Page 101 of 180
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 10 May 2016, 09:51
by mike1989
thank you. I just do not understand the last two files ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 10 May 2016, 09:56
by computerrank
mike1989 wrote:
thank you. I just do not understand the last two files ???
They show:
- the all time ratings of all men - pound 4 pound - sorted by htr
-- hr = all time points in official ratings
-- htr = all time points in test ratings (our last test version)
- the all time ratings of all heavyweights men - sorted by htr
-- hr = all time points in official ratings
-- htr = all time points in test ratings (our last test version)
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 10 May 2016, 10:08
by mike1989
computerrank wrote:mike1989 wrote:
thank you. I just do not understand the last two files ???
They show:
- the all time ratings of all men - pound 4 pound - sorted by htr
-- hr = all time points in official ratings
-- htr = all time points in test ratings (our last test version)
- the all time ratings of all heavyweights men - sorted by htr
-- hr = all time points in official ratings
-- htr = all time points in test ratings (our last test version)
thank you . and when will the new redakcija rating of all-time ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 10 May 2016, 11:52
by computerrank
mike1989 wrote:
thank you . and when will the new redakcija rating of all-time ???
Did you find any other errors? I would like to wait, until you have finished ...
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 10 May 2016, 15:16
by mike1989
computerrank wrote:mike1989 wrote:
thank you . and when will the new redakcija rating of all-time ???
Did you find any other errors? I would like to wait, until you have finished ...
even a little look. ) I am very pleased )
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 11 May 2016, 02:47
by mike1989
mike1989 wrote:computerrank wrote:mike1989 wrote:
thank you . and when will the new redakcija rating of all-time ???
why is the maximum value of the rating of boxers of the era of the past 20-60 years practically has not changed that is not true of the subsequent era in Boxing ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 13 May 2016, 06:54
by mike1989
computerrank wrote:mike1989 wrote:
thank you . and when will the new redakcija rating of all-time ???
Did you find any other errors? I would like to wait, until you have finished ...
I don't see any bug...when will be the release of a new rating of all time ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 13 May 2016, 09:31
by mike1989
Naoya Inoue :
2014 for Omar Andres Narvaez should not get 33 point and 66 point at the time of the fight Omar Andres Narvaez was in cluster No. 2 ( 2-1=1)= 66 point ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 13 May 2016, 16:57
by computerrank
mike1989 wrote:Naoya Inoue :
2014 for Omar Andres Narvaez should not get 33 point and 66 point at the time of the fight Omar Andres Narvaez was in cluster No. 2 ( 2-1=1)= 66 point ???
I will look into this in a week ... I am off from home ...
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 13 May 2016, 17:01
by mike1989
computerrank wrote:mike1989 wrote:Naoya Inoue :
2014 for Omar Andres Narvaez should not get 33 point and 66 point at the time of the fight Omar Andres Narvaez was in cluster No. 2 ( 2-1=1)= 66 point ???
I will look into this in a week ... I am off from home ...
okay .
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 19 May 2016, 16:35
by computerrank
mike1989 wrote:computerrank wrote:mike1989 wrote:Naoya Inoue :
2014 for Omar Andres Narvaez should not get 33 point and 66 point at the time of the fight Omar Andres Narvaez was in cluster No. 2 ( 2-1=1)= 66 point ???
I will look into this in a week ... I am off from home ...
okay .
Narvaez was #3 in the bantam / super flyweight cluster:
Code: Select all
432 Juan Francisco Estrada
414 Shinsuke Yamanaka
343 Omar Andres Narvaez
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 19 May 2016, 17:45
by mike1989
computerrank wrote:mike1989 wrote:computerrank wrote:
Narvaez was #3 in the bantam / super flyweight cluster:
Code: Select all
432 Juan Francisco Estrada
414 Shinsuke Yamanaka
343 Omar Andres Narvaez
the Juan Francisco Estrada belongs to the Flyweight division, and this cluster minimumweight+light Flyweight+ Flyweight, Omar Andreas narvales fighter cluster suer Flyweight + Bantamweight.???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 19 May 2016, 17:53
by computerrank
mike1989 wrote:computerrank wrote:mike1989 wrote:
the Juan Francisco Estrada belongs to the Flyweight division, and this cluster minimumweight+light Flyweight+ Flyweight, Omar Andreas narvales fighter cluster suer Flyweight + Bantamweight.???
Estrada's bout on 2014-12-06 was assigned to super flyweight.
But I understand, Estrada should better be assigned to flyweight at this time.
To keep all this up to date, I would have to check all boxers in all top lists before every bout - that might be too much effort for the progam
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 19 May 2016, 18:31
by mike1989
computerrank wrote:mike1989 wrote:computerrank wrote:
the Juan Francisco Estrada belongs to the Flyweight division, and this cluster minimumweight+light Flyweight+ Flyweight, Omar Andreas narvales fighter cluster suer Flyweight + Bantamweight.???
