A points ranking system to legitimise boxing rankings and championships!!

The TBRB responded to my idea, and I'll post my rebuttal here. It's essay-length and too long to fit in one post so I've broken it up into four posts. I think I covered every argument that can be made against a points system. Again, please let me know what about this system you think wouldn't work, or pledge your support


–There has yet to be a points-based system, computerized or otherwise, to produce a set of boxing rankings that stands the test of reason or common sense. What you identify as “anomalous” rankings are usually far too common to be considered anomalous. The IBO has Adrien Broner as the #1 welterweight with its computerized system. BoxRec has Lucian Bute #3 at light heavyweight. No one in their right minds believes that those rankings reflect anything like reality. The examples are fairly legion.


The IBO and Boxrec systems are so flawed that they can’t be used as evidence against a points ranking system. But your two examples are poor cases even against them!

You’re actually mistaken about Broner. The IBO lists four other alphabet titlists on top of their rankings, but they in fact exclude titlists from those organisations from their rankings. But I agree with them that Broner shouldn’t be highly ranked. With my system there would actually be no chance that Broner would be in the top-10, where you undeservedly have him, and he wouldn’t even be ranked above Malignaggi. One win in a new division shouldn’t mean a boxer should be ranked over the boxer he beat. I’m sure you’ll say that isn’t necessarily the case with the TBRB rankings, but I think in a fair system it should almost never happen.

It’s difficult to find details about the IBO’s ranking criteria, but their most heavily weighted factor seems to be a boxer’s last six fights. I think a time period rather than specific number of fights has to be used: activity should be rewarded, and a boxer who otherwise would have the same ranking as another should likely be ranked higher if he fights more often. There are also many inaccuracies in their rankings, for example ranking Gamboa number 1 at lightweight. They probably allow fights in other divisions to count towards a ranking, which shouldn’t be allowed. But their rankings are innacurate because they use an unfair standard, not because they use measurable standards. I think if the IBO champions weren’t a complete joke and they allowed other titlists to be ranked their rankings wouldn’t be significantly worse than yours.

As far as Bute’s Boxrec ranking, though I can’t be sure my system could very well have Bute with a top-20 ranking at light heavy. His last fight was a win in a fight in which both boxers were over the super middleweight limit so it would be considered a light heabyweight fight, and Grachev would definitely be top-30 at light heavy, so Bute would earn ranking points. If Juergen Braehmer’s best opponent Eduard Gutknecht wasn’t in the top-30 Bute’s win could even earn him a top-10 ranking. My system would definitely not allow Bute to be as high as number 3 however.

Boxrec’s rankings seem to be not too different from the IBO’s. So again while I like that they’re using an objective method, their method results in inaccurate rankings because they’re using factors that shouldn’t be used. The Boxrec system is overambitious and uses too much information. It assigns and scales ranking points based on method of victory, the scorecards of a decision, and the ranking of his opponent’s opponents, amongst other confusing factors. Their system somehow gives Wilder a number 13 ranking, but as his only opponent with a remote chance of being in the top-30 would be Liakhovich, Wilder would at best be barely top-30 with my system. It’s not fair to lighter punchers to award more ranking points for knockouts. A ranking should be determined only by wins and losses and opponent rank.

There has been one points system that produced fair rankings. The Boxing Tribune had a sliding points system like mine, but it was discontinued because it didn’t get enough support and only didn’t prove itself because it didn’t last long enough to do so. Only when a more logical points system than Boxrec’s or the IBO’s is given a chance can anyone say it won’t work. My system of course would reqire over three years to prove definitively that it would work. Like any other political change, it can’t happen overnight! There only hasn’t been significant interest in a points-based system because it hasn’t been tried!

 

–Part of the reason is because different kinds of systems work for different kinds of sports. While a points-based rankings system might work for tennis, for example, that sport is fundamentally less subjective in how a winner is determined than boxing, i.e. points are points, and while their might be disputes about when a ball is “in,” all in all a hard set of objective figures determines a victor. In boxing, scoring is significantly more subjective. (We actually did have discussions about how tennis rankings work when we were forming our board, and rejected them.) What that means is that sometimes a winner isn’t really the winn in boxing in the court of public opinion and doesn’t prove he’s better except to three judges who might be 100 percent at odds with everyone else watching, which is why we have a robbery clause. In other words, an official decision win in boxing as ruled by a judge is, exactly, the “opinion poll” you are criticizing — an opinion poll of three people.


