Ratings - please read before commenting - Archived

Locked
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

All time rating forumula changed to sum of
- 50% of current all time rating - career peak slidung averaged value (by some 14 bouts)
- 5% of sum of all annual ratings
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

Sergio Gabriel Martinez just lost 2 ranking spots after winning a 40-36, 40-36, 39-35 fight.

That doesn't seem correct.... You can't win a 4 round decision by much wider scorecards than that.
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

JCS wrote:Sergio Gabriel Martinez just lost 2 ranking spots after winning a 40-36, 40-36, 39-35 fight.

That doesn't seem correct.... You can't win a 4 round decision by much wider scorecards than that.
... there is no problem in the calculation these 4 points per judge are much more than needed for a clear decision - you only need 2 points per judge for 4 rounds boxed.

There must have been some impact from some other change.
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

Cobwebcat wrote:
computerrank wrote:All time rating forumula changed to sum of
- 50% of current all time rating - career peak slidung averaged value (by some 14 bouts)
- 5% of sum of all annual ratings
I'd like to say they look better...but they don't. Cooper better than Lewis?...give me a break!!
Yeah, that's pretty bad.

The all-time ratings are a lost cause as are the P4P. We ought to just scrap them.
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

Cobwebcat wrote:
JCS wrote:
Cobwebcat wrote: I'd like to say they look better...but they don't. Cooper better than Lewis?...give me a break!!
Yeah, that's pretty bad.

The all-time ratings are a lost cause as are the P4P. We ought to just scrap them.

That would be a shame as there is nothing like them on the net. Boxrec is the only site that has access to the information. No way should all that work go to waste. If it was used in a meaningful way they could draw people to the site.

I don't know why there was a change and I'm not against change for the better but these look silly now and they were reasonable before.
There are so many intangibles and era differences involved..... I don't know..
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

Yes, I see - when I tested this linear combination some weeks ago - it was pretty better. Seems to be an impact of the recent no-win correction of the current ratings.

I will reduce the part of the sum of the annual ratings - this should work.

Look at the previous test:

http://www.boxrec.com/forum/viewtopic.p ... 37#1011737
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

A fix is proposed for launch ...
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

MatthewS wrote:
JCS wrote:
computerrank wrote: ... I implemented a fix for this problem. Wins agains winless opponents will not be honored any longer - even in lower rating regions.

I proposed the new release for immediate launch.
:TU:

This one was long overdue. This helps some overrating of the exclusively foreign fighters building up their ratings with soft opposition.
Is Julio Cesar Chavez fcuked then? Or is that only Females? A lot of the older records are fighters who show up only as 0-2-0 or similar
This fix would be a hindrance to fighters like him... yes.
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

computerrank wrote:A fix is proposed for launch ...
... done
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

computerrank wrote:
computerrank wrote:A fix is proposed for launch ...
... done
John is on top of things these days, the updates used to take days or weeks!
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

JCS wrote:
computerrank wrote:
computerrank wrote:A fix is proposed for launch ...
... done
John is on top of things these days, the updates used to take days or weeks!
... turbo days ...
jomothepure
Heavyweight
Heavyweight
Posts: 2660
Joined: 24 Oct 2004, 08:43

Post by jomothepure »

I am ashamed to say I am turbo unnainititaed in the ways of the rankings.

I also keep clear of heart stoppping threads such as "what put x ahead of y," but seriously, what has put Sprott ahead of Skelton?
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

jomothepure wrote:I am ashamed to say I am turbo unnainititaed in the ways of the rankings.

I also keep clear of heart stoppping threads such as "what put x ahead of y," but seriously, what has put Sprott ahead of Skelton?
And Cintron ahead of Margarito :) Its just the way they shuffle.

Personally, I think there should be a bit more mandatory space between the winner and loser to better prevent this from happening.
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

The basic message is, they are at same level - these boxer are within +-50 points.

11 Lamon Brewster 446
12 Alexander Dimitrenko 444
13 Calvin Brock 423
14 Michael Sprott 416
15 Matt Skelton 409
16 Chris Arreola 404
17 Andrew Golota 401
18 Danny Williams 393 37
19 Hasim Rahman 391
20 James Toney 373
IKSRTFO
Heavyweight
Heavyweight
Posts: 4753
Joined: 09 Dec 2007, 17:14

Post by IKSRTFO »

Hopkins #1 P4P?. Cotto 25. Haye and Klit best male boxers?
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

John has excluded active boxers from the all time ratings - too much confusion before - as active boxers are not assigned to their career top division but to the division according with their current weight.
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

Cobwebcat wrote:
computerrank wrote:John has excluded active boxers from the all time ratings - too much confusion before - as active boxers are not assigned to their career top division but to the division according with their current weight.
I can see the logic in that. However, how is it that time can be found to do that but not re-introduce the all-time value figures? It's been nearly a year now! Time toadmit they aren't coming back?
... maybe, he doesn't want to see them ...
The Irish Assassin
Heavyweight
Heavyweight

Re: www.spam

Post by The Irish Assassin »

ehtrend wrote:check out this site - live and archived boxing, shows, news, audio and video interviews and alot more - great site.
Image
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

Ratings broke?
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

The daily update is down - John will be back on Monday.
Asterix
Heavyweight
Heavyweight
Posts: 7128
Joined: 02 Nov 2007, 10:42

Post by Asterix »

So that explains why things look a little funny. :TU:
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

Daily update is back ...
computerrank
Editor
Editor
Posts: 2492
Joined: 04 Jan 2003, 18:59

Post by computerrank »

Cobwebcat wrote:Site half updated?? Calzaghe moved weight class but also still in old one. Bhop rating unchanged after loss? Joe still behind Bhop...

Matter of time I suspect?
Todays update didn't work, I guess.
gregor
Heavyweight
Heavyweight
Posts: 422
Joined: 27 Feb 2002, 20:00

Post by gregor »

Adamek wins with Bell (who was not ranked) by TKO.... and he drops from 9the to 10th place.

Why? Take a guess... because Bell is back in the ranking, at 3rd position!
JCS
Heavyweight
Heavyweight
Posts: 6243
Joined: 17 Dec 2004, 13:27

Post by JCS »

gregor wrote:Adamek wins with Bell (who was not ranked) by TKO.... and he drops from 9the to 10th place.

Why? Take a guess... because Bell is back in the ranking, at 3rd position!
Read the damn page. The ratings didn't even run last night.
Locked