Ratings - please read before commenting - Archived

[gilgamesh]

Dears,

Why Deontay Wilder is not in this list ?!

https://boxrec.com/en/ratings?ZtQ%5Bcou ... %5D=&r_go=

Thanks
AmirD

Doesn't matter because the list and the criteria for making the list don't matter.

Boxrec's ratings are not now, nor have they ever been accurate.

They're KINDA in the ballpark. That's the best you can give 'em, and sometimes they ain't even that.

[computerrank]

Dears,

Why Deontay Wilder is not in this list ?!

https://boxrec.com/en/ratings?ZtQ%5Bcou ... %5D=&r_go=

Thanks
AmirD

He is #73 - not in the top 50 so far

[gilgamesh]

I didn't say I have a better system. But I sure ain't impressed by the system they got going.

I personally have better ratings yeah. At least at Heavyweight where I still bother anymore.

[gilgamesh]

I didn't say I have a better system. But I sure ain't impressed by the system they got going.

I personally have better ratings yeah. At least at Heavyweight where I still bother anymore.

No public ranking system will please everyone so the only way to decide which is best is via predictive results. The present system, which Computerrank is trying to improve on, has a 76% success rate. I haven’t heard of a better one. It’s not fair to say it’s never been accurate...it has...objectively.

It certainly is fair. Because I've looked at it for years, and it's never been accurate. David Haye and Usyk are not Top 25 All Time Heavyweights, and that's just 1 of many examples I could point out. If I cared to bother, which I don't.

[gilgamesh]

It certainly is fair. Because I've looked at it for years, and it's never been accurate. David Haye and Usyk are not Top 25 All Time Heavyweights, and that's just 1 of many examples I could point out. If I cared to bother, which I don't.

It really isn’t. You can do what you have just done with any ranking system going. You need to test them using a ranking probability score ideally.

Any system does need to make sense to the average eye though so I agree with you there but I’m guessing that’s why he publishes prospective rankings for people to comment on.

“Never been accurate” is a throw away statement without meaning or empirical proof. However most people do the same. If you have put together your own rankings in the past and did so objectively then you will know that if you publicised them the first post would be telling you they were bull crap and inaccurate and you would be right to say...prove it.

Watching Boxing proves it, and on that note. I'm done with you, and this conversation.

[computerrank]

The WHR ratings winner raw prediction acurracy improved from 76.8 percent to 78.7 percent by the more sophisticated implementation now. - based on the ratings points only.

[JCS]

Chasing *only* prediction rate is only good for some type of secondary system or a way to help verify otherwise seemingly arbitrary tweaks. We tried that before... a system based only on prediction rate... many years ago. I thought it was the best way to go about things.. because it was objective and that couldn't be argued against. What happened? It got chased away almost unanimously..

Any of these systems have Beterbiev over Gvodzyk? Basically a 50/50 match in the eyes of oddmakers... No. Why not? Because the data just wasn't there. I mean Beterbiev did finish all of his opponents and never lost.. I suppose you could start looking at trends like that.

Combat sports don't have schedules and matchmaking to support a data-driven, prediction-based only system... and let's not even mention weight divisions and various other things.

[JCS]

I mean.. am I reading the test set right?

Ruiz Jr would be #6 at HW currently?

[computerrank]

I mean.. am I reading the test set right?

Ruiz Jr would be #6 at HW currently?

right

[JCS]

I mean.. am I reading the test set right?

Ruiz Jr would be #6 at HW currently?

right

These may be highly predictive.. but they just aren't "boxing style" rankings.. At absolute worst, he's #3.

[JCS]

These may be highly predictive.. but they just aren't "boxing style" rankings.. At absolute worst, he's #3.

Disagree 6 is right for me and I’ve never liked a rule where a boxer has to jump above one he has just beat either.

This wasn't a fluke win though.. and there was no injury.

Ruiz was a notable contender... Who got up from a knockdown to beat the shit out of the unanimous #1...

I realize the software can't watch the fights.. but in a sport where top guys fight twice a year.. you have to give the utmost respect to the most recent fight.

[computerrank]

right

Is home advantage built into the rankings in any way or is that too difficult to calculate given all of boxing' s other variables not least of which are dodgy decisions?

It is not clear, what home advantage should be.

[computerrank]

Now 79.9 percent RAW winner prediction accuracy for WHR current ratings. With more sophisticated weight division move algorithm.

http://151.boxrec.com/~martin/ratings.php
http://151.boxrec.com/~martin/ratings_at.php

[margaret thatcher]

Awaiting our boi Manny to shut those down

[computerrank]

Now 79.9 percent RAW winner prediction accuracy for WHR current ratings. With more sophisticated weight division move algorithm.

http://151.boxrec.com/~martin/ratings.php
http://151.boxrec.com/~martin/ratings_at.php

Current very good.

All time still poor. Don't tell Froch he's in the top 20 fighters p4p all time or we will never hear the end of it. So many great fighters near the bottom like Ezzard Charles and Roy Jones is sinking without trace.

The basic WHR ratings depend on a division independent strength of a boxer. But this strength naturally is lower for boxers in lower weight divisions.

This lets the weight division rankings and prediction accuracy intact. But I have to introduce a weight division dependent enhancement factor for the P4P ratings. This could be lower or higher. If lower, the boxers in lower weight divisions would rank lower in the P4P ratings.

[Manrae]

Now 79.9 percent RAW winner prediction accuracy for WHR current ratings. With more sophisticated weight division move algorithm.

http://151.boxrec.com/~martin/ratings.php
http://151.boxrec.com/~martin/ratings_at.php

Subjective survey...
Raise your hand if you believe Tom Schwarz is a top 25 heavyweight right now...

