Ratings - please read before commenting - Archived

[computerrank]

What I'm saying is that using WHR is a bad idea, full stop. If you want to create a boxing handicapping system, it's great. If you want to create something that looks like traditional (The Ring, BLH, ESPN) boxing rankings, it is not. Hence, why you've already had to make modifications to it to placate BoxRec users, because maximizing predictivity creates a set of rankings that people will simply not use (except maybe for gambling) or accept.

jujigatame,,

I see WHR is not your's, it's a bad idea from your perspective. But I am convinced the boxing community will accept a rating not because it is traditional, but because it matches with the competitiveness of the boxers. And competitiveness can be measured by matching bout result prediction and bout result.

Everything else at the end is unfounded talking about things with no means to get it verified in any way. And BoxRec wants to have it verified, not only declared.

[jujigatame]

The idea of a mathematically "verifiable" ranking system is a pipe dream. It is an impossibility, and striving towards it is ultimately a waste of your time. You can see this in the way you have already had to modify the WHR system in ways that reduce predictivity. Maximizing predictivity leads to a system that the boxing community will never accept.

[computerrank]

The idea of a mathematically "verifiable" ranking system is a pipe dream. It is an impossibility, and striving towards it is ultimately a waste of your time. You can see this in the way you have already had to modify the WHR system in ways that reduce predictivity. Maximizing predictivity leads to a system that the boxing community will never accept.

jujigatime,

The two different kind of prediction ratios, you mention, are apples and oranges, do not compare them.

The first one weighs all bouts the same.

The second one weighs bouts with close rated boxers higher. As those bouts are more difficult to predict, the resulting prediction ratio naturally is lower. But that says nothing about the quality of the second kind of ratings in relation to the ratings of the first kind.

So what you can decide is, do I want to look at all bouts weighed equal - or do I want to look at all bouts weighed by the closeness of the ratings of the boxers involved.

But you cannot mix that up. The second concept has a lower prediction ratio, and the two kinds are not comparable by value.

And in addition, I fear you are caught in a trap to avoid to be involved in a gambling or betting concept, setting equal predicting and gambling or betting Yes, gambling and betting is connected to prediction. But verifying ratings is connected to prediction too, and that is not belittled by betting and gambling using the same concept.

[jujigatame]

No, I simply don't agree that any type of predictivity is a useful primary metric for evaluating boxing rankings. I don't think it ever has been. I've only followed boxing for ~20 years but every time I looked at a credible set of rankings, they were based around who HAS beaten who, not who WOULD beat who. There is a correlation there but they are still quite different.

Once you put high importance on predictivity you almost by definition must start overvaluing some fighters with weak resumes, because there will be many instances (especially with prospects) when a weak resume will defeat a stronger one. This can have undesirable effects throughout the system, which I believe is what we're seeing here.

[jujigatame]

On a sidenote specifically pertaining to Tank and Nakalitha, is it possible that the system is simply overvaluing KOs? Like is a KO of LSC significantly more valuable than a decision over Lomachenko? This would also have benefited Nakalitha as he KO'd a ton of marginal opponents in Africa.

[computerrank]

No, I simply don't agree that any type of predictivity is a useful primary metric for evaluating boxing rankings. I don't think it ever has been. I've only followed boxing for ~20 years but every time I looked at a credible set of rankings, they were based around who HAS beaten who, not who WOULD beat who. There is a correlation there but they are still quite different.

Once you put high importance on predictivity you almost by definition must start overvaluing some fighters with weak resumes, because there will be many instances (especially with prospects) when a weak resume will defeat a stronger one. This can have undesirable effects throughout the system, which I believe is what we're seeing here.

jujigatame,

It is a common solution advice 'rank as who has beaten whom'.

But the realization of that concept leads to nowhere in such a ladder system. Set the winner just above the loser. But beyond that, that system lets you alone with more questions than answers.

#1 what to do, when the winner will not fight appropriate opposition after that (quite common in proboxing, systems like WHR will compensate for that)
#2 what about obvious erratic upsets (systems like WHR will compensate for that)
#3 what about weight division moves (systems like WHR will compensate for that)

The common solution advice 'rank as who has beaten whom' in it's essence is empty of content. Or it leads to a ranking team, 'improving' those rankings - and then deal with 20,000 active boxers - look at all the ABC, The Ring, BLH, ESPN, any boxing board rankings - never ....

Or it leads to a points reward system, where you reward points for defeating a rank. But i lets you alone too with questions #1 #2 #3.

Any such system will fail because the main issue with proboxing is an essential lack of appropriate matched competition.

[jujigatame]

I think the traditional BoxRec system (similar to the system used by FightMatrix for MMA) solved those problems just fine. And it had the added bonus of being comprehensible to any person who wanted to understand how any particular ranking came to be. The importance of that should not be understated.

[Manrae]

I think that I found a bug

https://boxrec.com/en/ratings?r%5Brole% ... 5D=a&r_go=

#8 Kareem Martin ranked way too high... unless I'm missing something

Not a bug:
Martin defeated Ananyan and Ananyan defeated Matias within the last 18 months. So the BoxRec winner above loser rule is applied.