I’m not attempting to apply anything from the tennis system that isn’t relevant to boxing. I just think the idea of ranking based on the last three year’s accomplishments (compared with one year in tennis), and on quality of opposition (compared with levels of success in different levels of tournaments in tennis) is sensible. I didn’t take anything else from the tennis system. Don’t get hung up on the tennis comparison, but many tennis fans felt Rafael Nadal shouldn’t have been ranked number 5 and David Ferrer number 4 after the French Open, because Nadal beat Ferrer in the final and was on a streak of winning almost every tournament the last six months. But Ferrer was ranked lower because Nadal was injured and missed tournaments less than one year ago, while Ferrer had maintained consistent success throughout the year, and Nadal won the French Open the previous year so wouldn't gain points, while Ferrer improved on his previous year success. Nadal’s potential, ‘common sense’ that he’s better, and his star-power don’t have a place in tennis rankings, and those qualities in boxers shouldn’t in boxing rankings either.

Yes, there is subjectivity in judging, and in this boxing isn’t analogous to tennis. But though it’s unfortunate an objective ranking system has to accept the subjectivity of judging as definitive. That you have a ‘robbery clause’ is actually the one stated rule of yours that I vehemently disagree with. A bad decision should have to be accepted. Overturning decisons for ranking purposes is a slippery slope. If you do that you might as well rank a boxer higher if you thought he deserved a 119-109 decision rather than a 116-112 decision. If you respect a close decision that you think should be wider, than you should respect a decision that was given to the boxer you thought lost. For ranking purposes a decisive line has to be drawn, and that line should be to respect the official decision.

I think it’s worthwhile looking to other sports for a fresh perspective. Boxing over the years has been far too inward-looking. I’m sure boxing fans would laugh to compare boxing with figure skating, but though I don’t know anything about it it’s a judged sport, and I was able to find out that except for a 2002 decision reversal olympic figure skating has always accepted judges decisions. Boxing rankings have to consider a win a win.

 

–Wouldn’t the system you describe mean that in some cases, a fighter who beats another could still be ranked beneath the fighter he just beat? As in, Darren Barker has faced two top opponents, and beaten one of them, for x amount of points. Daniel Geale has faced a greater number of top opponents, and beaten more of them. In your system, wouldn’t Barker be ranked below Geale despite beating him? And if he wouldn’t in this case, isn’t that a potential outcome? Also, wouldn’t someone who was used repeatedly as an “opponent” to top fighters be ranked more highly than a boxer with a couple quality wins, merely for fighting ranked contenders more often? You raise a point about arbitrariness — isn’t their arbitrariness in the length of time you choose, for, say, the period of a fighter’s career measured?


Yes, looking at their recent records it's very possibile that Geale would be ranked above Barker with my system. One win should not necessarily rank a boxer over the boxer he beat most recently. Geale has beaten at least three opponents in the last three years that my system would likely consider top-30, while the only likely top-30 opponent Barker fought in the last three years would be Martinez. It’s a consistent problem even with rankings outside of the alphabet organisations that the most recent fight is given far too much weight over a boxer’s body of success over the last year or few years. It’s possible that a boxer who just beat a higher ranked boxer would earn a higher ranking with my system, but it should not be almost guaranteed that this will happen. The TBRB too often ranks a slightly less accomplished boxer one spot above a more accomplished boxer after he beats him.

My system might seemingly award more points than deserved to boxers used as opponents by top contenders. I think I’ll try to test this soon by using past results based on a sample top-30, and if it would make for more accurate rankings losses could deduct ranking points from a boxer. But I suspect that won’t be necessary, and it might unfairly punish losses. With my system a boxer would go down in the rankings if he lost while boxers ranked directly below him beat top-30 ranked boxers. You might think my system would give a journeyman like Darnell Boone too much credit for losing to top-30 boxers, but a ranking system that deducts points for losses could conversely allow a boxer with a padded record but only one or two victories over top-30 opponents to surpass boxers who had fought more top-30 contenders. It’s imperative to my system that losses to top opponents are not punished, because I think a fair system needs to reward quality of opposition.

The three-year period is actually not arbitrary at all. To clarify, it’s not three calender years, but to be fair to all boxers it would be the previous exact three years. For example, a rankings update on the 1st of December 2016 would be for fights that took place from the 1st of December 2013 to the 30th of November 2016. Points would be lost as boxers ranking points go below the ranking points they had three years ago. For example if a boxer fought lower ranked opponents than he did three years ago, or if he lost to a top-contender but beat a similarly ranked contender three years ago. Three years seems fair to me because it would include enough fights to give an accurate reflection of a boxer success, but it wouldn’t go back so far that an older boxer’s prime or a younger boxer’s wins over unranked opponents were included.

