How can Boxrec Rankings change to actually work?

Gotcha, Gaz.

I'd say they are at least trying something different and it completely removes all accusations of bias towards individual fighters.

It does throw up some crazy ones, I'll give you that.

 

Cheers mate. I apologize if any of my posts in this thread came across as a personal "attack" (for want of a better word) on the guys behind the system. That wasn't my intent at all.:good

 

I think that 1 reason you get some strange rankings using their formula is that boxers compete so infrequently compared to other sports with much more accurate individual rankings like tennis, snooker or chess.

The more competition, the more often the formula is used & the mean result from all of these will be closer to a 'real' rating that more people would agree with.

I also think that the gain/losses are far too high from individual results & should somehow be reduced by factoring in their recent records. For example if you have won your previous 5 fights then lose, that shouldnt lead to a rating reduction as large as someone who has a worse recent record against similar competition. Paul Williams dropped a lot after a single loss to Martinez despite his previous high standing.

Currently, the ratings as they are calculated favour unbeaten guys over guys with good records & the occasional loss against good opposition. This accounts for a lot of the difference betwenn the common perception of how good a fighter is & their boxrec rating.

 

I agree with you Gaz and I'm not saying it's as staright forward as the D&L method but my point remains, although we dont have the formula it does exist and it's just a matter of time before somebody finds it, the fact we dont know what it is shouldn't slow our pursuit.

Imagine what cricketers before the D&L method was around thought.

 

imo ranking top 10s is hard enough and causes enough arguements so i think a formula to rate boxers in the top 300 makes sense. As long as you take the rankings with a pinch of salt they have as much value as others

 

Too subjective. I don't think a computer can do it.

You need honest people to decide the rankings based on fights in the ring, not pumped up records, down payments from promoters, what country the org is based in etc

 

I don't see it this way. I could be wrong though.

If you're 20-0 against tomato cans, you're nowhere. If you're 10-10 against decent blokes you should be better off. You get **** all for wins against lads with no points.

 

Don't you mean 'objective'? The computer removes all subjectivity (other than that which is in the formula).

 

You lose so many points for losing the ratings as calculated by boxrec favour guys with an 0. Who is the top active fighter? (Floyd though Manny has a much better resume) There are lots of unbeaten guys in the top 25s of their divisions who have fought no-one.

 

Of the top 10s in 17 divisions, 44 are currently unbeaten. So around 1 in 4. I'm not making any judgment on this - but that's what the current situation is.

 

I'll take your word for it

I'm making a generalisation but an example would be Chisora (14-0 other than Sexton has beat no-one) ranked #9 & Arreola is #10(32-2).