EXPLANATION OF DATA AND
Kentucky Opens Against NKU
I am sure that you are as happy as me that a new Kentucky basketball season has started. This season I will continue my basic game tracking of efficiency and the related parameters discussed and presented in these pages. I invite you to comment and raise questions.
During the 2004-05 season, I will expand my data collection and analysis to address the use of the RPI Strength of Schedule factor as a normalizing parameter in making direct comparison of data for two teams. I began looking at this relationship during the last month of last season. Deductively, a relationship should exist between the relative strength of two teams' schedule and their respective efficiencies. The NCAA is attempting to make similar comparisons of teams in recognition the significance of strength of schedule.
I am looking at using teams offensive and defensive efficiencies together with each teams RPI Strength of Schedule values. In theory, the team with a stronger schedule will understate its relative strength with its average efficiencies. Similarly, the team with the weaker schedule will overstate its relative strength with its average efficiencies. The appropriate algorithm to describe this relationship remains unknown at this time.
I am also going to track the number of points scored by each team during bonus possessions due to offensive rebounding. Based on anecdotal evidence from a limited number of games last season, it appears that these second [and third, ...] chance possessions may be much more productive than any other type of possession.
Last night, Kentucky tested its new roster against Northern Kentucky University, a D II team for one of its two pre-season exhibition game. This is a departure from traditional pre-season competition, and I question whether Northern Kentucky or Kentucky Wesleyan are as strong as the competitive Nike Elite, EA Sports, California Allstars, and AIA teams.
Kentucky's debut performance against NKY is comparable to Kentucky's debut performances against the more traditional exhibition opponents. See the tabulation below:
However, there are some aspects of last night's game, when compared to the prior exhibition games, that are worth note. First, the offensive rebounding differential is the third lowest [2 here v. -4 in 02 and -5 in 00]. Second, the pace of the game  is slower than the average pace . Third, net game efficiency is third highest [0.183 this year v. 0.247 EA Sports in 04 and 0.294 Nike in 02].
If NKU is comparable to the traditional exhibition opponent, then this team posted a better than average start last night but could be more aggressive on the offensive boards, and could pick up the pace about 10 percent.
If NKU is a weaker opponent than the traditional exhibition opponent, then the weakness on the board could be an achilles heel for this team, and the optimistic forecast by Coach Smith that this team would play at a faster pace would appear unfounded.
Have the exhibition game performances provided a reliable indication of season long average performance? Probably not for net game efficiency, and perhaps for offensive rebounding difference. In the five previous seasons, the year end average net game efficiency was higher three times an lower twice. The average variance between the two is about 0.05 points per possession higher for the total season as compared to the exhibition games.
However, offensive rebounding in these exhibition games may indicate the strength of offensive rebounding for the season. In each of the previous five seasons, Kentucky's offensive rebounding was stronger against their exhibition game opponents than the corresponding season average. The average variance between exhibition and season play is about 4.5 offensive rebounds per game. This is an alarming warning about the upcoming season. If the average occurs, then a 2 offensive rebound difference against exhibition opponents would predict a -2.5 against all opponents for the season.
Therefore, let's watch the offensive rebounding ability of this team over the next 4 or 5 games, and revisit the question.
Copyright 2004 Richard Cheeks