Kentucky and Louisville both had pretty darn good years in 2013-14. Kentucky played for the national title, and Louisville outscored teams by 20 points per game on average on their way to a #1 seed in the NCAA tournament, and a #1 rating at KenPom.com. Both teams outscored their opponents on average, but they did so using distinctly different playing styles.
Nearly every TV broadcast of a college basketball game starts with the announcers unveiling their wisdom in the form of “keys to the game”. These usually take the form of clichés like “take care of the ball”, “win the free throw battle”, and “get hot from 3”. I want to find out the truth…what were the keys to the game for UK and UofL in 2013-14?
To get to this, I’m going to look at every time a lineup played at least 5 possessions together in a game. For each lineup, I’m interested in their scoring margin per possession, and some underlying stats. I don’t just care about how UofL or UK did, however; I’m interested in how their performance compared to the opponents. If UK or UofL does better in one of these categories, does that ultimately mean they’re more likely to win the game? We’re about advanced basketball stats here, so we’ll use the following:
True Shooting Percentage (adjusts for extra value of 3 pointers and accounts for free throws)
FTA/FGA (a measure of how often teams draw free throws, which are a more efficient way to score)
Offensive/Defensive Rebounding Percentage (what % of available rebounds does the team get)
Assist Rate (what % of a team’s baskets assisted)
Turnover Rate (what % of a team’s possessions end in a turnover)
Foul Rate (what % of a team’s possessions have a foul drawn, or committed when on defense)
Block Rate (what % of a team’s shot attempts are blocked)
For the 2013-14 season, some of these stats weren’t very useful in determining whether UofL or UK outscored the other team. Let’s step back and consider this….
FTA/FGA: UK’s ability to draw more free throws than their opponent didn’t have much to do with them winning or losing. For example, UK outshot UConn on FT’s 24-10 in the NCAA title game, and outshot Florida 26-17 in the SEC final, but lost both. UofL’s free throw margin did matter a little bit, with an interesting twist…more on that in a bit.
Assist Rate: Everyone is taught to pass the ball, and that assists are good. However, the ability for either team to assist on more baskets than their opponent didn’t matter in determining their overall performance. UK lived on the offensive glass, and UofL had Russ Smith darting through the lane…and those don’t improve your assist total.
Foul Rate: At every level of basketball, fans complain when their team gets called for more fouls. However, this is often a sign of aggressiveness, and has a strong relationship with forcing turnovers. For UofL especially, drawing fouls tends to lead to forcing turnovers. Next time you hear someone complain about the foul count in a game, you’ll know that getting called for more fouls doesn’t lead to losing a game, necessarily.
Block Rate: Blocking a lot of shots looks awesome, but it can often lead to the shooting team getting the ball back. If you blocked every shot a team took, you’d win every game. But D-1 teams don’t block enough shots for it to really help determine the winner of a game.
So, the following stats were useful in predicting whether UK or UofL would outscore the other team:
|For UK:||For UofL:|
|True Shooting %||True Shooting %|
|Offensive Reb %||FTA/FGA|
|Defensive Reb %||Offensive Reb %|
|Turnover rate||Defensive Reb %|
For each stat, UK or UofL’s performance compared to the opponent is the key to predicting their scoring margin. For UK, the 4 stats above predict about 80% of the variance in the scoring margin per possession between lineups. For UofL, the 5 stats predict about 83% of the variance. The really cool stuff, however, is how significant the factors are for each team, and how that reflects each team’s playing style.
True Shooting % (TS%): This was the most important factor for both teams, although it was a little more important for UK to outshoot their opponents than UofL. This is driven by the fact that UK tends to have a worse turnover margin, so they have a few more wasted possessions per game. In order to outscore their opponent, UK had to be better at scoring on their turnover-free possessions. UofL had a bit more cushion, because they could generate a few extra possessions by winning the turnover battle. TS% adjusts for threes and free throws, so whether a team likes to bomb away or live in the paint, they need to score more efficiently than their opponent.
Turnover Rate (TO%): This was the 2nd most important factor for both teams, but it was much more significant for UofL than for UK. UofL plays a style that relies heavily on turnovers, so losing the turnover battle is a major sign that they’re going to lose the game. UK relied on strengths elsewhere, so losing the turnover margin wasn’t as big of a deal for them.
Offensive Rebounding % (OR%): The 3rd most important factor for both teams, but again, a big discrepancy between the significance for each team. This was much more crucial for UK, as they needed the ability to rebound their missed shots and generate extra scoring opportunities. UofL wasn’t as reliant on offensive boards, because they generated extra opportunities through turnovers.
Defensive Rebounding % (DR%): The 4th most important factor for both teams, but again, more important for UK than UofL. UK needed to be better than their opponents at denying extra scoring opportunities off rebounds, because they gave away opportunities elsewhere.
FTA/FGA: This is the one stat that was significant for UofL, but not for UK. UK drew a lot of free throws in nearly every game; they didn’t tend to lose game because their opponent shot more free throws. This stat was much less significant in predicting scoring margin for UofL, but I promised a twist, and here it is: when UofL shot FTs more often than their opponent, they tended to be outscored. Let that sink in. I would propose that this is another effect of UofL’s aggressiveness; if their opponent is shooting a lot of free throws, UofL is being aggressive and probably forcing turnovers. So, if you knew a UofL fan who complained about fouls being called on their team in 2013-14, please let them know that UofL tends to be a little more successful when they do send their opponents to the line.
So, know we’ve figured out what the real keys to the game are for both UK and UofL. UK depends on rebounding and making shots, while UofL needs to force turnovers and hit their shots. So, the next time an analyst preaches about their “keys to the game”, you’ll know what is really important.