Lomachenko dominates Compubox Ratings...

[Eff Pandas]

This is the exact conclusion I came to, man. I mean anyone with an iota of sense can see than compubox can give you 100% connect by throwing and landing just 1 punch a round whereas throwing 100 and landing 50 nets you only 50%... which fighter would you consider superior?

LOL thats the exact example I was gonna throw into my post til I noticed I rambled on too long. But yea thats a -50 +/- result & the -50 guy CLEARLY had a great night & the +50 guy did not. +/- is a massively flawed stat.

My system is slightly different from yours but gives a better result (IMHO of course!). Instead of taking the difference in the number of punches landed as you suggest (which would give the same apparent result in a close 550 vs 600 punch fight as it would in a wide 25 vs 75 punch fight) I expressed the number of punches landed / the number of opponents punches landed as a ratio... in the case of Loma vs Rigo this gives a 3.66 in favour of Loma. It's a while since I've crunched any numbers cos I got better things to do with my time these days, but I do remember GGG and Kovalev both having a ratio in the 2.2ish sort of range (averaged over a number of fights) when they were tearing through their respective divisions - despite Kovalev never having particularly impressive +/- numbers. In general anything over about a 2.0 ratio represents a landslide win and anything over about 1.5 is pretty decisive and inarguable..

Anyway just my 2p, man. Make of it what you will.

Hmmm. Interesting. I'm not fully sold on my method myself tbqh & am working it out still, but I think I might be more sold on your method already cuz its simpler first & foremost + seems to have less variance in the figures if what you are saying holds true with most fights + those figures could theoretically tell the story of a specific fight & the fighter overall + vs other fights with other guys in other divisions in easily translated way. I like it

I'm gonna look at some fights stats over the next few days & see what it looks like. Might be hi******* this idea from you lol. Respect.

[Eff Pandas]

It depends what you're using 'em for, man. Like any stats. They're just a resource. If you're part of the fanboy/hater brigade seeking to prove that this is or that guy is the best ever/a bunch of wank then of course it won't settle any arguments - but then what will?

The inaccuracy issue is of course valid, so punch stats produced based on replays are of course preferable, but one would imagine that (contrary to the claims of the paranoid and partisan on NSB) the inaccuracies will tend to average out for the fighters giving a result which ought to be at least relatively representrative in the main. Again I suspect here that your main problem with the stats is that you don't like the way people use 'em. That ain't the fault of the stats though.

This.

I can't believe we still gotta have these dumb talks of "compubox is useless/boxing stats are ******".

I've never heard anyone I respect saying any stat means everything about a fight or fighter. But I've heard too many people saying stats mean nothing. There is middle ground between those two opinions folks.

Boxing stats are merely a tool that says what happened in a fight to the best ability of its human contributors. Its nothing to start a war over or to act like they are completely meaningless. Its merely information that can in theory, & if one can find meaningful numbers or trends, give more information about a fighter, a fight or boxing as a whole than most of the more speculative stuff we talk about 99% of the time that has an even broader range of logic if we are straight shooting about it.