Ratings - please read before commenting - Archived

[JCS]

Its tough to get a stranglehold on the lowest of the low, while still doing well for the fighters in the spotlight. You almost need separate systems.

[KOin5]

KOin5, I'm also British. The P4P system is more beneficial to boxers who dominate a weight class than boxers who have moved through multiple weigh categories quickly.

Surely someone who dominates a weight class has acheived less than someone who hops between divisions.

And the boxers who acheive most are generally the best.

And if the boxers who are best are best p4p...

Then why isn't Pac top?

There, is an obvious flaw in the system ^_^

[Asterix]

KOin5, I'm also British. The P4P system is more beneficial to boxers who dominate a weight class than boxers who have moved through multiple weigh categories quickly.

Surely someone who dominates a weight class has acheived less than someone who hops between divisions.

And the boxers who acheive most are generally the best.

And if the boxers who are best are best p4p...

Then why isn't Pac top?

There, is an obvious flaw in the system ^_^

Yes, the P4P chart is very flawed. It should be taken down, imo, as it detracts from the divisional charts (which are very good).

[JCS]

It should be renamed to division dominance, because that's what it is.

[hitman09]

It should be renamed to division dominance, because that's what it is.

Thats one way of measuring P4P though isn't it? Perhaps the only objective method

Not when some divisions are weaker than others.

[conan_the_cribber]

Ratings at the other of the scale.

I was wondering if there might be scope for improvements at the other end of the scale i.e. down amongst the tomatos.

Some thread showed the Zelenoff vs Hartley clip and so I went scanning for the ranks of these fighters. Both were quoted as unranked, which is understandable for Zelenoff, but questionable for Hartley. I mean, the guy has had 15 or so fights now and so that's quite a data set to rank him with.

Anyhow, I went to the ratings link and selected all the U.S Welters. Sure enough, all the welters are there, with their points total. Many have the same points total, and yet seem to be sorted in a reasonable order, so I guess there is an unpublished number helping this sorting.

Anyhow, amongst the curiosities that I saw.

Carrine Hammlet http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer has a draw and a SD loss, so I'm surprised he landed on 0 points

William Deboe Thomas http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer who has only fought twice, and lost both, ended up a) with a rank and b) with 3 pts. Probably the MD is valued higher than the SD, whereby, the SD was actually better becuase Hammlet won on one judges card, whereas Thomas did not.

Josh Durden http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer is also on 0 pts, despite a run of 5 wins and 0 losses. Admittedly, he has fought the worst of the worst, but nevertheless, 5 wins and 4kos against this opposition would mean to me, that he is probably a tick better than the other 0pt fighters.

Rafael Estradahttp://www.boxrec.com/list_bouts.php?hu ... &cat=boxer same goes for Rafael who has at least some wins (including a win against someone who won a fight) and a draw against Sook what's his name, who is obviously not complete dirt either.

Getting back to the original two fighters. Intuitively I’d say, that someone who beats someone , even if it is a debuting fighter, is generally better than them. I’d rank Hartley above Zelenoff. Naturally you shouldn’t be able to rocketing up the ladder by KOing a bunch of 0-0-0 or 0-1-0 fighters, but I’d still give you at least 1 pt for the win. And I’d probably rank someone higher, who has beaten 5 such debuting fighters as someone who has beaten just one.

Probably all this is reflected in a mystery hidden number somehow. Still I think some improvements are possible.

conan

Any comments computerrank? Surely improving the bottom ends up improving the top at some point.

[John]

I think computerrank is on holiday for the next week or so

[conan_the_cribber]

I think computerrank is on holiday for the next week or so

thanks john. i'll let it rest a while.

cheers

conan

[computerrank]

@conan

Only full points are presented in the ratings. But the ranking (look at the rank independent of rated or not) depends on the part points too.

All boxers below 1 point are 'not rated'.

Hartley vs Zelenoff

Hartley 0.010
Zelenoff 0.001

But both below 1 and so unrated.

Carrine Hammlet 0.631 - 4 rounder (only 1/3 points compared with 12 rounder) against opponent with only 6 points

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2007-08-03|Light Middleweight  |Hamlett             |Carrington          |L |UD |    0|    0|    0|    0|    0
2008-03-08|                    |Hamlett             |O'Hare              |D |MD |    0|    0|    0|    0|    0
2008-04-26|Light Middleweight  |Hamlett             |Woods               |L |UD |    0|    0|   23|   23|    0
2008-09-19|Super Middleweight  |Hamlett             |Seivert             |L |SD |    0|    0|    6|    5|    0

William Deboe Thomas 3.352 - 4 rounder against opponent with 60 points

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2008-10-10|Welterweight        |Thomas              |Dolton              |L |UD |    0|    0|   27|   27|    0
2008-11-14|Lightweight         |Thomas              |Lester              |L |MD |    0|    3|   60|   57|    0

