Ratings - please read before commenting - Archived

[mike1989]

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

the system is interesting but too difficult to understand, it is more difficult than the actual rating, the ratings will increase all fighters will appear the bunch of extra digits; and how You this option, simple and straightforward :
1. to unsubscribe from cluster systems to ispolzovat only for the fighters move from one category to another;
2. make integrirovannoi the system of calculating the annual rating of the eyes:
200 basis points if/ in place in ranking No. 10 weight category has 150 points
next :
№10 №5 №1 yes.% no.%
150
200 400 600 2,5+ 0
250 500 750 5+ 2,5+
300 600 900 7,5+ 5+
350 700 1050 10+ 7,5+
400 800 1200 15+ 10+
450 900 1350 20+ 15+
500 1000 1500 25+ 20+
600 1200 1800 30+ 25+
700 1400 2100 35+ 30+
800 1600 2400 40+ 35+
900 1800 2700 45+ 40+
1000 2000 3000 50+ 45+
1100+2200+3300+55+ 50+
for a fall:
№10 №5 №1 yes.% no.%
100 200 300 15- 30-
80 160 240 20- 40-
60 120 180 25- 50-
60- 120- 180- 30- 60-
3. 5 % of the maximum number of points in the career of the fighter

[computerrank]

the system is interesting but too difficult to understand, it is more difficult than the actual rating, the ratings will increase all fighters will appear the bunch of extra digits; and how You this option, simple and straightforward :
1. to unsubscribe from cluster systems to ispolzovat only for the fighters move from one category to another;
2. make integrirovannoi the system of calculating the annual rating of the eyes:
200 basis points if/ in place in ranking No. 10 weight category has 150 points
next :
№10 №5 №1 yes.% no.%
150
200 400 600 2,5+ 0
250 500 750 5+ 2,5+
300 600 900 7,5+ 5+
350 700 1050 10+ 7,5+
400 800 1200 15+ 10+
450 900 1350 20+ 15+
500 1000 1500 25+ 20+
600 1200 1800 30+ 25+
700 1400 2100 35+ 30+
800 1600 2400 40+ 35+
900 1800 2700 45+ 40+
1000 2000 3000 50+ 45+
1100+2200+3300+55+ 50+
for a fall:
№10 №5 №1 yes.% no.%
100 200 300 15- 30-
80 160 240 20- 40-
60 120 180 25- 50-
60- 120- 180- 30- 60-
3. 5 % of the maximum number of points in the career of the fighter

The new system fixes some weak points of the old system:

- a boxer gets no points for just being high rated from old times without boxing a value opponent in the same year
- the points depend on the the deepness of the division, not only on the ranks

Sorry, I do not understand your example - would you please explain?

[mike1989]

The new system fixes some weak points of the old system:

- a boxer gets no points for just being high rated from old times without boxing a value opponent in the same year
- the points depend on the the deepness of the division, not only on the ranks

Sorry, I do not understand your example - would you please explain?

All Time ("Inactive") Ratings
The All Time Rating for a boxer is the sum of 33% of the sum of annual rank points he gets for his annual ratings - and 33% of the sum of annual rank points of best defeated opponents - and 14% of his career top rating:
here 14% replacement 5%

the annual rating is the rating at the end of every year the boxer was active
annual rank points = 200 / annual rank in division cluster
№10 №5 №1 yes.% no.%
150
200 400 600 2,5+ 0
250 500 750 5+ 2,5+
300 600 900 7,5+ 5+
350 700 1050 10+ 7,5+
400 800 1200 15+ 10+
450 900 1350 20+ 15+
500 1000 1500 25+ 20+
600 1200 1800 30+ 25+
700 1400 2100 35+ 30+
800 1600 2400 40+ 35+
900 1800 2700 45+ 40+
1000 2000 3000 50+ 45+
1100+2200+3300+55+ 50+
example 2015 Welterweight No. 10 - 288 points , No. 5 - 636 points , No. 1 - 1641 point total : 200 annual rank points+ 5 % and 210 on that, too, and now in the ranking with the exception of the use of cluster

the value of 200 annual points for the top boxer is reduced, if the annual rating of #10 in the division cluster is less than 100 for men,
here is not 100 but 150, and the reference to the following table :
№10 №5 №1 yes.% no.%
100 200 300 15- 30-
80 160 240 20- 40-
60 120 180 25- 50-
60- 120- 180- 30- 60-
2015 : Minimumweight No. 10 - 83 point number 5 - point 138, No. 1 - 243 point total : 200 annual rank points of -40 % and 120 annual rating points on everything as always.

the formula used in opredelenii clusters use only for changes of the actual rating in connection with a change in weight category.