Estrada's bout on 2014-12-06 was assigned to super flyweight.
But I understand, Estrada should better be assigned to flyweight at this time.
To keep all this up to date, I would have to check all boxers in all top lists before every bout - that might be too much effort for the progam
the Estrada in any case, the Flyweight ????
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 19 May 2016, 18:36
by mike1989
on a brush of clusters:
this system is extremely unfair to kruzerveyt , I would suggest to shield the average rate that goes between the cluster and apply them to weight categories then everything will be in equal conditions ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 23 May 2016, 12:50
by computerrank
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 03:55
by mike1989
write the formula by which you shield rating ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 04:28
by computerrank
mike1989 wrote:
write the formula by which you shield rating ???
Each division is ranked by itself.
Boxers are ranked in their best division last year. Opponents are ranked in the division, where the bout happened.
The ratings conversion formula for division changes is the same as before.
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 04:54
by mike1989
computerrank wrote:mike1989 wrote:
write the formula by which you shield rating ???
Each division is ranked by itself.
Boxers are ranked in their best division last year. Opponents are ranked in the division, where the bout happened.
The ratings conversion formula for division changes is the same as before.
this wording is taken as a basis or remains previous ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 05:39
by computerrank
mike1989 wrote:
this wording is taken as a basis or remains previous ???
Did you look into it? It looks better for me.
I guess, I could give the annuals rank points only, if the boxer had a bout against a ranked opponent in the same year ...
Here it is:
http://backup.boxrec.com/~martin/ann_tops_3.txt
http://backup.boxrec.com/~martin/ann_tops_years_3.txt
http://backup.boxrec.com/~martin/bouts_ranks_3.txt
http://backup.boxrec.com/~martin/all_time_men_3.pdf
http://backup.boxrec.com/~martin/all_ti ... vies_3.pdf
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 06:24
by mike1989
computerrank wrote:mike1989 wrote:
this wording is taken as a basis or remains previous ???
Did you look into it? It looks better for me.
I guess, I could give the annuals rank points only, if the boxer had a bout against a ranked opponent in the same year ...
haven't delved into the rating. But pomory write.
I want to share with You their thoughts naschet rating:
1. your formula for the definition of clusters and determined the average value , it will give a single coefficient to be used for lowering or raising the rating if a fighter changes weight class (this will be true with respect to all weight categories) ;
2. leave the clusters on your wording but opredelyaem the dominant division in the cluster for ranking fighters, for example :
in 2015 , the cluster Super Featherweight+Lightweight, saddlebag top 20 Super Featherweight divided by sum of top 20 Lightweight get coefficient 0,849
in favor of the Super Featherweight to determine the annual rating point; this coefficient use for the determination of the best female defeated opozicii for all of 2016;
how about this ????
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 06:36
by mike1989
unable to describe in more detail the new formulas ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 06:46
by computerrank
mike1989 wrote:
haven't delved into the rating. But pomory write.
I want to share with You their thoughts naschet rating:
1. your formula for the definition of clusters and determined the average value , it will give a single coefficient to be used for lowering or raising the rating if a fighter changes weight class (this will be true with respect to all weight categories) ;
2. leave the clusters on your wording but opredelyaem the dominant division in the cluster for ranking fighters, for example :
in 2015 , the cluster Super Featherweight+Lightweight, saddlebag top 20 Super Featherweight divided by sum of top 20 Lightweight get coefficient 0,849
in favor of the Super Featherweight to determine the annual rating point; this coefficient use for the determination of the best female defeated opozicii for all of 2016;
how about this ????
The clusters prefer the highest division. So I will go with the rankings in each distinct division.
For opponent ranking the ratings of both opponents will be converted to the bout's division. This mostly depends on boxers weights and not on population of the divisions.
A cruiserweight boxing at heavyweight will be much less competitive than an original heavyweight, just due to the weight difference. The difference is much smaller at lower divisions as the weight difference is much smaller.
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 08:22
by mike1989
mike1989 wrote:
unable to describe in more detail the new formulas ???
and on a brush ???
Re: Boxrec Ratings - Read first before commenting on the ratings
Posted: 24 May 2016, 10:40
by computerrank
mike1989 wrote:mike1989 wrote:
unable to describe in more detail the new formulas ???
and on a brush ???
- annual year's end rank in division with best performance
-- 100 points / rank * weight
-- weight = number of division top 10 with rating higher than 150 points / 10
-- annual year's end rank points only if there was a ranked best opponent in the same year
- annual best opponent rank converted to bout division a time of bout
-- 100 points / (rank_opp-1) * weight_opp ; when rank_opp is 1 or 2 then use 2
-- weight_opp = number of bout division top 10 with rating higher than 150 points / 10 at time of bout
all time rating is the sum of these annual points over career