you're being counter productive than helpful.

Cheers.
the records show that dan hill beat your man :TU:

8-1 somehow being superior to 6-1 , just because it's more wins , is seriously lol. surely you cant be that naive if youre handling someone's career!!
margaret thatcher
Super Bantamweight
Posts: 38085
Joined: 22 Jul 2019, 15:43

Re: Ratings - please read before commenting

Post by margaret thatcher »

danny dubois new boxrec #1 hw, fury #2, usyk #3 :yay:
mike1989
Super Middleweight
Posts: 323
Joined: 06 Feb 2012, 10:00

Re: Ratings - please read before commenting

Post by mike1989 »

Why Usyk No. 3, Fury No. 2 and Dubois No. 1???? Adeye Fury and Dubois lost to Usyk. Just like Joshua lost???? What is it???))))...
JCS
Heavyweight
Heavyweight
Posts: 6174
Joined: 17 Dec 2004, 13:27

Re: Ratings - please read before commenting

Post by JCS »

This has probably been said 20 times.. but the post-fight rankings are approximate.

A full update takes place overnight at some point....

So wake up tomorrow and complain :)

I am quite curious to see them.

Before the fight it was...

Joshua, Usyk, Fury, Kabayel, Dubois
JCS
Heavyweight
Heavyweight
Posts: 6174
Joined: 17 Dec 2004, 13:27

Re: Ratings - please read before commenting

Post by JCS »

All the complaining must've prompted an early refresh of the ratings...

Joshua, Usyk, Fury, Kabayel, Dubois
becomes
Usyk, Dubois, Fury, Kabayel, Dubois

I'd probably go
Usyk, Fury, Dubois, Joshua, Kabayel
Andrew Brown
Minimumweight
Posts: 6
Joined: 02 Jun 2024, 06:05

Re: Ratings - please read before commenting

Post by Andrew Brown »

margaret thatcher wrote: 21 Sep 2024, 17:20
1628268 wrote: 20 Sep 2024, 22:42 Okay so you are not from Boxrec!

You are just swinging statements and throwing irrelevant names and scores without actually explaining yourself or anything.

I have provided evidence and records.
But you wish to ignore that.

Then you go into Boxrec manipulation, which no one mentioned other than you.

I have asked legitimate questions and if you don't have any answers for me which you clearly don't, please mind your business and let someone who knows what they're talking about handle this.

Thank you for trying to answer this query, but you're being counter productive than helpful.

Cheers.
the records show that dan hill beat your man :TU:

8-1 somehow being superior to 6-1 , just because it's more wins , is seriously lol. surely you cant be that naive if youre handling someone's career!!
Actually, our boxers and athletes love us because we go in to bat for them and ask questions when needed, that's what they want, not self serving managers in bed with corrupt promoters and officials. The question is why are you so butt-hurt :witzend:
margaret thatcher
Super Bantamweight
Posts: 38085
Joined: 22 Jul 2019, 15:43

Re: Ratings - please read before commenting

Post by margaret thatcher »

1628268 wrote: 22 Sep 2024, 12:49
margaret thatcher wrote: 21 Sep 2024, 17:20
1628268 wrote: 20 Sep 2024, 22:42 Okay so you are not from Boxrec!

You are just swinging statements and throwing irrelevant names and scores without actually explaining yourself or anything.

I have provided evidence and records.
But you wish to ignore that.

Then you go into Boxrec manipulation, which no one mentioned other than you.

I have asked legitimate questions and if you don't have any answers for me which you clearly don't, please mind your business and let someone who knows what they're talking about handle this.

Thank you for trying to answer this query, but you're being counter productive than helpful.

