Rankings are an important but controversial part of boxing.  I began to list my own rankings on myWomen's Boxing Page web site in 1998 because I saw signs of favoritism in the rankings done by some boxing federations. For example, boxers who were about to fight on the cards of certain promoters appeared to rise in the rankings for no other reason, while boxers who had fought for titles of other federations were sometimes ignored.  Some boxers who had long retired, and even a few who had never fought, were being ranked ... while active boxers who didn't have the "right" connections weren't.

Since 1998 I've produced my own independent rankings, based solely on the fight results. I'm not connected to any boxer, trainer, manager, promoter or federation in any way. I am a reporter on, and supporter of, this sport who's been in contact many of its people since I started the Women's Boxing Page web site in 1997.  

For over ten years I have compiled a database which now has over 7000 fight results for over 3000 female professional boxers worldwide and is as complete as possible for pro fights since 1996.  Since 2005, Sue TL Fox and I have maintained this database together --- it is the basis for all of the boxer and fight results data pages on the WBAN Records Site.   

The number of pro female boxers and pro women's bouts happening per year makes ranking all the boxers "by hand" a pretty daunting task. In 2001 I realized that it was getting hard to be sure I was completely consistent when reviewing so many results for so many boxers, so I began to explore ways to have the computer do the rankings directly from my database.  I've learned that this approach has both advantages and disadvantages, which I'll describe below.  

Pros and cons of computer rankings 

The advantage of using a computer program for rankings is that it can automatically and impartially assess every fight result. A computer program has no favorites --- it is guaranteed to apply its rules to all boxers equally.  Once a set of ranking rules is established, those rules will be applied impartially to every boxer and every fight in the database.  

The disadvantage of computer ranking is that it can't account for subtleties that humans can see in the fights. It uses ONLY the actual fight results. As I don't have the detailed scorecard for every fight, my computer program uses only information that's available to me for EVERY fight: number of rounds, KO or TKO results,  evidence of closeness in the fight from split or majority decisions, or from draws. 

There are also nuances that could go into a ranking done by a knowledgeable person, e.g., ignoring blatantly unfair decisions, which the program doesn't do (yet).

So I do not claim that any computer program can be a panacea for women's boxing rankings.  But the complete impartiality and uniformity of computer rankings should make them interesting to anyone who wants a system that treats all boxers the same way, regardless of their country or affiliation, when assessing their records.

Since I joined WBAN in April 2004, Sue Fox and I have discussed how to use computer rankings to provide an independent WBAN ranking system, and the computer program now includes rules that we have both agreed on to determine  eligibility for being ranked (based on activity by the boxers and on setting minimum computer scores in order to be ranked )  

Here are some details on how the computer rankings are produced:

Data Quality 

For any computer ranking program work at all, its database must be as complete as possible for every fighter, and it must record her name exactly the same way every time.  

This isn't as easy to do as it sounds!  Not all sources of fight results are equally careful about checking and reporting the boxers' names.  So we cross-index the WBAN database by weight class and boxer name in various ways to weed out any inconsistencies. The WBAN database is also regularly checked against the data at Boxrec.com.  I believe that both of these databases are similarly complete for fight results, especially a few weeks  after the fights, but there may be differences in their spellings of fighters' names, or other fight details.  If you see any errors, omissions or inconsistencies in any ranking list produced from the WBAN database, please send me an email, and I'll investigate and make changes as soon as possible when needed.      

Weight classes

A difficult issue with rankings for pro female boxing is that even the top boxers may take fights in several different weight classes in order to keep active and to get enough good fights.  There aren't as many females as males in each weight class, and promoters are often more stingy about paying expenses for female boxers.  A female boxer is more likely to fight out of her preferred weight class to get a fight at a good time or place than is a comparable male boxer.

The computer program requires that each boxer is ranked in just one weight class.  Where there's a lot of choice I will assign her to the class that she's fought in most often recently, or (if it's different from that) the one in which she holds what I think is her most reputable title belt, or (if it's different yet again and I think it's justified) the weight class that she or her management have personally asked me to list her in.  (Boxers: if you would like to be ranked in a weight class other than the one I'm ranking you in now, please send me an email).

How it's done

The computer program's rules are sophisticated and not easily described in just a few lines or formulas. An in-depth account of is given here; what follows is just a summary ...

This program doesn't just look at won-lost records, but it rates the "importance" of every fight based on both boxers' fight records. It learns who the tough opponents are and it gives bigger scores for winning big (competitive) fights and smaller scores for wins over easier opponents.  In that sense, it ranks the fights as well as the fighters.  (This principle is essential to make a rankings system work well!)

This program does its rankings in two stages. The first stage makes a "short list" of rankable boxers. The second stage adjusts the scores for the rankable boxers so that the results of the most recent head-to-head fights between them determine the final ranking order. This second stage is particularly important to "rising stars" of the sport -- it guarantees progress up these rankings whenever a boxer defeats higher-ranked opponents, with the most recent results getting absolute precedence. 

The listings

WBAN lists the "top 10" boxers from my computer rankings in each weight class, with their numerical ranking scores. It also lists my full rankings  and recent "relevant fights" for each weight class, on the "details" pages.

These listings are updated monthly, close to the beginning of the month.  Rankings from earlier months are also archived on WBAN.

What's next

The ranking program is still being tweaked, but now I have been running it for years I keep any changes small to ensure month-to-month consistency. I'll be happy to have your feedback about any oddities that you see coming out of it.

I've tried to make this ranking system as fair, and as responsive to recent results, as I can, and I've been encouraged by how its results have been received by the women's boxing community.  Several women's federations now use these rankings as the basis for theirs, and I have given them permission to do so --- but that is my only connection with those federations.  I do not accept ... and never will ... any compensation from anyone for doing these rankings.  No boxing federation, promoter or manager has, or will ever have, any direct influence over them.   

However, I am open to suggestions for how to improve the general rules inside the computer program. I welcome input of that sort from anyone who is interested in women's boxing. My only requirement is that future changes to the rules must be ones that can be applied to all boxers and to all fights in the database, in order to preserve the program's fairness and impartiality. Please send me any suggestions or comments about the ranking system by email.

© Dee Williams 2008

Page last updated: Friday, 04 April 2008