MMA-AI.net

This dataset uses ufcstats.com. For each fight, we calculated the technique differential between the opponents and also the fighter's historical average technique differential.

Differential stats are calculated by diving the two individual's technique stats. For example, if Conor hits Khabib 10 times, and Khabib hits Conor 5 times, then Conor's differential stat is 10/5 = 2. Khabib's differential stat is 0.5.

This dataset tracks the following basic stats:

date fight_url event_url result fighter opponent division stance dob method total_comp_time round time fighter_url opponent_url referee time_format reach height age knockdowns sub_attempts reversals control takedowns_landed takedowns_attempts sig_strikes_landed sig_strikes_attempts total_strikes_landed total_strikes_attempts head_strikes_landed head_strikes_attempts body_strikes_landed body_strikes_attempts leg_strikes_landed leg_strikes_attempts distance_strikes_landed distance_strikes_attempts clinch_strikes_landed clinch_strikes_attempts ground_strikes_landed ground_strikes_attempts KO_losses days_since_last_comp lose_streak win_streak win_loss_ratio total_comp_time stamina num_fights trueskill elo 

It then parses those stats starting at reach and creates several new categories for each:

stat_differential stat_per_min stat_accuracy stat_def stat_peak stat_valley stat_vs_peak stat_vs_valley 

Then it further modifies those categories above with:

avg_stat recent_avg_stat 

Then for each of those it does:

change_stat stat_vs_opp 

Exploring the data

See all of Jon Jones' stats df.loc[df['fighter'] == 'jon jones']

See all of Jon Jones' striking differential stats df.loc[df['fighter'] == 'jon jones'].filter(like='sig_strikes_differential')

See Jon Jones' postfight Elo scores and who his opponents were df.loc[df['fighter'] == 'jon jones'].loc[:, ['opponent', 'elo']]

See Jon Jones' postfight ground control time, average ground control, and peak ground control df.loc[df['fighter'] == 'jon jones'].loc[:, ['opponent', 'control', 'avg_control', 'control_peak']]

See Jon Jones' prefight average ground control vs his opponent's average ground control df.loc[df['fighter'] == 'jon jones'].loc[:, ['opponent', 'precom_avg_control_vs_opp']]