 

–While there are elements of opinion poll to what we are doing, it is not strictly that. The rankings are actually decided through a vigorous debate where Board members state their case and try to (and in some cases do) win over other Board members. The chairs and oversight can act as arbiters and tiebreakers, which makes this more of a “checks and balances” system than a pure opinion poll. As for bias: The only way to assess that is to look at our rankings for evidence of it. While some have criticized elements of our rankings, we accept that because no one will ever 100 percent agree with every single ranking we end up with; it’s simply impossible. However, we have not been accused of any bias to date and I think it would be hard for anyone to demonstrate there was any. Whatever prejudices or varying degrees of awareness our Board members have, they have all averaged out to produce a set of rankings that are highly credible.


Regardless of how vigorous the debate or how informed the members contributing to the board might be, it is still reached by combined personal judgement, which will leave potential for corruption at worst and confusion at best. I suppose you admit any inherent psychological bias, but it’s not sufficient to claim that a group will average this out. All members of the board will naturally have a much deeper awareness of boxers who are more visible and frequently fight on television. They’ll have a more informed view of those boxers, and because they would have less knowledge of less visible boxers, the more visible boxers would often be favoured and ranked higher than their actual accomplishments should deserve.

The TBRB’s high rankings of Mares, Darchinyan, and Juergen Braehmer are examples of this.

Mares should not be number 3 at featherweight. He’s had just one fight at featherweight. His two wins over top junior featherweights and his resume at bantam could keep him in the top-5 in both of those divisions with my system, but no way does he deserve the high ranking at featherweight you give him. He would earn significant ranking points with my system for his one quality win, but wouldn’t get a ranking of more than about 10, which should be where any fair ranking would have him. He would only have a ranking of higher than that if he’s given credit for potential or his marquee name.

It’s even worse to have Darchinyan ranked number 6 at junior featherweight with just two fights in the division. Neither win was against a proven contender in the division, one was against a prospect, the other against a much lower level boxer. Both of them might have ranking points with my system but most likely not enough to be top-30 ranked contenders. Darhinyan could only have this ranking if he was credited for past success in other divisions.

As I suggested when I compared his ranking to Bute’s, Juergen Braehmer is also too highly ranked at number 9. He shouldn’t be ranked quite as high because he’s only fought regional-level opposition. He’s fought three opponents who might be considered top contenders, but two of them five and eight years ago. He’s only ranked where he is because he’s gained top visibility in Germany for beating almost only overmatched Germans and holding one of the alphabet titles that you should be ignoring. He should be top-20, but Campillo, Grachev, Chilemba, and indeed Bute should definitely be ranked above him, so he shouldn’t be ranked in the top-10.

I’ll add that your system of oversight only shows that your rankings are at best less biased, it does not completely remove bias. I think it’s outrageous that three chair members are given power as arbiters. That you have three individuals as final decision makers is exactly why I described the TBRB as an ‘oligarchy’. Similarly, I’m strongly against an American-style presidential system of government, as the executive branch which was supposed to be a fair arbiter has become so all-powerful that democracy in the United States is now a sham (I guess this shouldn’t get too political, but I’m a leftist anarchist and I’m against all forms of hierarchy).

If I’ve made a compelling case… confessing your mistakes and scrapping the TBRB would leave a great opportunity for an objective, points-based ranking system to be put into practice and have an impact towards truly reforming boxing.

 

My article at the excellent Boxing Tribune..

http://theboxingtribune.com/2013/08/friendly-fire-opinion-poll-boxing-rankings-must-go/

 

It's great to hear that someone else not only agrees that an objective ranking system is needed, but has tried to make one too!

I'm on the west coast and I'm exhauted because I've been up all night, but I'll get back with more detail later today. I hope we can keep a conversation going. I think as many people who agree on the general idea of such a points ratings system should debate together to try to organise the best system. What matters is that results in fair and accurate rankings. If either you or I find that what we propose would not result in the most accurate rankings we should be willing to change. I really don't think knockouts should be given more weight though. But if this can be done, let's take our time so it's done right, then got it accepted and viable!

 

-Where there’s a fatal problem with your rankings proposal, as I see it, is with your refusal to rank a fighter who beat another fighter above that fighter immediately after that win. If you’re ranking someone by who’s better than the next guy, what more conclusive evidence do you need than that he actually beat that guy? It’s hard to take seriously the notion that Barker isn’t better than Geale and deserve to be ranked above him, given that he just proved he was better than him in the ring.