Josh Durden 0.001

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2006-11-11|Welterweight        |Durden              |Batts               |W |KO |    0|    0|    0|    0|    0
2007-03-24|Light Welterweight  |Durden              |Johnson             |W |TKO|    0|    0|    0|    0|    0
2007-09-28|                    |Durden              |Fairfax             |W |KO |    0|    0|    0|    0|    0
2007-11-17|                    |Durden              |Wiggins             |W |TKO|    0|    0|    0|    0|   99
2009-01-10|Lightweight         |Durden              |Shakir              |W |PTS|    0|    0|    0|    0|    0

Rafael Estrada 0.605

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2004-03-03|Middleweight        |Estrada             |Moreno              |W |TKO|    0|    0|    0|    0|    0
2004-03-26|Middleweight        |Estrada             |Hill                |D |PTS|    0|    0|    0|    0|    0
2004-04-10|Middleweight        |Estrada             |Silvey              |NC|ND |    0|    0|    0|    0|    0
2004-06-02|                    |Estrada             |Bartley             |W |TKO|    0|    0|    0|    0|    0
2004-06-25|Welterweight        |Estrada             |Boldridge           |W |UD |    0|    0|    0|    0|    0
2004-08-26|Welterweight        |Estrada             |Hinds               |L |KO |    0|    0|    0|   30|    0
2005-05-26|Welterweight        |Estrada             |Collins             |W |TKO|    0|    0|    0|    0|    0
2008-04-19|Welterweight        |Estrada             |Siyajuck            |D |MD |    0|    1|    5|    4|    0

[conan_the_cribber]

@conan

Only full points are presented in the ratings. But the ranking (look at the rank independent of rated or not) depends on the part points too.

All boxers below 1 point are 'not rated'.

Hartley vs Zelenoff

Hartley 0.010
Zelenoff 0.001

But both below 1 and so unrated.

Carrine Hammlet 0.631 - 4 rounder (only 1/3 points compared with 12 rounder) against opponent with only 6 points

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2007-08-03|Light Middleweight  |Hamlett             |Carrington          |L |UD |    0|    0|    0|    0|    0
2008-03-08|                    |Hamlett             |O'Hare              |D |MD |    0|    0|    0|    0|    0
2008-04-26|Light Middleweight  |Hamlett             |Woods               |L |UD |    0|    0|   23|   23|    0
2008-09-19|Super Middleweight  |Hamlett             |Seivert             |L |SD |    0|    0|    6|    5|    0

William Deboe Thomas 3.352 - 4 rounder against opponent with 60 points

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2008-10-10|Welterweight        |Thomas              |Dolton              |L |UD |    0|    0|   27|   27|    0
2008-11-14|Lightweight         |Thomas              |Lester              |L |MD |    0|    3|   60|   57|    0

Josh Durden 0.001

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2006-11-11|Welterweight        |Durden              |Batts               |W |KO |    0|    0|    0|    0|    0
2007-03-24|Light Welterweight  |Durden              |Johnson             |W |TKO|    0|    0|    0|    0|    0
2007-09-28|                    |Durden              |Fairfax             |W |KO |    0|    0|    0|    0|    0
2007-11-17|                    |Durden              |Wiggins             |W |TKO|    0|    0|    0|    0|   99
2009-01-10|Lightweight         |Durden              |Shakir              |W |PTS|    0|    0|    0|    0|    0

Rafael Estrada 0.605

Code: Select all

date      |division            |boxer               |opponent            |rs|dec|box_0>box_1|opp_0|opp_1|h_adv
----------|--------------------|--------------------|--------------------|--|---|-----|-----|-----|-----|-----
2004-03-03|Middleweight        |Estrada             |Moreno              |W |TKO|    0|    0|    0|    0|    0
2004-03-26|Middleweight        |Estrada             |Hill                |D |PTS|    0|    0|    0|    0|    0
2004-04-10|Middleweight        |Estrada             |Silvey              |NC|ND |    0|    0|    0|    0|    0
2004-06-02|                    |Estrada             |Bartley             |W |TKO|    0|    0|    0|    0|    0
2004-06-25|Welterweight        |Estrada             |Boldridge           |W |UD |    0|    0|    0|    0|    0
2004-08-26|Welterweight        |Estrada             |Hinds               |L |KO |    0|    0|    0|   30|    0
2005-05-26|Welterweight        |Estrada             |Collins             |W |TKO|    0|    0|    0|    0|    0
2008-04-19|Welterweight        |Estrada             |Siyajuck            |D |MD |    0|    1|    5|    4|    0

Hi computerrank,

Thanks for the behind the scenes info. A lot more makes sense, however I am left with the following question marks.

Josh Durden 5-0 (against unknown scrubs) and Charlie Zelenoff 0-1 (against ultimate known scrub Andrew Hartley) both have 0.001 Pts?