??????????????????????7
sorry for my English , I badly know English, use translator ... )))) waiting for your reply .

[computerrank]

Thanks mike1989,

The intention for using only the ranks is, the all time ratings should be basically independent of the varying point levels over time.

The intention for cutting the maximum annual value is, to compensate for divisions of too low quality ( if not all top 10 at 150+ points). This is for historical times mostly and for some very weak divisions such as minimumweight.

1 Wanheng Menayothin 208
2 Knockout CP Freshmart1 87
3 Byron Rojas 166
4 Hekkie Budler 157
5 Jose Argumedo 145

Here #1 only gets 40 points = 40% of the maximum 100 points, as only 4 boxers are at 150+ points.

I don't think it is helpful to increase the maximum annual value for overshooting annual levels.

[mike1989]

Thanks mike1989,

The intention for using only the ranks is, the all time ratings should be basically independent of the varying point levels over time.

The intention for cutting the maximum annual value is, to compensate for divisions of too low quality ( if not all top 10 at 150+ points). This is for historical times mostly and for some very weak divisions such as minimumweight.

1 Wanheng Menayothin 208
2 Knockout CP Freshmart1 87
3 Byron Rojas 166
4 Hekkie Budler 157
5 Jose Argumedo 145

Here #1 only gets 40 points = 40% of the maximum 100 points, as only 4 boxers are at 150+ points.

I don't think it is helpful to increase the maximum annual value for overshooting annual levels.

a change in the rating which I predlojil would allow to estimate the fair era, and the influence of the maximum value of the rating of a boxer in the rating of all-time ???

[computerrank]

a change in the rating which I predlojil would allow to estimate the fair era, and the influence of the maximum value of the rating of a boxer in the rating of all-time ???

5% of 6000 in peak rating would be 300 points - which is still ways too much.

But I feel better and straighter to go with the ranks only.

[mike1989]

a change in the rating which I predlojil would allow to estimate the fair era, and the influence of the maximum value of the rating of a boxer in the rating of all-time ???

5% of 6000 in peak rating would be 300 points - which is still ways too much.
But I feel better and straighter to go with the ranks only.

for a fair evaluation of rating fighters not to abandon the forces of an era or specific year, 5% of the maximum rating of a fighter and variacija annual rating points from baseline -60% to +60%, making the gap in the show-off between soldiers from different eras but also evaluates them according to merit. plus it is easy. Make your rating more simple and fair taking into account the effect of each year of the epoch... ???? As you look at it ???

[computerrank]

for a fair evaluation of rating fighters not to abandon the forces of an era or specific year, 5% of the maximum rating of a fighter and variacija annual rating points from baseline -60% to +60%, making the gap in the show-off between soldiers from different eras but also evaluates them according to merit. plus it is easy. Make your rating more simple and fair taking into account the effect of each year of the epoch... ???? As you look at it ???

I will add the square root of the career top rating for every boxer to his all time rating.

80 points for a top rating of 6400
30 points for a top rating of 900
10 points for a top rating of 100

[mike1989]

I will add the square root of the career top rating for every boxer to his all time rating.
80 points for a top rating of 6400
30 points for a top rating of 900
10 points for a top rating of 100

this is interesting, but would have meaning and weight in the overall rating if you leave everything like this :
- 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

plus it will not be possible to realistically assess the rating of guys with top 50, top 100, top 200.... ) How to look at it ????

[computerrank]

this is interesting, but would have meaning and weight in the overall rating if you leave everything like this :
- 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

plus it will not be possible to realistically assess the rating of guys with top 50, top 100, top 200.... ) How to look at it ????