I'm sure you'll be disappointed to know that with my system there could actually be very occasional circumstances when a boxer ranked above another would move BELOW the lower ranked boxer after beating him. This could only happen if the two were close in ranking points but the winner managed to lose more points from exactly three years ago than the winner. But this would not be an anomaly that should be prevented, but proof of the legitimacy of the system. I mentioned that the ATP ranked Nadal number 4 and Ferrer number 5 before the French Open, but after Nadal beat Ferrer in the final Nadal moved below him in the rankings.

My system wouldn't refuse to move a lower-ranked boxer above a higher-ranker boxer he beat, and I never said anything to suggest that. I only said that a win wouldn't necessarily move the winner above the other. With my system Donaire would likely have moved above Nishioka when he beat him, Bellew would likely have moved above Chilemba when he beat him, and Mikey Garcia would have moved above Juanma Lopez when he beat him (though not before). But this likely wouldn't happen more often than not, because the higher ranked boxer going in would have earned his ranking. One win shouldn't be given predominant weight in a ranking. Earning a higher rank should depend on consistent success and quality of opposition. It's opinion-based rankings which as a rule rank the winner above the boxer he beat, like Barker above Geale, which lack credibility.

Your suggestion that rankings should indicate who's better than the next guy is problematic and misunderstands the purpose of rankings. Rankings should strictly be a reflection of performance. A boxer shouldn't get a higher rank simply because of potential.

 

Darchinyan is probably the most overranked boxer of any division in your rankings! His only win against a top-15 opponent since he beat Arce well over four years ago was against Yonnhy Perez. He's lost handily to every better fighter than that he's fought in that period. His record above 115 is just 5 and 4. I already mentioned that of his only two fights at junior feather one was against a nobody, but what led me to believe that it was Darchinyan's visibility that was a factor in your ranking of him was that what likely got him to number 6 in your rankings was ONE win against an unproven prospect in Del Valle. Upsets of prospects when they first step up are not too uncommon. But your rankings suggest that if it wasn't an established name on a fairly heavily promoted, HBO-televised event, he would not earn the ranking Darchinyan did. I could come up with more examples, but Jonathan Maicelo was surely a similarly rated prospect, but Rustam Nugaev was not given a prominent ranking at lightweight when he upset him. On top of Darchinyan's lack of recent accomplishments, he couldn't even be given credit for potential, because he's severely flawed and shown he can't handle elite technical boxers, and he's past his prime.

It's also no evidence that Darchinyan's visibility isn't credited to say a boxer who's fought 20 fights in the division over the last four years but is less visible is ranked higher by you. But what shows that Darchinyan is overranked, and likely overranked because of his higher visibility, is that he's ranked above many more proven junior featherweights. All of the following generally less visible boxers who you rank below Darchinyan or not at all should be ranked above him: Scott Quigg, Kiko Martinez, Jorge Arce, Wilfredo Vazquez, Takalani Ndlovu, Rendall Munroe, Jhonathan Romero, Alejandro Lopez, Chris Avalos.

Another boxer that seems overranked because of visibility is Gamboa at lightweight. Though Gamboa's top 20-ish opponent Darley Perez was better than Darchinyan's opponent, it was just one unimpressive win, but Gamboa seems to have been given the benefit over more deserving established lightweights because he's a big name fighting in a big big event.

In the case of Braehmer, I was trying to make the point that relative dominance in any area of the world when exposed on television can give a distorted impression of the quality of a boxer. My point was that to Germans or anyone around the world who's able to watch can only see what mass media presents to them. Boxing is popular outside of the US, so indeed heavily promoted and exposed domestic talent can be presented as and perceived to be of a higher class than they actually are. Your German and European board members especially may have overranked him based on his visibility. Campillo, Chilemba, and Grachev have all had more success against higher-level opposition and should be ranked above Braehmer. Indeed Bute, and possibly Erdei, with your 'robbery clause' from his official loss to Grachev, should possibly also be ranked above him.

 

Mares losing tonight kind of proved one win at a weight shouldn't outweigh three years of consistent success in the rankings!!! Mares had a quality win previously, but Gonzalez showed why his body of success probably would rank him above Mares going in, and proved that my three year period is the best length of time to go with.

 

With my system Liakhovich would be the only opponent of Wilder's who would have minimal ranking points, though he still wouldn't be in the top-30. So Wilder would have fought no one in the top-30, would have no points, and wouldn't in the top 50. My simpler system is therefore more sensible than boxrec's, which allows KOs over mismatched opponents to contribute towards ranking.