Cheers.
the records show that dan hill beat your man :TU:

8-1 somehow being superior to 6-1 , just because it's more wins , is seriously lol. surely you cant be that naive if youre handling someone's career!!
Actually, our boxers and athletes love us because we go in to bat for them and ask questions when needed, that's what they want, not self serving managers in bed with corrupt promoters and officials. The question is why are you so butt-hurt :witzend:
not butt hurt at all, just pointing out the utter silliness and naivety of your understanding of records :TU:
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

Cobwebcat wrote: 05 Oct 2024, 13:30 Computerrank…If you were to compare the raw WHR predictive percentage with the WHR plus Boxrecs modifications predictive percentage how to they compare?
The only specific BoxRec modification is the winner-above-loser rule, which on tendency has a bit lower predictive percentage.
JCS
Heavyweight
Heavyweight
Posts: 6174
Joined: 17 Dec 2004, 13:27

Re: Ratings - please read before commenting

Post by JCS »

I'm curious if the Japanese boxers benefit in the ratings by having a strong "hometown" advantage. That'd be fun to analyze...

Maybe it is just me... but these top guys NEVER leave the country... Over the years, it has felt like they've performed so much better at home.... and vice versa.
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

JCS wrote: 12 Oct 2024, 22:21 I'm curious if the Japanese boxers benefit in the ratings by having a strong "hometown" advantage. That'd be fun to analyze...

Maybe it is just me... but these top guys NEVER leave the country... Over the years, it has felt like they've performed so much better at home.... and vice versa.
BoxRec found and uses advantage corrections for the winner prediction:

$this->common['adv_height'] = 1.28; #more than 10 cm difference
$this->common['adv_stance'] = 1.16; #southpaw against orthodox
$this->common['adv_prom'] = 1.22; #event promoted by own promoter
$this->common['adv_country'] = 1.36; #home advantage
$this->common['adv_koed'] = 1.22; #disadvantage for being koed in bout before

The rating of the advantage owner is multiplied by this factor and the rating of his opponent is divided.
JCS
Heavyweight
Heavyweight
Posts: 6174
Joined: 17 Dec 2004, 13:27

Re: Ratings - please read before commenting

Post by JCS »

computerrank wrote: 13 Oct 2024, 05:17
JCS wrote: 12 Oct 2024, 22:21 I'm curious if the Japanese boxers benefit in the ratings by having a strong "hometown" advantage. That'd be fun to analyze...

Maybe it is just me... but these top guys NEVER leave the country... Over the years, it has felt like they've performed so much better at home.... and vice versa.
BoxRec found and uses advantage corrections for the winner prediction:

$this->common['adv_height'] = 1.28; #more than 10 cm difference
$this->common['adv_stance'] = 1.16; #southpaw against orthodox
$this->common['adv_prom'] = 1.22; #event promoted by own promoter
$this->common['adv_country'] = 1.36; #home advantage
$this->common['adv_koed'] = 1.22; #disadvantage for being koed in bout before

The rating of the advantage owner is multiplied by this factor and the rating of his opponent is divided.
This is used in rating calc or just to gauge prediction?
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

JCS wrote: 13 Oct 2024, 10:45
computerrank wrote: 13 Oct 2024, 05:17
JCS wrote: 12 Oct 2024, 22:21 I'm curious if the Japanese boxers benefit in the ratings by having a strong "hometown" advantage. That'd be fun to analyze...

Maybe it is just me... but these top guys NEVER leave the country... Over the years, it has felt like they've performed so much better at home.... and vice versa.
BoxRec found and uses advantage corrections for the winner prediction:

$this->common['adv_height'] = 1.28; #more than 10 cm difference
$this->common['adv_stance'] = 1.16; #southpaw against orthodox
$this->common['adv_prom'] = 1.22; #event promoted by own promoter
$this->common['adv_country'] = 1.36; #home advantage
$this->common['adv_koed'] = 1.22; #disadvantage for being koed in bout before

The rating of the advantage owner is multiplied by this factor and the rating of his opponent is divided.
This is used in rating calc or just to gauge prediction?
It is only used in prediction. It improves the prediction ratio my more than 1 percent point.

But application to the ratings results in a lower predicition ratio :verysad:
JCS
Heavyweight
Heavyweight
Posts: 6174
Joined: 17 Dec 2004, 13:27

Re: Ratings - please read before commenting

Post by JCS »

That makes sense.. if used in the ratings.. it is sort of like punishing a boxer for being tall or being southpaw... yet this is an advantage they have in every fight. Same applies if they almost always fight a home.

