Ratings - please read before commenting - Archived

[JCS]

A couple of related notes:

1) JMM is above Chris John despite John remaining undefeated and JMM only beating Terdsak Jandang since losing to him.

2) Vlad Virchis is above Chagaev despite Chagaev remaining undefeated and Virchis only beating Vidoz since losing to him.

If you're not trying to make a predictive system anymore, I feel like you should be required to beat more than some lower-level guy you were expected to beat anyway to vault someone over someone you've lost to recently.

I wonder if Sergei should still be above Lamon?


I can see the difficulties in adding such barriers.. but it should probably be done.

[jujigatame]

It's tough to say, I think The Ring's current ranking of Briggs-Sergei-Lamon in the 5-6-7 spots is perfectly valid, but I don't really mind BoxRec's system pushing Sergei down so low since he hadn't really done much before he beat Brewster, and Briggs was not even in the top 25. I just think that more should be done to adhere to boxing's age old "if you beat the man, you are the man" system. I realize this could be difficult to program. Is there anything in place that makes it harder to advance, the closer to the top you are?

[JCS]

It's tough to say, I think The Ring's current ranking of Briggs-Sergei-Lamon in the 5-6-7 spots is perfectly valid, but I don't really mind BoxRec's system pushing Sergei down so low since he hadn't really done much before he beat Brewster, and Briggs was not even in the top 25. I just think that more should be done to adhere to boxing's age old "if you beat the man, you are the man" system. I realize this could be difficult to program. Is there anything in place that makes it harder to advance, the closer to the top you are?

No, its all a numbers game... doesn't matter where you are, you advance the same. It all depends on who you beat, how you beat them, and where the fighters are rated around you.

In this type of system, I think you'd agree Sergei shouldn't have been knocked entirely out of the Top 20.

Either way, I've argued your point before and I think its a very valid argument. The whole "where should Sergei" type of point is another one.. but its a little more complex I do believe.\


BTW - Anyone notice Brian Viloria is 29th!! We should just bring the predictor back online :)

[computerrank]

you're making all the brits mad since Hatton isnt in Top 25 P4P!!

Shows what a nonesense the current system is.

Also PBF will probably retire undefeated and be about 79th in the all time list behind some complete bums because he went thru the weights!

Where's Conan now?!

PBF currently is not in the division, where he achieved most.

He will be assigned to his top division after he retired - and then will rank much higher.

[JCS]

I still say we need to implement some kind of protection criteria. If it hinders prediction rate, so be it. That is not even the focus any longer.. only a guidance statistic.

[JCS]

Martin,

I think the hometown factor might be a bit too aggressive now. Guys like Felix Sturm and Brian Viloria are really getting their rating trashed by outsiders.

Also, I'd like to hear your thoughts on the "protection" factors.. in a scenario that involves fighters like Briggs/Liakhovich and Brewster.

Most "experts" would rank them Briggs before Liakhovich before Brewster. Because of the close proximity of their fights, and the chronological analysis of the fights.

Liakhovich beats Brewster
Briggs beats Liakhovich (Brewster does nothing in meantime)

[computerrank]

Martin,

I think the hometown factor might be a bit too aggressive now. Guys like Felix Sturm and Brian Viloria are really getting their rating trashed by outsiders.

Also, I'd like to hear your thoughts on the "protection" factors.. in a scenario that involves fighters like Briggs/Liakhovich and Brewster.

Most "experts" would rank them Briggs before Liakhovich before Brewster. Because of the close proximity of their fights, and the chronological analysis of the fights.

Liakhovich beats Brewster
Briggs beats Liakhovich (Brewster does nothing in meantime)

Jason,

1. Home Advantage

The home advantage impact is not as big as you might think. the home clear decision loser would lose

- 47% of his points with current release version
- 40% of his points in the last release version
- 34.5% without home advantage

But I think, the home advantage factor is quite correct with the assumption that it makes the home boxer a 2 by 1 favorite, if they have equal boxrec ratings.

2. Protection

This would introduce this odd A def B def C def A issue. Keep it simple - we cannot solve this problem anyway.

Best regards
Martin

[JCS]

Martin,

I think the hometown factor might be a bit too aggressive now. Guys like Felix Sturm and Brian Viloria are really getting their rating trashed by outsiders.

Also, I'd like to hear your thoughts on the "protection" factors.. in a scenario that involves fighters like Briggs/Liakhovich and Brewster.

Most "experts" would rank them Briggs before Liakhovich before Brewster. Because of the close proximity of their fights, and the chronological analysis of the fights.

Liakhovich beats Brewster
Briggs beats Liakhovich (Brewster does nothing in meantime)

Jason,

1. Home Advantage

The home advantage impact is not as big as you might think. the home clear decision loser would lose

- 47% of his points with current release version
- 40% of his points in the last release version
- 34.5% without home advantage

But I think, the home advantage factor is quite correct with the assumption that it makes the home boxer a 2 by 1 favorite, if they have equal boxrec ratings.

2. Protection

This would introduce this odd A def B def C def A issue. Keep it simple - we cannot solve this problem anyway.

Best regards
Martin

I still think a 2 to 1 home advantage is a bit severe. I mean look what happened to Viloria and Sturm after their losses. Both fell out of the Top 25.

I think we should deliberate harder on the protection since this is now a "traditional based" rating system.

[computerrank]

With the smaller home advantage = 40% advantage instead of 100%:

Viloria and Sturm would have 20% points more, which would only help them with about 5 ranks.

[jomothepure]

Sorry to intrude, but can someone tell me why JC Gomez has just entered the heavyweight rankings again? I thought after a year of innactivity a fighter drops out of the rankings, he doesn't enter them?