Here the square root of the career top rating for every boxer is included in the all time ratings ...

http://backup.boxrec.com/~martin/ann_tops_4.txt
http://backup.boxrec.com/~martin/ann_tops_years_4.txt
http://backup.boxrec.com/~martin/bouts_ranks_4.txt
http://backup.boxrec.com/~martin/all_ti ... _men_4.pdf
http://backup.boxrec.com/~martin/all_ti ... vies_4.pdf

[mike1989]

Here the square root of the career top rating for every boxer is included in the all time ratings ...
http://backup.boxrec.com/~martin/ann_tops_4.txt
http://backup.boxrec.com/~martin/ann_tops_years_4.txt
http://backup.boxrec.com/~martin/bouts_ranks_4.txt
http://backup.boxrec.com/~martin/all_ti ... _men_4.pdf
http://backup.boxrec.com/~martin/all_ti ... vies_4.pdf

What I would have been more understandable back to Vladimir Klitschko ...) why the 2001 and 2003 he did not receive rating points ????
the influence of the maximum value points of the fighter too much??? this will reduce the accuracy of podset leaderboard all-time boys top 100...top 1000 ) ???

[computerrank]

Here the square root of the career top rating for every boxer is included in the all time ratings ...
http://backup.boxrec.com/~martin/ann_tops_4.txt
http://backup.boxrec.com/~martin/ann_tops_years_4.txt
http://backup.boxrec.com/~martin/bouts_ranks_4.txt
http://backup.boxrec.com/~martin/all_ti ... _men_4.pdf
http://backup.boxrec.com/~martin/all_ti ... vies_4.pdf

What I would have been more understandable back to Vladimir Klitschko ...) why the 2001 and 2003 he did not receive rating points ????
the influence of the maximum value points of the fighter too much??? this will reduce the accuracy of podset leaderboard all-time boys top 100...top 1000 ) ???

Wladmir Klitschko didn't box an opponent with at least 150 points in these years. So he didn't get annual points at all, neither for his ranking nor for an opponent. Boxers should not get points for a year, just resting on their points without a valid bout.

From my perspective the all time ratings are even more accurate now. For all time ratings only real top achievemens by boxing valid opponents should count.

[mike1989]

Wladmir Klitschko didn't box an opponent with at least 150 points in these years. So he didn't get annual points at all, neither for his ranking nor for an opponent. Boxers should not get points for a year, just resting on their points without a valid bout.
From my perspective the all time ratings are even more accurate now. For all time ratings only real top achievemens by boxing valid opponents should count.

an interesting formulation of the ranking, she now likes me more than three days ago ))).... but there is naneskolko questions :
1. how to shield a rating of fighters with the top 50...top 1000 ????
2. now 150 points have an average of 20-30 top fighters of the division when in the years 1940-1970, the number of such fighters have become exactly top 100 ???? agree it's not fair ???
3. as you look at something that would be counted against point no 1 defeated the best fighter of the year, and defeated the best fighters : in one year, the fighter can beat No. 1, No. 2, No. 3 division unfair to give him points just for one soldier????
4. in this version of the rating of the clusters is not working ????

[computerrank]

an interesting formulation of the ranking, she now likes me more than three days ago ))).... but there is naneskolko questions :
1. how to shield a rating of fighters with the top 50...top 1000 ????
2. now 150 points have an average of 20-30 top fighters of the division when in the years 1940-1970, the number of such fighters have become exactly top 100 ???? agree it's not fair ???
3. as you look at something that would be counted against point no 1 defeated the best fighter of the year, and defeated the best fighters : in one year, the fighter can beat No. 1, No. 2, No. 3 division unfair to give him points just for one soldier????
4. in this version of the rating of the clusters is not working ????

1. I think, boxers below say top 20 may be neglected for the all time ratings
2. good point - I will restrict it to the top 20
3. I do not understand this point - please explain
4. this version is on single divisions now

[mike1989]

1. I think, boxers below say top 20 may be neglected for the all time ratings
2. good point - I will restrict it to the top 20
3. I do not understand this point - please explain
4. this version is on single divisions now

1.a rating of excellent boxrec the fact that it evaluated all the fighters , not the individual elite, if this be neglected, the rating will lose badly ?????????????
2. boxrec this is the ranking of all the fighters because it is impossible to restrict top 20 ????
3. example: Floyd Mayweather 2013 opposition Robert Guerrero 665 point, Saul Alvarez 872 point, it's not fair to charge only a rating for Alvarez you need for both ???????????

[mike1989]

[quote="mike1989"][quote="computerrank"][quote="mike1989"]
old version I like more, it just need to adjust ????

