Post by db1972 on Jan 21, 2020 16:57:57 GMT -6
I was a little bored today, so I took some time to examine UE's pace of play and efficiency metrics (all data from Kenpom against D-1 opponents). This is a simple analysis containing only 4 variables: Offensive Efficiency, Defensive Efficiency, Pace, and Result from all of UE's 18 D-1 games so far this season. If I have more time and there is demand, perhaps I will go more in depth to look at more specific statistics or players rather than a general overview.
Anyway, enough with the boring details, here's what I found. UE's offensive efficiency is directly linked to UE's pace. For every extra possession in a game, UE's offensive efficiency increases by .0177 points per possession on average. You might be thinking that .0177 doesn't sound like a lot, but it adds up over 65+ possessions in a game. Basically what this means is that the faster UE plays, the better the offense looks. I think most teams post better offensive numbers in transition than in half court, but UE takes it to the extreme this season. UE is 7-1 in games with a pace of over 70 possessions per 40 minutes, but only 2-9 in games with a pace of 69 possessions or fewer. Our average offensive rating in games with a 70+ pace is 108.17, while it is a paltry 87.98 in games with a 69- pace. For a little bit of perspective, the 108.17 mark would be good for 50th in the country overall, while 87.98 would put us 347th in the country overall out of 353 teams. Basically, we need to play in transition or our offense belongs in the SWAC.
Please let me know if you have any questions about what I did, or how to interpret it, or if you would like me to answer a specific question you have. I will do my best to answer the forum's questions about analytics as well as I can since I believe analytics are an extremely important part in understanding how good teams and players actually are.
*** DISCLAIMER ***: I did not attempt to make an adjustment for missing McCarty or DeAndre in any of these analyses. Every game without McCarty falls in the 69- category, as does every game without DeAndre.
*** Heavy math section***: For all the math nerds out there, the only variables with a statistically significant correlation to each other were pace and offensive efficiency with a p-value of .0377, which is why I mostly focused on that. Offensive efficiency was statistically significant in predicting the result as well, with a p-value of .0311. Pace was almost significant at predicting the result, with a p-value just outside of the acceptable range at .0545. When using Pace and Offensive Efficiency together to predict the result, however, neither one was significant as they are highly correlated with each other. There is no significant correlation between defensive efficiency and any of the other metrics which I examined. I am also looking at raw, unadjusted data for each of our games, whereas the overall offensive efficiencies listed on Kenpom are adjusted for the opponents' defensive strength. Given that a lot of our best offensive games came against bad defenses, and our worst offensive games came against good defenses in the Valley, 108.17 is probably a slight overstatement, while 87.98 is probably a slight understatement. Adjusting for defenses certainly does not erase a 20 point efficiency gap, though, so I believe the validity of my results still stands.
Anyway, enough with the boring details, here's what I found. UE's offensive efficiency is directly linked to UE's pace. For every extra possession in a game, UE's offensive efficiency increases by .0177 points per possession on average. You might be thinking that .0177 doesn't sound like a lot, but it adds up over 65+ possessions in a game. Basically what this means is that the faster UE plays, the better the offense looks. I think most teams post better offensive numbers in transition than in half court, but UE takes it to the extreme this season. UE is 7-1 in games with a pace of over 70 possessions per 40 minutes, but only 2-9 in games with a pace of 69 possessions or fewer. Our average offensive rating in games with a 70+ pace is 108.17, while it is a paltry 87.98 in games with a 69- pace. For a little bit of perspective, the 108.17 mark would be good for 50th in the country overall, while 87.98 would put us 347th in the country overall out of 353 teams. Basically, we need to play in transition or our offense belongs in the SWAC.
Please let me know if you have any questions about what I did, or how to interpret it, or if you would like me to answer a specific question you have. I will do my best to answer the forum's questions about analytics as well as I can since I believe analytics are an extremely important part in understanding how good teams and players actually are.
*** DISCLAIMER ***: I did not attempt to make an adjustment for missing McCarty or DeAndre in any of these analyses. Every game without McCarty falls in the 69- category, as does every game without DeAndre.
*** Heavy math section***: For all the math nerds out there, the only variables with a statistically significant correlation to each other were pace and offensive efficiency with a p-value of .0377, which is why I mostly focused on that. Offensive efficiency was statistically significant in predicting the result as well, with a p-value of .0311. Pace was almost significant at predicting the result, with a p-value just outside of the acceptable range at .0545. When using Pace and Offensive Efficiency together to predict the result, however, neither one was significant as they are highly correlated with each other. There is no significant correlation between defensive efficiency and any of the other metrics which I examined. I am also looking at raw, unadjusted data for each of our games, whereas the overall offensive efficiencies listed on Kenpom are adjusted for the opponents' defensive strength. Given that a lot of our best offensive games came against bad defenses, and our worst offensive games came against good defenses in the Valley, 108.17 is probably a slight overstatement, while 87.98 is probably a slight understatement. Adjusting for defenses certainly does not erase a 20 point efficiency gap, though, so I believe the validity of my results still stands.