You should look at age and reach!
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

JCS wrote: 13 Oct 2024, 11:40 That makes sense.. if used in the ratings.. it is sort of like punishing a boxer for being tall or being southpaw... yet this is an advantage they have in every fight. Same applies if they almost always fight a home.

You should look at age and reach!
I did and there was not significant influence additional to height.
Manrae
Heavyweight
Heavyweight
Posts: 341
Joined: 28 Nov 2002, 18:57

Re: Ratings - please read before commenting

Post by Manrae »

computerrank wrote: 13 Oct 2024, 15:44
JCS wrote: 13 Oct 2024, 11:40 That makes sense.. if used in the ratings.. it is sort of like punishing a boxer for being tall or being southpaw... yet this is an advantage they have in every fight. Same applies if they almost always fight a home.

You should look at age and reach!
I did and there was not significant influence additional to height.
What about amount of fights (experience)?
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

Manrae wrote: 13 Oct 2024, 23:45
computerrank wrote: 13 Oct 2024, 15:44
JCS wrote: 13 Oct 2024, 11:40 That makes sense.. if used in the ratings.. it is sort of like punishing a boxer for being tall or being southpaw... yet this is an advantage they have in every fight. Same applies if they almost always fight a home.

You should look at age and reach!
I did and there was not significant influence additional to height.
What about amount of fights (experience)?
There is a correction for the rating of boxers with less than 6 bouts.
Manrae
Heavyweight
Heavyweight
Posts: 341
Joined: 28 Nov 2002, 18:57

Re: Ratings - please read before commenting

Post by Manrae »

https://boxrec.com/en/iba-pro/733152

Is there a plan to add star ratings to IBA pro fights? All of them are unrated, except for this one...
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

Manrae wrote: 05 Nov 2024, 00:54 https://boxrec.com/en/iba-pro/733152

Is there a plan to add star ratings to IBA pro fights? All of them are unrated, except for this one...
Currently no ratings for IBA_pro fights are calculated.
Manrae
Heavyweight
Heavyweight
Posts: 341
Joined: 28 Nov 2002, 18:57

Re: Ratings - please read before commenting

Post by Manrae »

computerrank wrote: 05 Nov 2024, 11:42
Manrae wrote: 05 Nov 2024, 00:54 https://boxrec.com/en/iba-pro/733152

Is there a plan to add star ratings to IBA pro fights? All of them are unrated, except for this one...
Currently no ratings for IBA_pro fights are calculated.
So, why does this one have a 4 star rating? :neutral:
computerrank
Editor
Editor
Posts: 2481
Joined: 04 Jan 2003, 18:59

Re: Ratings - please read before commenting

Post by computerrank »

Manrae wrote: 06 Nov 2024, 08:36
computerrank wrote: 05 Nov 2024, 11:42
Manrae wrote: 05 Nov 2024, 00:54 https://boxrec.com/en/iba-pro/733152

Is there a plan to add star ratings to IBA pro fights? All of them are unrated, except for this one...
Currently no ratings for IBA_pro fights are calculated.
So, why does this one have a 4 star rating? :neutral:
It seems to have been entered as box-pro first - now the rating fields contain the ratings.
Manrae
Heavyweight
Heavyweight
Posts: 341
Joined: 28 Nov 2002, 18:57

Re: Ratings - please read before commenting

Post by Manrae »

computerrank wrote: 06 Nov 2024, 10:43
Manrae wrote: 06 Nov 2024, 08:36
computerrank wrote: 05 Nov 2024, 11:42
Currently no ratings for IBA_pro fights are calculated.
So, why does this one have a 4 star rating? :neutral:
It seems to have been entered as box-pro first - now the rating fields contain the ratings.
It does?

https://boxrec.com/en/iba-pro/856311

I'm assuming it'll show after a refresh (tomorrow)?
Locked