[JCS]

When Hatton fights next it will be back at light welterweight.

What would be his rating at that weight now. Presumably he would be above Castillo and therefore in top 20 P4P?

I Believe he would fall in at #3. Perhaps a dominant win then pushes him up to #2, possibly in the Top 25 P4P.

[JCS]

How can he be 2nd at Welterweight, move DOWN a weight, and be ranked even lower???!

I supposed Hatton wasn't in the top 25 as he was being recorded outside his normal weight. If returning to LW makes no difference then what's going on?

What does he have to do?! He's in everyone's top 10 but doesn't appear in Boxrec's top 25?? What is Computerrank doing?

The points you see are based on division.. not on P4P. This is a limitation (and a stupid one) which has not been resolved.

As fighters move through divisions, they lose points. Hatton's case deviates from the norm, as he went from 140, to 147, back to 140. Unfortunately (I still believe this is true) he'll lose points each time.

This is hurting his P4P standing, as those rankings are based on divisional points.. which is a flaw. Now, the divisional points are normalized based on the available talent, etc., but I still dont see how this makes the P4P proper.

So in summary, fighters who move divisions on a regular basis are hindered in the P4P rankings.

[JCS]

In summary then the P4P list penalises fighters who prove themselves at different weights.

This system gets more and more bizarre everytime I look at it. So if PBF beats ODH he will move down as he has moved up weights?

Beyond stupid. I've lost all faith in these ratings. What a shame. I thought Conan had some good points but the way they have been interpreted by the new system is ludicrous.

Now you see my argument all along. You cannot make a perfect computerized system, so why not strive for something different?

This system has become a hybrid of its prediction-based predecessor anyhow.. So now we have to deal with it just because some small-minded, short-sighted, whining, complaining individuals had to keep bitching.

[mattyp151]

In summary then the P4P list penalises fighters who prove themselves at different weights.

This system gets more and more bizarre everytime I look at it. So if PBF beats ODH he will move down as he has moved up weights?

Beyond stupid. I've lost all faith in these ratings. What a shame. I thought Conan had some good points but the way they have been interpreted by the new system is ludicrous.

I wouldn't exactly call Rick proven at 147...more luck extremely lucky...but anyway, I was going to ask the same thing about Oscar/PBF.

[computerrank]

Boxers' points are transformed in relation to the square of the divisions' weight limits, when they move. So they lose no points, just for moving up and down.

[computerrank]

But that's the point. You shouldn't lose any points for moving DOWN a weight!!!

You win points moving down, and lose points moving up - this means in relation.

You are at the pointage after moving up and down - and no effective bout in between.

So there is no inherent loss of points, just by moving.

[JCS]

That doesn't seem to be what JCS is saying. Is he wrong then?

What's Hatton's likely rating back at LW then? He was top before he moved up to welter..

I think what he's trying to say is that he's trying to normalize points when a fighter moves between divisions based on the talent, etc, at that division. Technically, their points would change, but only in relation to the division as a whole.

I still think Hatton would enter at 3rd.

[pundit]

Why does Lacy lose -- and Tsypko win -- points for a Lacy win over Tyspko?

[JCS]

Why does Lacy lose -- and Tsypko win -- points for a Lacy win over Tyspko?

Because he won by the skin of his teeth.. which proves these two are closer together in rating than pre-fight thinking would've suggested.

Not to mention that Lacy had a hometown advantage so that's factored in as well.

Makes sense to me anyway..

[pundit]

Why does Lacy lose -- and Tsypko win -- points for a Lacy win over Tyspko?

Because he won by the skin of his teeth.. which proves these two are closer together in rating than pre-fight thinking would've suggested.

Not to mention that Lacy had a hometown advantage so that's factored in as well.

Makes sense to me anyway..

Makes sense, but how does a computer know this?

[JCS]

Why does Lacy lose -- and Tsypko win -- points for a Lacy win over Tyspko?

Because he won by the skin of his teeth.. which proves these two are closer together in rating than pre-fight thinking would've suggested.

Not to mention that Lacy had a hometown advantage so that's factored in as well.

Makes sense to me anyway..

Makes sense, but how does a computer know this?

Scorecards

[pundit]

Because he won by the skin of his teeth.. which proves these two are closer together in rating than pre-fight thinking would've suggested.

Not to mention that Lacy had a hometown advantage so that's factored in as well.

Makes sense to me anyway..

Makes sense, but how does a computer know this?

Scorecards

That's an impressive level of detail.

Other question: when a figher moves up the rankings, does this also affect the fighters who beat him recently or narrowly lost?

[JCS]

Makes sense, but how does a computer know this?

Scorecards

That's an impressive level of detail.

Other question: when a figher moves up the rankings, does this also affect the fighters who beat him recently or narrowly lost?

No.

Thats one of those things though, that you can heavily argue for, or against.

[probert24]

... as he has only 1 annual rating in welterweight 2005, he is just #205 til now in welterweight.

In lower divisions his ranking would be much higher, but the editors have him at welterweight now - which is quite correct.

This and later posts tell me why I can't find Hopkins and why Virgil Hill is so low career-wise in the alltime list.
So, would it be possible to to make a guy's career ranking include the division below and above where he rests at. For example; Virgil Hill's career cruiser ranking would take into effect fights between light heavy and heavy - albeit none at heavy but it would give a better idea of his standing in total.

One detraction with only taking into account one weight class is it would seem to benefit older fighters who didn't have the opportunity to bounce up or down the 3-25lbs needed to change divisions as they were not available.

[probert24]

Can Iran Barkley be moved out of heavyweight? 1201st all time is pretty weak, especially when DM Qawi is listed as a light heavy.