The pure WHR ratings would be:
- Matias 66.66
- Ananyan 11.07
- Martin 2.915

Winner above loser rule makes:
- Martin 66.66
- Ananyan 55.66
- Matias 54.97

I see...

[margaret thatcher]

what would gerovnta davis be rated if he were in the 140 pound rankings

[computerrank]

what would gerovnta davis be rated if he were in the 140 pound rankings

about 10 % less points there

[margaret thatcher]

so #1 by a big margin.....you've got a bit of a tank davis affair going on with the ratings, the system absolutely loves him

where would he be rated at 147 and 154?

[computerrank]

so #1 by a big margin.....you've got a bit of a tank davis affair going on with the ratings, the system absolutely loves him

where would he be rated at 147 and 154?

77% and 67% of Lightweight points. But it makes no sense to do that across large differences between weight classes.

[margaret thatcher]

tank #1 at 154 baby

[computerrank]

tank #1 at 154 baby

... but Alvarez will stop him ...

[Manrae]

tank #1 at 154 baby

... but Alvarez will stop him ...

I'm sorry, I was going to let it go but... something needs to change

https://boxrec.com/en/ratings?r%5Brole% ... 5D=a&r_go=

Just look at #15... 4-0... 4 round scheduled fights.... #15 in the world....

https://boxrec.com/en/proboxer/912048

[computerrank]

tank #1 at 154 baby

... but Alvarez will stop him ...

I'm sorry, I was going to let it go but... something needs to change

https://boxrec.com/en/ratings?r%5Brole% ... 5D=a&r_go=

Just look at #15... 4-0... 4 round scheduled fights.... #15 in the world....

https://boxrec.com/en/proboxer/912048

It is the BoxRec winner above loser rule:
Jorge Armando Sanchez Hernandez defeated Mario Gutierrez
Mario Gutierrez defeated Daniel Valladares

pure WHR ratings:
Valladares 6.312 at Minimumweight
Gutierrez 0.460 at Flyweight
Hernandez 0.050 at Light Flyweight

BoxRec winner above loser rule makes
Valladares 4.796 at Minimumweight ( = 4.407 at Light Flyweight)
Gutierrez 4.037 at Flyweight ( = 4.502 at Light Flyweight )
Hernandez 5.801 at Light Flyweight

[Manrae]

... but Alvarez will stop him ...

I'm sorry, I was going to let it go but... something needs to change

https://boxrec.com/en/ratings?r%5Brole% ... 5D=a&r_go=

Just look at #15... 4-0... 4 round scheduled fights.... #15 in the world....

https://boxrec.com/en/proboxer/912048

It is the BoxRec winner above loser rule:
Jorge Armando Sanchez Hernandez defeated Mario Gutierrez
Mario Gutierrez defeated Daniel Valladares

pure WHR ratings:
Valladares 6.312 at Minimumweight
Gutierrez 0.460 at Flyweight
Hernandez 0.050 at Light Flyweight

BoxRec winner above loser rule makes
Valladares 4.796 at Minimumweight ( = 4.407 at Light Flyweight)
Gutierrez 4.037 at Flyweight ( = 4.502 at Light Flyweight )
Hernandez 5.801 at Light Flyweight

I understand that... but when a 4 round fighter, with only 4 fights (3-0-1) is ranked #15 in the world in a division over former world champions and such...

How does this not ring any alarm bells for you?

Please consider adjusting or getting rid of the rule

[JCS]

Oh that's a pretty cool one!

[jujigatame]

Has any thought been given to what I asked about KOs?

Tank's point total has gotten pretty outrageous. He now has ~400 points at 140 pounds while his biggest win was over a guy with 88 points at 130 pounds.

Assuming I don't have the power to convince you to abandon WHR, this seems like the top thing to look at. It would explain why Tank is rated so outrageously high, why Nakalitha was overrated to such a degree, and maybe even why Crawford inexplicably has 1200+ points while Spence is stuck at ~700.

Another serious oddity is Kell Brook's rating somehow RISING from 2016 to 2020.

[computerrank]

It is the BoxRec winner above loser rule:
Jorge Armando Sanchez Hernandez defeated Mario Gutierrez
Mario Gutierrez defeated Daniel Valladares

pure WHR ratings:
Valladares 6.312 at Minimumweight
Gutierrez 0.460 at Flyweight
Hernandez 0.050 at Light Flyweight

BoxRec winner above loser rule makes
Valladares 4.796 at Minimumweight ( = 4.407 at Light Flyweight)
Gutierrez 4.037 at Flyweight ( = 4.502 at Light Flyweight )
Hernandez 5.801 at Light Flyweight

I understand that... but when a 4 round fighter, with only 4 fights (3-0-1) is ranked #15 in the world in a division over former world champions and such...

How does this not ring any alarm bells for you?