Raise your other hand if you believe Tom Schwarz is better than Otto Wallin right now...

Now, objectively, let's look at these 2 men...
Both of these fighters have had 1 major or meaningful opponent in their career. Both of which happen to be the same fighter, in the same year. If I were to ask you "Which guy was dominated and stopped in a few rounds and which guy actually made it a competitive fight in which he won rounds and hurt said opponent?" based solely on looking at their placement of the rankings, which would you choose?

I don't know exactly how to fix it, I'm no math genius at all, but these calculations do not reflect objective reality.

No matter how I look at it, I cannot stretch the realm of plausibility to believe that TS > OW. Perhaps TS beats a name fighter in the future and proves me wrong, but until then...

I'll look at the other weight classes later. But from what I see at heavyweight, there needs to be a tweak to how meaningful bouts are weighed as opposed to regular bouts.

[Manrae]

http://151.boxrec.com/~martin/ratings.php

Let's continue...

175--- Maxim Vlasov over Gvozdyk, Ramirez (who beat him), and Beterbiev (too low anyways)
168--- Jeurgen Braehmer is too high. A win over Brant 2 years ago and everything else being 3-5 years ago
154--- Liam Smith and Michel Soro are a bit too high. Arguable top 10 guys, but not in the top 5
140--- Terry Flanagan in the top 10 at 140. Perhaps his work from 135 got him here, but he's done nothing in significant bouts at 140
130--- Jhonny Gonzalez at #11............ thank you?
108--- Elwin Soto deserves a bit more respect than #19. He's a world champion and he KO'd the champ to become so. No disrespect to Acosta, he's also a great fighter and it was an absolute slugfest that he lost, but he came up short
105--- Sam situation, Wilfredo Mendez needs a little bump

With all of that said, these are the best ratings so far. However, I want to say something about the weight movement variables...

Moving up in weight should be weighted differently depending on what weight changes are being made. For example, a fighter going from 105 to 108 should be weighted differently than a fighter going from 175 to 200. Much more danger the higher you go. Also, there should not be an automatic addition of points for fighters moving down in weight. The vast majority of fighters who move down in weight do NOT have success. The opposite is actually a lot more common. Moving up has its dangers but it has its benefits due to being more hydrated and less focused on making weight.

TLDR-- Moving up should incur a point penalty which depends on from what class they move. Moving down should either also incur a penalty or nothing at all. There should definitely NOT be an addition. Moving down in weight is one of the most dangerous things a boxer can do.

[Manrae]

http://151.boxrec.com/~martin/ratings_at.php

Klitschkos are way too high p4p, they were great heavy----- wait....

Is that Alexander PEDvetkin at #23 all time p4p?

I don't need to continue, that's ridiculous

[computerrank]

http://151.boxrec.com/~martin/ratings_whr.php

I extended the ratings inspection procedures for WHR
- you can click on the boxer ID, and you will get the record
- the record shows the rating points development (current release and WHR)
- at the end of every line you see a value, which indicates the impact of the bout to the boxers WHR rating points, positiv and negativ

So you can inspect, how bouts influence the WHR ratings.

@Manrae: Regarding your indicated cases for the current ratings, I can not find striking inconsistencies in the values shown.

I fixed a bug in the WHR weight division change algorithm. This influenced the WHR all time ratings - @Cobwebcat: divisional all time ratings better now, I guess.

[JCS]

Personally, I still can't accept the fact that Ruiz Jr is so low at HW.

[JCS]

Personally, I still can't accept the fact that Ruiz Jr is so low at HW.

I’m guessing that one of the guys you think should be lower than Ruiz is Joshua as he beat him. But looking at it another way which fighter has the better record? That could be a one off victory. Who do the bookies think will win the return? I genuinely don’t know but I’d guess Joshua.

Just a different way of looking at it I guess.

The only two I could see arguing over Ruiz are Wilder and Fury.

[JCS]

The only two I could see arguing over Ruiz are Wilder and Fury.

Do you think Ruiz will beat Josha in the return?

No clue. Obviously I thought Joshua would crush Ruiz like everyone else, but I was wrong there....

[JCS]

No clue. Obviously I thought Joshua would crush Ruiz like everyone else, but I was wrong there....

Does the fact that you can’t see a clear winner suggest Joshua may be the better fighter overall given that he has just lost to him by knockout?

Yes. However.. just because I think he is better.. doesn't mean he should be ranked higher.

To me.. ranking/rating should be based on recent accomplishment and prediction rate should only be used in the absence of data.

[computerrank]

I’m guessing that one of the guys you think should be lower than Ruiz is Joshua as he beat him. But looking at it another way which fighter has the better record? That could be a one off victory. Who do the bookies think will win the return? I genuinely don’t know but I’d guess Joshua.

Just a different way of looking at it I guess.

The only two I could see arguing over Ruiz are Wilder and Fury.

Joshua lost to Ruiz, but Joshua defeated W. Klitschko, Povetkin, Whyte, Parker, Breazeale, Martin, all of them better than Ruiz' other defeated opponents. And Ruiz lost to Parker. Ruiz has to confirm his result. Until then Joshua may well be recognized a member in an other league.

[JCS]

The only two I could see arguing over Ruiz are Wilder and Fury.

Joshua lost to Ruiz, but Joshua defeated W. Klitschko, Povetkin, Whyte, Parker, Breazeale, Martin, all of them better than Ruiz' other defeated opponents. And Ruiz lost to Parker. Ruiz has to confirm his result. Until then Joshua may well be recognized a member in an other league.

I think this supports Joshua being ranked higher on an all-time list... but not on a current list.

Right now Ruiz is #1 or #3 for me... I'm sure that applies on any respected non-software ranking.... Find me a respected list that has Ruiz lower than #3.