Studying The Four Factors At State

In past seasons, I've done a Hubie Brown like analysis of the games of the state tournament.  In that analysis I looked at turnovers, free throws and shooting percentage.  Since it is the era of analytics, I've decided to advance that study.  One very similar way is to use The Four Factors of Dean Oliver.  If you watch the Timberwolves broadcasts, you'll see Jim Petersen reference these and you likely know what they are.  For those who aren't familiar here's what they are.

• Effective Field Goal Percentage
• Turnover Percentage
• Rebounding Percentage
• Free Throw Rate

One fundamental idea you'll see is that all of the values are percentages.  For the 3 items that aren't shooting %, that normalizes the data to take away differences in number of opportunities.  20 turnovers sounds bad but if you add the context of 100 possessions vs 50 possessions, it has far more meaning.  That's in contrast to my previous studies where I only looked at the final result without any manipulation.  For the detailed explanation of each factor and how its calculated, read this link.

So now that we have a reasonable tool, let's apply it to my subject, high school hoops.  No better time than the present with the state tournament just finishing up.  The data for the state tournament is easily found and I have reams and reams of it laying around.  That allows me to dig up the last 5 years of games at the boys state tournament.  That's 172 games of data with 43 games in each of the 4 classes.  I deem that a big enough sample size to provide reasonable value

Let's start with a chart of how winning teams did with respect to the factors

YEAR	EFG%	FT Rate	TO%	Reb%
2017	79.55%	77.27%	59.09%	68.18%
2016	90.63%	87.5%	65.63%	62.5%
2015	71.88%	78.13%	75%	68.75%
2014	84.38%	68.75%	65.63%	50%
2013	81.25%	62.5%	62.5%	59.38%
OVERALL	81.4%	75%	65.12%	62.21%

So when a team won a game in the last 5 years, 81.4% of the time they had the higher effective field goal percentage. This is in line with my previous simpler studies and something I've been saying for years.  If you shoot it better, odds are very very good you'll win.  Period.  Basket->Ball.  Put ball in basket.  In this case is just over 4 out of 5.  Don't overthink it, just make shots. 

So what happened in the other 18.6% of wins where the winning team lost the EFG% battle?  The winning team won the TO% battle 29 out of 32 times.  That's 90.63% of the time, a very insane number.  Here's how that EFG% result broke down per class.

Class A: 88:37%
Class AA: 81.4%
Class AAA: 79.07%
Class AAAA: 76.74%

Let's look at how many factors each winning team won

YEAR	1	2	3	4
2017	4.55%	25%	52.27%	18.18%
2016	0%	25%	43.75%	31.25%
2015	3.13%	25%	46.88%	25%
2014	3.13%	31.25%	59.38%	6.25%
2013	3.13%	37.5%	50%	9.38%
OVERALL	2.91%	28.49%	50.58%	18.02%

So 68.6% of the time a team won at least 3 or 4 of the factors.  More importantly, when a team won only 1 factor, they only won 2.91% of the time.  In those 5 wins, 2 of the wins were 20+ point wins thanks to winning the EFG% category.  The other 3 wins were by 3, 3 and 2 points and the winning team won the turnover battle in each.

Winning a 2nd factor wasn't a guarantee as winning 1 or 2 factors meant a win only 31.4% of the time.  That's less than 1 out of 3.  There are 6 combinations of winning 2 factors.  But the combination of EFG% and FT Rate accounted for 34.69% of those alone, double the amount of an even distribution.    The average margin of victory was 9.9 PPG with almost 60% of the games being single digit finishes.

When a team won 3 factors, effective field goal percentage was included 87.36% of the time.    The margin of victory expanded to 14.8 PPG.  The number of single digit games decreased to 37.9%

When a team won 4 factors, a blowout was in order as expected.  The average margin of victory was 21.45 PPG.  Only 12.9% of those games were single digit finishes.  That exactly matches the number of games that had a margin of 44 points or more.

Here's a breakdown of the number of factors won across classes.  Percentages are percent of the column total.

# Factors Won	A	AA	AAA	AAAA
1	2.33%	0%	2.33%	6.98%
2	13.95%	37.21%	25.58%	37.21%
3	55.81%	48.84%	55.81%	41.86%
4	27.91%	13.95%	16.28%	13.95%

83.72% of the games in class A were 3 or 4 factor games.  So in single A, odds are very high that you're going to see a game that ends up lopsided.  On the other end of the spectrum was class AAAA with 55.81% of the games seeing a team win 3 or 4 of the factors.

Here's how margin of victory related to the 4 factors in each class.  The values in the table are the average margin of victory.

# Factors Won	A	AA	AAA	AAAA
1	2	N/A	3	16
2	11.33	9.25	13.27	9.9
3	15.75	13	14.75	14.83
4	25.25	19.33	22.29	21.45

The class A 4 factors won number jumps up because of Minneapolis North's dominance the last 2 seasons.  Interesting that 2 of the 5 wins were blowout by doing nothing more than shooting better.  To win only 1 factor and steal a close game makes sense but 40% of the outcomes being blowouts is interesting.  Granted that's a very small sample size of 5 games.  Otherwise its what you expect, the more factors you win, the bigger you win.

Overall conclusions.  Shooting is big, rebounding can be overcome.  Dominate the factors, dominate the game.