[computerrank]

1.a rating of excellent boxrec the fact that it evaluated all the fighters , not the individual elite, if this be neglected, the rating will lose badly ?????????????
2. boxrec this is the ranking of all the fighters because it is impossible to restrict top 20 ????
3. example: Floyd Mayweather 2013 opposition Robert Guerrero 665 point, Saul Alvarez 872 point, it's not fair to charge only a rating for Alvarez you need for both ???????????

1. We are talking about all time ratings here, not about current ratings. In all time ratings the focus should be on top 20 performances. The rest of the boxers will be represented by their careeer peek rating here ...
2. We are talking about all time ratings here - and annual ranking points and annual top oppnent ranks should be restricted to top 20 - you showed me pointing to the 1930s to 1970s
3. Mayweather got 100 opponent points AND Alvarez got 100 opponent points for 2013 - where is the problem, please?

[computerrank]

Now I like the new version - with your appreciated improvements - much more, than the old one

- no overemphasizing overshooting ratings of the 1930s to 1970s
- validation in single divisions and not in clusters (so avoiding the advantages of the higher cluster divisions)
- annual rating points only for boxers with valid opposition in the year

I like this

[mike1989]

3. Mayweather got 100 opponent points AND Alvarez got 100 opponent points for 2013 - where is the problem, please?

I mean that the fighter for one year and can beat the No. 1 and No. 2 No. 3 to get 100 points, the other fighter can win only the No. 1 and get the same 100 points ... not fair ???????????))))))

[mike1989]

Now I like the new version - with your appreciated improvements - much more, than the old one

- no overemphasizing overshooting ratings of the 1930s to 1970s
- validation in single divisions and not in clusters (so avoiding the advantages of the higher cluster divisions)
- annual rating points only for boxers with valid opposition in the year

I like this

what I suggested before 10-15 posts ago ), is more fairly evaluated reytig fighters, of course with your modifications ???????????

[computerrank]

3. Mayweather got 100 opponent points AND Alvarez got 100 opponent points for 2013 - where is the problem, please?

I mean that the fighter for one year and can beat the No. 1 and No. 2 No. 3 to get 100 points, the other fighter can win only the No. 1 and get the same 100 points ... not fair ???????????))))))

You think, a boxer with MORE than 1 annual bout against top opponents should get the points for EACH of these bouts? Nice idea

[mike1989]

You think, a boxer with MORE than 1 annual bout against top opponents should get the points for EACH of these bouts? Nice idea

Yes.. that's what I wanted to say .))

[mike1989]

[quote="computerrank"][quote="mike1989"][quote="mike1989"]

- 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

any idea on this point:-- weight = number of division top 10 with rating higher than 150 points / 10
true, the figure is a shield by the average number 20 of each category for each year separately. For example 2016 year:
Heavyweight-145
Cruiserweight-110
Light Heavyweight-149
Super Middleweight-143
Middleweight-215
Light Middleweight-155
Welterweight-215
Light Welterweight-155
Lightweight-173
Super Featherweight-146
Featherweight-173
Super Bantamweight-141
Bantamweight-139
Super Flyweight-115
Flyweight-108
Light Flyweight-74
Minimumweight-58
total = 2352/17= 138,35=138 point
then this figure can be justified on averages, I am the shield that's fair. As dumaesh ????
if you take this idea then I will say that the new version of the ranking for best old ??? )))

[computerrank]

then this figure can be justified on averages, I am the shield that's fair. As dumaesh ????
if you take this idea then I will say that the new version of the ranking for best old ??? )))

I will keep the simple method:

10% value, when 1 boxer is above 150 points
20% value, when 2 boxers are above this value ...
100% value, when 10 boxers are above this value.

I am not convinced I should give extra points for more than 10 boxers above this value.

No boxer above this value will mean no points for this division in this year. More boxers above this value is same value for the annual ratings.


But I will test a version where multiple valid annual opponent ranks will be added.

[mike1989]

I will keep the simple method:

10% value, when 1 boxer is above 150 points
20% value, when 2 boxers are above this value ...
100% value, when 10 boxers are above this value.
I am not convinced I should give extra points for more than 10 boxers above this value.
No boxer above this value will mean no points for this division in this year. More boxers above this value is same value for the annual ratings.
But I will test a version where multiple valid annual opponent ranks will be added.

I don't understand the point possible an example please ???