Hartley who has only beat two 0-0-0 fighters and lost a lot more, is ranked above Durden.

That seems a little unfair on Josh. I mean, Hartley has proved that he very beatable, yet Charlie couldn't do it. That has to rank Chrlie lower than Durdan who has at least won five times against opponents, who have not yet become so low as Andrew Hartley.

(@Andrew, if your reading, all respect to you, you're just an example here, not having a go at you).

I dunno how I'd do it, but I'd probably start by doing a first pass of the database, looking for absolute zeros. Fighters who in their career never won or drew a match. These fighters can barely contribute to another figthers record. Then you do a second normal sequential pass.

When you fight such a marked fighter, you get say 1pt for beating the first one,0,5pts for beating the second one, 0,25 pts for beating the third. So you can max out at 2pts, for beating 20 0-0-0 fighters in a row.

More later. Important it to cheat and do two passes of the db.

cheers

conan

[computerrank]

@conan

more in detail and corrected for a decimal error (Hartley):

Durden 0.0010050
Hartley 0.0010005
Zelenoff 0.0009999

[computerrank]

New release launched - bugs fixed:

- scores for decisions on points - bouts with scores from referee and judges were not considered correctly

- weight division change without limit by last rating in division returned to

[Sledgehammer]

I joined this forum just to try to figure out the rating system. I have obtained alot of good information, but would like to ask, how would a fighter that was tring to move up in the ratings do so. I'm a mid level fighter. looking to move up in the most efficient mannor, and would appreciate any advise.

[jujigatame]

I don't think there's any secret or trick to it. Just try to fight the highest ranked guys possible, and win.

[Asterix]

I don't think there's any secret or trick to it. Just try to fight the highest ranked guys possible, and win.

Yeah, pick the guys with the most points who you think you can beat by knockout. Try to keep as active as possible.

[Sledgehammer]

Thank you guys. I appreciate it.

[conan_the_cribber]

Thank you guys. I appreciate it.

Oh and fight rated guys. Don't pad your record with unranked guys. You can see in a boxer's profile on boxrec if they are ranked or not.

conan

[conan_the_cribber]

Hi computerrank,

continuing on with my bottom trawling. I looked at some US heavies.

US Rank 311 Freddie Miller http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer got his first win by UD against an 0-0-0 opponent.

However, he lands ahead of

US Rank 315 Brandon Mayweather http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer who got his first win via KO1 against a different 0-0-0 fighter. I would've assumed that Mayweather's win would've been worth more as it is more conclusive.

And I'm surprised that both land ahead of

US Rank 319 Newbie Jackie Saye http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer who has three KO wins against lacklustre opposition. One of the three wins is against a 0-0-0 guy, so effectively Saye has achieved what Mayweather and Miller have achieved and then some more, yet he is ranked behind them.

Any clues on that.

thanks

conan

[computerrank]

@conan

US Rank 311 Freddie Miller http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer got his first win by UD against an 0-0-0 opponent.

=> should be 0.0010009 - is 0.001

However, he lands ahead of

US Rank 315 Brandon Mayweather http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer who got his first win via KO1 against a different 0-0-0 fighter. I would've assumed that Mayweather's win would've been worth more as it is more conclusive.

=> should be 0.0010009 - is 0.001 - UD and KO are the same value

And I'm surprised that both land ahead of

US Rank 319 Newbie Jackie Saye http://www.boxrec.com/list_bouts.php?hu ... &cat=boxer who has three KO wins against lacklustre opposition. One of the three wins is against a 0-0-0 guy, so effectively Saye has achieved what Mayweather and Miller have achieved and then some more, yet he is ranked behind them.

=> should be 0.0010027 - is 0.001

The bug will be fixed with the next release.

[JCS]

Martin,

Are you still factoring unsanctioned bouts at half value?

[computerrank]

Martin,

Are you still factoring unsanctioned bouts at half value?

Yes - I do!

[JCS]

Martin,

Are you still factoring unsanctioned bouts at half value?

Yes - I do!

I am worried about the Thailand results...

Only the high profile fights get set to unsanctioned (like Wonjongkam v Miranda), while many of the mid-level fights do not. I was informed by Scott Mallon that all Thailand bouts should be to unsanctioned.

However, this would have quite an impact on some fighters' ratings.. like Pongsaklek.

I think this should be reconsidered...

[JCS]

How do you feel about this, Martin?

[computerrank]

How do you feel about this, Martin?

As I can see, nearly all bouts in Thailand and Philippines are unverified - this is not the same as unsanctioned.

Only unsanctioned bouts are accounted with 50% only.

[computerrank]

Computerrank: Can you have a look at the all-time light welter rankings please. No 13 Len Davies does not look like he should feature so high looking at his record.

I will see ...