Ratings - please read before commenting - Archived

[mike1989]

It turns out that Cruiserweight is not the most beneficial position among all weight categories, even compared to Minimumweight where the average number of active athletes in 4 times less than at Cruiserweight. And the situation will be suhubiette Cruiserweight, but the average weight in the Heavyweight will grow.

[computerrank]

I understand your point with the increasing weight gap from cruiserweight to heavyweight.

The clusters are only relevant for the all time ratings. So current cruiserweights might be handicaped.

But there is another problem for the first years for every new weight division - the first years after 1983 for the cruiserweights. There were very few boxers in the new weight division in these first years. So these historic boxers would be in heavy advantage for a historic all time top rating.

Cruiserweights segregated from the heavyweights not from the light heavyweights. ... and the other new divisions also segregated from the respective classic division above.

So this is no easy problem.

[mike1989]

but in the future this problem will be solved ? have any thoughts on the matter ????

[computerrank]

but in the future this problem will be solved ? have any thoughts on the matter ????

This might be a solution:

Cluster the divisions as done now.

AND now - not convert the rating points to the upper division in the cluster anyway, but to convert the rating to the division with the best performance in the year.

Heavyweights (boxers with best performance in heavyweight in the year) have their rating converted to heavyweight points.
Cruiserweights (boxers with best performance in cruiserweight in the year) have their rating converted to cruiserweight points.

The points are collected in the cluster and ranked then.

[computerrank]

And the result would be this:

Code: Select all

+---------------+-------------+------------------+------+
| division      | first_name  | last_name        | rd   |
+---------------+-------------+------------------+------+
| Cruiserweight | Roy         | Jones Jr         | 1130 |
| Cruiserweight | Carlos      | De Leon          |  356 |
| Cruiserweight | Yoan Pablo  | Hernandez        |  234 |
| Cruiserweight | Ossie       | Ocasio           |  231 |
| Cruiserweight | Orlin       | Norris           |  216 |
| Cruiserweight | Juan Carlos | Gomez            |  208 |
| Cruiserweight | Johnny      | Nelson           |  204 |
| Cruiserweight | O'Neil      | Bell             |  198 |
| Cruiserweight | S.T         | Gordon           |  183 |
| Cruiserweight | Marco       | Huck             |  181 |
| Cruiserweight | Jean Marc   | Mormeck          |  177 |
| Cruiserweight | Anaclet     | Wamba            |  168 |
| Cruiserweight | Marvin      | Camel            |  165 |
| Cruiserweight | Vassiliy    | Jirov            |  153 |
| Cruiserweight | Jeff        | Lampkin          |  150 |
| Cruiserweight | Nate        | Miller           |  142 |
| Cruiserweight | Piet        | Crous            |  132 |
| Cruiserweight | John        | Odhiambho        |  130 |
| Cruiserweight | Danny       | Green            |  127 |
| Cruiserweight | Juan Carlos | Gimenez Ferreyra |  126 |
+---------------+-------------+------------------+------+
20 rows in set (0.13 sec)

[JCS]

Jones #1 at Cruiser? What the hell?

[computerrank]

Jones #1 at Cruiser? What the hell?

Active boxers are included. Their later all time rating may be in an other division ...

[mike1989]

and what hinders to count the rating as you wrote above ???

[Manrae]

Hello guys, I'm a long time fan of this website and I had a question about the new "odds" section.

How exactly are these odds calculated? I'm assuming it's based on the boxer's individual ratings, but I can't figure out how it works.

I clicked on Jessie Vargas' "odds" box and noticed that it was 80/20 for #86 ranked fighter with a rating of "88" and 90/10 for #87 ranked fighter with a rating of "86" ---- (Junior Witter is STILL boxing?)

It seems very specific, how does it work exactly?

Keep up the good work guys

[computerrank]

and what hinders to count the rating as you wrote above ???

Nothing, but the question, what is better?

[computerrank]

Hello guys, I'm a long time fan of this website and I had a question about the new "odds" section.

How exactly are these odds calculated? I'm assuming it's based on the boxer's individual ratings, but I can't figure out how it works.

I clicked on Jessie Vargas' "odds" box and noticed that it was 80/20 for #86 ranked fighter with a rating of "88" and 90/10 for #87 ranked fighter with a rating of "86" ---- (Junior Witter is STILL boxing?)

It seems very specific, how does it work exactly?

Keep up the good work guys

It depends on the absolute values of the individual ratings compared and their ratio. And the more uneven odds are rounded down to 60/40, 70/30, 80/20 and 90/10.

