Luke's Minute Analysis

DATA

I created (with the assistance of Shane) a series of metrics to show who has ran the most balanced distribution of playing time this year. For each of the three metrics, I also calculate the average across all 3 teams, as well as the theoretical maximum a team could achieve via our stated rules. I’ll go through each column briefly. All data used is from the final 5 weeks of the season, when rosters were officially settled so we can compare apples to apples without needing to account for trades. Green rows represent teams in the playoffs.

“Range” is simply the difference between the minutes of the players who played the most and the least. Lower is better. The theoretical maximum range is 15 (if one player was set at 35 and another at 20), but that is scaled down to 12.5 minutes since we play 40 minute games instead of 48 minute games. This number could be exceeded further if a team has ghosts.

“Gini” refers to the Gini coefficient; it is often used to quantify income inequality by economists, but can be used simply to reflect inequality in a series of numbers overall. Lower is better. Instead of comparing only the first and last values, every single value is compared. The theoretical maximum (if you assigned 7 players to play 20 minutes, 1 player to play 30, and 2 players to play 35 minutes) would be .121. This does not need to be scaled as it is a proportional answer. This could also be exceeded if a team was playing ghost minutes to players.

“Owner” reflects how many minutes the owner of a team played relative to their teammates. Higher is better. 1 means they played the more than all of their other teammates, 10 means they played the least. The theoretical maximum (or in this case minimum) was therefore 1.

Then, I went through and color-coded each team’s results. If a team had a result that was equal-to-or-worse than the average, I put it in orange. If a team had a result that was somehow equal-to-or-worse than the theoretical maximum, I put that in red.

“Rank” is my attempt to quantify these three metrics. For every orange metric a team had, they received one penalty point. For every red metric, two points. Lower is better. Lastly, I sorted the table by this rank, and sorted by Gini coefficient as the tiebreaker between teams with the same ranking. New York, Atlantic City, and Ohio were the only teams above-average in all three metrics. San Antonio and Oakland were the only two below-average in all three metrics, with Oakland being the only one at or beyond the theoretical maximum in all three.

ANALYSIS

The easiest conclusion is that the five most unequal teams all missed the playoffs, so therefore being unequal is a bad move tactically. Yet there is a chicken and egg problem. Did the teams really struggle because they played their top guys a ton of minutes? Or were they struggling already, and needing to ride their stars to have even a chance to remain competitive? Perhaps a third culprit (a young bench, multiple ghosts) explains both to an extent.

There are a lot of asterisks you can attach to this analysis, and every team probably thinks they are very fair and equitable. Anyone can claim “well XYZ gets takeover a lot so he stays in longer” or “well I had him only playing 5 minutes but the CPU didn’t sub him out for 7 minutes this game, that’s not on me, I keep it balanced.” However, this analysis compares to the rest of the league, so those issues should effect everyone equally; if you’re a team at the top or bottom, it’s not because of takeover, I guarantee it.

I’d imagine each team has their own story to tell. As the owner of Baltimore, I take personal pride in being the only owner who is dead last on their team in minutes, preferring to give them more to the rest of my players. Coach Tim Riggins, who sets our rotations, is also only 7th on our team in minutes, and still somehow our leading scorer. In a similar vein, honorable mention goes to co-owners like Love Lockdown (10th on their team, but not on this list since Hoe Jarris at 4th was higher) and Werbadgamers (who does not have a player at all, but has Ingram as a co-owner who is 2nd on the team).

Seattle should take pride in their distribution as well, with by far the tightest range and Gini coefficient. Their only knock is that Steele Stern’s is “first” on the team in minutes. Yet at only 22.6 minutes per game, he would be 6th on other teams like Michigan who play all their other starters more.

There is an easily-visible trend amongst some teams to have 5 set players getting heavy minutes (presumably their starters), and 5 set backups getting very few. Oakland as the least equal team is a case study in this: Stank, Klank, Colakovic, and Glisack and Storm all get 24-27 MPG. Stevens, Bourne, Nuff, Clarke, and Jones all get between 10-17 MPG. No one on the team received between 17-24 MPG. Other teams following this mold (to a lesser degree) include New Orleans and Kentucky.  San Antonio, Michigan, Atlantic City, and Virginia all do it to an extent as well, but not with quite as defined of a 5-5 split. For anyone who wishes to see the data yourself, here is a screenshot of every team’s minutes distribution, with the range included (thanks to Shane for fully putting this part together by himself).

Stay tuned for more statistical analysis from me on non-rotation questions later on in the year as we progress through the playoffs!