Please consider adjusting or getting rid of the rule

It wasn't my decision, it was the decision of the BoxRec board. They say, it is an iron law in proboxing - the winner has to be ranked above the loser - at least for a certain time period.

I would prefer to list the pure WHR ratings, which would require that upsets have to be confirmed by further achievements in such cases before such winners can pass their opponents.

[computerrank]

I'm sorry, I was going to let it go but... something needs to change

https://boxrec.com/en/ratings?r%5Brole% ... 5D=a&r_go=

Just look at #15... 4-0... 4 round scheduled fights.... #15 in the world....

https://boxrec.com/en/proboxer/912048

It is the BoxRec winner above loser rule:
Jorge Armando Sanchez Hernandez defeated Mario Gutierrez
Mario Gutierrez defeated Daniel Valladares

pure WHR ratings:
Valladares 6.312 at Minimumweight
Gutierrez 0.460 at Flyweight
Hernandez 0.050 at Light Flyweight

BoxRec winner above loser rule makes
Valladares 4.796 at Minimumweight ( = 4.407 at Light Flyweight)
Gutierrez 4.037 at Flyweight ( = 4.502 at Light Flyweight )
Hernandez 5.801 at Light Flyweight

I understand that... but when a 4 round fighter, with only 4 fights (3-0-1) is ranked #15 in the world in a division over former world champions and such...

How does this not ring any alarm bells for you?

Please consider adjusting or getting rid of the rule

There is an option. The winner above loser rule only says, that the winners in such a queue have to be above the losers - somewhere inside their current maximum and minimum WHR ratings.

Currently in case of an upset the winner is set to 100% of the loser's WHR rating and the loser is set to 10% of the difference lower.So a big upset winner can get that high as here.

An option would be to set the upset winner to only to 70% of the difference and the upset loser to 10% of the difference lower.

Then Hernandez would only be at #61 at Light Flyweight and not #15 as now.

And Valladeres at the end of the queue would be dumped from #11 Minimumweight now to #39 Minimumweight then.

[JCS]

Winner above loser rule is part of the traditional ranking recipe... I don't think you should remove this.

The problem comes when there are chains of multiple boxers and split/majority decisions.... That is when things get strange.

[computerrank]

Has any thought been given to what I asked about KOs?

Tank's point total has gotten pretty outrageous. He now has ~400 points at 140 pounds while his biggest win was over a guy with 88 points at 130 pounds.

Assuming I don't have the power to convince you to abandon WHR, this seems like the top thing to look at. It would explain why Tank is rated so outrageously high, why Nakalitha was overrated to such a degree, and maybe even why Crawford inexplicably has 1200+ points while Spence is stuck at ~700.

Another serious oddity is Kell Brook's rating somehow RISING from 2016 to 2020.

I looked into that.

It is not the KO/TKOs as the bouts are over 12 rounds and a clear win on points then would have the same weight.

It is more the point already discussed above. Bouts between close rated boxers should be weighed higher in the ratings calculation than bouts between very uneven rated boxers.

The ratings Davis / Lopez change from 440 / 235 to 119 / 92, or ratio 1.87 to ratio 1.29.

And Crawford / Spence change from 1248 / 697 to 158 / 204 or ratio 1.79 to 0.77!

While Nakalitha is quite where he was #10.

[computerrank]

Winner above loser rule is part of the traditional ranking recipe... I don't think you should remove this.

The problem comes when there are chains of multiple boxers and split/majority decisions.... That is when things get strange.

I won't remove that (BoxRec board is watching ), just rearrange the pointage section where the queue of boxers is sorted in.

[jujigatame]

Has any thought been given to what I asked about KOs?

Tank's point total has gotten pretty outrageous. He now has ~400 points at 140 pounds while his biggest win was over a guy with 88 points at 130 pounds.

Assuming I don't have the power to convince you to abandon WHR, this seems like the top thing to look at. It would explain why Tank is rated so outrageously high, why Nakalitha was overrated to such a degree, and maybe even why Crawford inexplicably has 1200+ points while Spence is stuck at ~700.

Another serious oddity is Kell Brook's rating somehow RISING from 2016 to 2020.

I looked into that.

It is not the KO/TKOs as the bouts are over 12 rounds and a clear win on points then would have the same weight.

It is more the point already discussed above. Bouts between close rated boxers should be weighed higher in the ratings calculation than bouts between very uneven rated boxers.

The ratings Davis / Lopez change from 440 / 235 to 119 / 92, or ratio 1.87 to ratio 1.29.

And Crawford / Spence change from 1248 / 697 to 158 / 204 or ratio 1.79 to 0.77!

While Nakalitha is quite where he was #10.

But don't you think there is some inherent problem in the fact that a fighter can attain 400 points at 140 when his biggest win is over an 88 point fighter at 130? This to me seems like a massive red flag.

Perhaps the problem is not KOs in particular (if a KO is weighted equally to a dominant 120-108 type win) but the difference in weighting between dominant and decisive (but not dominant) wins.