[Manrae]

Hello guys, I'm a long time fan of this website and I had a question about the new "odds" section.

How exactly are these odds calculated? I'm assuming it's based on the boxer's individual ratings, but I can't figure out how it works.

I clicked on Jessie Vargas' "odds" box and noticed that it was 80/20 for #86 ranked fighter with a rating of "88" and 90/10 for #87 ranked fighter with a rating of "86" ---- (Junior Witter is STILL boxing?)

It seems very specific, how does it work exactly?

Keep up the good work guys

It depends on the absolute values of the individual ratings compared and their ratio. And the more uneven odds are rounded down to 60/40, 70/30, 80/20 and 90/10.

Thank you for responding!

I used Welterweights...
#10. Jessie Vargas-368
#86. Little Tyson-88 80/20
#87. Junior Witter-86 90/10

...as an example, but I couldn't figure out the math that changes 80/20 odds to 90/10 odds. Using their absolute values, I tried using division, ratios, medians.... I can't figure it out

Can you explain to me how it works?

[computerrank]

There is an upper and lower ratio limit for every odds category.

The ratio limits for Vargas' 368 points and the odds categories 80/20 and 90/10 is between 88/368 and 86/368

The ratio limits correspond with the winning probabilities based on the bouts recorded in Boxrec.

[Manrae]

There is an upper and lower ratio limit for every odds category.

The ratio limits for Vargas' 368 points and the odds categories 80/20 and 90/10 is between 88/368 and 86/368

The ratio limits correspond with the winning probabilities based on the bouts recorded in Boxrec.

50/50
60/40
70/30
80/20
90/10
95/5
99/1

So for each ratio category, there is a preset ratio limit (upper/lower) that applies to each boxer, regardless of their absolute value, correct?

So for example, If the boxer's rating was 100, what would the ratio limit be for each category?

[computerrank]

There is an upper and lower ratio limit for every odds category.

The ratio limits for Vargas' 368 points and the odds categories 80/20 and 90/10 is between 88/368 and 86/368

The ratio limits correspond with the winning probabilities based on the bouts recorded in Boxrec.

50/50
60/40
70/30
80/20
90/10
95/5
99/1

So for each ratio category, there is a preset ratio limit (upper/lower) that applies to each boxer, regardless of their absolute value, correct?

So for example, If the boxer's rating was 100, what would the ratio limit be for each category?

It depends on the absolute values too. I am off from home currently. I will send you a PM when I am back - till end of Feb.

[mike1989]

how to understand this line of All Time rating : 33% of the sum of annual rank points of best defeated opponents, the opponents in 1 year or top 5 or top 10 for career ???

[Manrae]

There is an upper and lower ratio limit for every odds category.

The ratio limits for Vargas' 368 points and the odds categories 80/20 and 90/10 is between 88/368 and 86/368

The ratio limits correspond with the winning probabilities based on the bouts recorded in Boxrec.

50/50
60/40
70/30
80/20
90/10
95/5
99/1

So for each ratio category, there is a preset ratio limit (upper/lower) that applies to each boxer, regardless of their absolute value, correct?

So for example, If the boxer's rating was 100, what would the ratio limit be for each category?

It depends on the absolute values too. I am off from home currently. I will send you a PM when I am back - till end of Feb.

Ok cool, appreciate it buddy

[mike1989]

how to understand this line of All Time rating : 33% of the sum of annual rank points of best defeated opponents, the opponents in 1 year or top 5 or top 10 for career ???

[JCS]

Is this an interrogation? lol

[computerrank]

how to understand this line of All Time rating : 33% of the sum of annual rank points of best defeated opponents, the opponents in 1 year or top 5 or top 10 for career ???

... sum of rank points for every best opponent in every career year

[mike1989]

)))... one of the best defeated opponent in a year ???

[computerrank]

)))... one of the best defeated opponent in a year ???

... the best in every year only ...

[mike1989]

what happened to awarding ranking points from Sunday rank doesn't update ???

[computerrank]

what happened to awarding ranking points from Sunday rank doesn't update ???

??????

[Romulus9]

Question regarding a recent result/rating:


Dave Markert (0 points) defeated Daniel Craycraft (1 point/1 star) in a six round decision. Craycraft went from 1 to 0 points, but Markert did not receive any points. Craycraft may have fought most of his career in lower classes but all of his fights since returning have been at light heavyweight, the same weight class.

Just curious as to why that was the case, and if Craycraft being correctly computed at Super Middleweight or Light Heavyweight (where he now competes) would make a difference.


Thanks!