How often do the losers gain more paints and when might it happen?

[gclarkii]

Greetings,

I've been looking through various welterweight boxers records(Why? Because my mentor Harrison "Young" Kennedy was ranked at #112, below people with worse records) and have ran across more than one instance where the loser ends up getting more points the winner, sometimes by a large ratio(e.g. almost 4:1 in the Pruitt vs Soriano match).

I can't seem to find any rim or reason for this happening.
Here are some examples of this:
Harper(96.81) vs Flanagan(105.2) 1954 -- Harper(23-3-0) beats Flanagan(68-8-2) in the 10th round by SD.
Wolak(27.65) vs Smith(19.68) 2018 -- Smith(19-3-0) wins over Wolak(21-0-0) after 10 rounds by UD.
Bundrage(48.82) vs Smith(32.33) 2013 -- Smith(24-5-0) wins over Bundrage(32-4-0) after 12 round by MD.
Bellows(5.389) vs Perez(3.194) 2016 -- Perez(15-5-1) wins Bellows(16-1-1) after 8 rounds by SD.
Berto(45.24) vs Karass(23.78) 2016 -- Karass(27-8-3) wins over Berto(28-2-0) in round 12 by TKO.
Roman(8.798) vs Karass (6.818) 2002 -- Karass(7-0-0) wins over Roman(9-0-0) in the 1st round by KO.
Pruitt(51.94) vs Soriano(13.81) 1971 -- Soriano(51-21-1) wins over Pruitt(43-9-2) in the 6th round by TKO.
And just to show it's not a new phenomena, here is one from 1949, it's also the first that I found where the person getting less points also had the worst record.
Savage(18.19) vs Harper(32.93) 1949 -- Savage(23-16-7) wins over Harper(5-0-0) in the 6th round by RTD.

And what's up with some of them displaying 3 digits(e.g. 5.389)of precision and others displaying 1(e.g. 105.2) or 2 digits(27.65)? They can be mixed within a bout(as displayed in first entry - 96.81/105.2) and/or an individual boxers stats(was seen in the two Karass entries(23.78/6.818). This kind of disparities also show up in a broad range of years with no consistency(2016 has an entries with both 2:3 and 2:2 digits. and 1954 has entries with 2:1, 2:2 and 2:3 digits) and classes,.
I'm a retired programmer(I've done desktop(under MS-Win and Unix with C, C++, and Perl) and web applications(under Unix with C++, Perl, and PHP) and this sort of worries me. To my mind this shows that the these ratings may not have been calculated using the same method.
My reasoning for this is that in PHP, the lines from the same database would be displayed using the same group of code(i.e. the function in question loops over the DB and displays each line as it's retrieved) and the actual displays are ALL over the place. I 've seen .000, .001(both for DEBUT's by the way), .021, X.XXX, X.XX XXX.X. Not good.



I

[computerrank]

Greetings,

I've been looking through various welterweight boxers records(Why? Because my mentor Harrison "Young" Kennedy was ranked at #112, below people with worse records) and have ran across more than one instance where the loser ends up getting more points the winner, sometimes by a large ratio(e.g. almost 4:1 in the Pruitt vs Soriano match).

I can't seem to find any rim or reason for this happening.
Here are some examples of this:
Harper(96.81) vs Flanagan(105.2) 1954 -- Harper(23-3-0) beats Flanagan(68-8-2) in the 10th round by SD.
Wolak(27.65) vs Smith(19.68) 2018 -- Smith(19-3-0) wins over Wolak(21-0-0) after 10 rounds by UD.
Bundrage(48.82) vs Smith(32.33) 2013 -- Smith(24-5-0) wins over Bundrage(32-4-0) after 12 round by MD.
Bellows(5.389) vs Perez(3.194) 2016 -- Perez(15-5-1) wins Bellows(16-1-1) after 8 rounds by SD.
Berto(45.24) vs Karass(23.78) 2016 -- Karass(27-8-3) wins over Berto(28-2-0) in round 12 by TKO.
Roman(8.798) vs Karass (6.818) 2002 -- Karass(7-0-0) wins over Roman(9-0-0) in the 1st round by KO.
Pruitt(51.94) vs Soriano(13.81) 1971 -- Soriano(51-21-1) wins over Pruitt(43-9-2) in the 6th round by TKO.
And just to show it's not a new phenomena, here is one from 1949, it's also the first that I found where the person getting less points also had the worst record.
Savage(18.19) vs Harper(32.93) 1949 -- Savage(23-16-7) wins over Harper(5-0-0) in the 6th round by RTD.

And what's up with some of them displaying 3 digits(e.g. 5.389)of precision and others displaying 1(e.g. 105.2) or 2 digits(27.65)? They can be mixed within a bout(as displayed in first entry - 96.81/105.2) and/or an individual boxers stats(was seen in the two Karass entries(23.78/6.818). This kind of disparities also show up in a broad range of years with no consistency(2016 has an entries with both 2:3 and 2:2 digits. and 1954 has entries with 2:1, 2:2 and 2:3 digits) and classes,.
I'm a retired programmer(I've done desktop(under MS-Win and Unix with C, C++, and Perl) and web applications(under Unix with C++, Perl, and PHP) and this sort of worries me. To my mind this shows that the these ratings may not have been calculated using the same method.
My reasoning for this is that in PHP, the lines from the same database would be displayed using the same group of code(i.e. the function in question loops over the DB and displays each line as it's retrieved) and the actual displays are ALL over the place. I 've seen .000, .001(both for DEBUT's by the way), .021, X.XXX, X.XX XXX.X. Not good.



I

Regarding the Boxrec Ratings you can post here:

viewtopic.php?p=5545482#p5545482

I will answer your post there