James Jeffries vs. James Corbett I: A Round-by-Round AI Analysis

Discussion in 'Classic Boxing Forum' started by Flavius, Jul 23, 2026 at 6:53 AM.


  1. Andre2000

    Andre2000 Member Full Member

    133
    191
    Jun 26, 2026
    A truly monstrous effort and a very interesting analytical approach.

    Very much appreciated. Thank you.

    One simplified strand of approach could be to view each newspaper as a single judge.

    As such, Corbett would win by split decision in terms of who was winning on the scorecards as at the time of the KO.

    Treated aa a stand alone judge, the newspaper that had Jeffries up by 7 rounds as at the time of the KO would be very much at odds with the other two judges/newspapers and would appear overly lopsided in Jeffries’ favour.

    That newspaper clearly pulled Jeffries right up in the composite stats.

    Using scorecards (including one lopsided card) from an actual fight and combining them could see a much different composite conclusion than held by most observers.

    For instance, for the Hagler - Leonard fight, two judges had it 115-113 apiece. Most might view those cards as reasonably close to the truth of the matter - including those who had it either way but understood that it was a very close fight - even close enough to be to be called a draw.

    Suffice to say, factoring in the crazily lopsided interpretations and 118 - 110 (10 rounds to 2) scorecard of Jo Jo Guerra with the other two scorecards for a combined, averaged out conclusion would only serve to move the fight further away from its true reality.