BrewCity...

[IrwinFletcher]

I read your response to Phantom's post on the Top 15 players and have a request for you.

I am not up to speed on some of the statistical analysis that you mentioned and that KenPom utilizes in his rankings. I do not subscribe to his site, likely due to not understanding everything that his site is about.

Would you be able to give some insight into what some of these stats measure and what they mean to evaluate a player/team? I think I would find it interesting and better help me understand some things. Much like baseball, basketball is turning into something that can be measured by some advanced numbers.

[Alan Bykowski, "brewcity77"]

Sure. His site has a lot of primer articles, but I'll try to sum up some of the key points. First, efficient field goal percentage, or eFG%, which is the absolute king of basketball metrics.

eFG% is calculated like this: (2-point field goals made) + 1.5x(3-point field goals made) / field goal attempts. So if a player goes 1/7 from inside the arc and 4/6 from three, his FG% would only be 38.7%, but his eFG% would be 53.8%. Generally, the benchmark for eFG% is 50%. Anything over that is usually going to be a positive, below is where you start questioning shot selection.

The math:
FG% 5/13 = 38.7%
eFG% (1 x 1) + (1.5 x 4) / 13 = 53.8%

eFG% gives a big boost to threes, and deservedly so. If you flip those numbers on the statline above (4/7 inside, 1/6 outside) his eFG% would be 42.3%. Essentially, it tells you how many points the player is scoring per shot attempt. The first example results in 15 points on 13 shots, while the latter is 11 points.

The math:
FG%: 5/13 = 38.7%
eFG%: (4 x 1) + (1.5 x 1) / 13 = 42.3%

On a game by game basis, there is no single statistic that is more indicative of winning and losing than the winner of the eFG% statistic. Yes, it can be offset with a turnovers, offensive rebounds (both of which give more attempts), or free throws, but in terms of predicting a winner, eFG% is king. The four factors (eFG%, turnovers, rebounding, free throws) were based on a paper written by Dean Oliver years ago, and he believed eFG% was 40% responsible for winning, but this Cracked Sidewalks article indicates eFG% is actually 63% responsible for predicting the outcome, making it more important than the other three factors combined.

[Alan Bykowski, "brewcity77"]

Pomeroy's numbers look at things on a tempo free basis. Instead of looking at quantity, it looks at quality. Why? Because the fastest paced team (The Citadel, 84.5 possessions/game) plays more than 25 possessions additional on a given night than the slowest paced team (Virginia, 59.3 possessions/game). Between those two, The Citadel averages over 90 points per game while Virginia averages under 70. However, Virginia is much better in terms of scoring per possession, so even though they are outscored by their faster opponents by more than 20 points/night, Virginia actually has the far better offense.

As far as the other three factors, here's a quick breakdown:

Turnover Percentage: Typically, broadcasters and pundits like to talk about things like turnovers per game. At the end of the day, what does that really mean? Not a lot because teams play at such disparate paces. Let's consider Wisconsin and Marquette. Right now, UW averages 11.1 turnovers per game while Marquette averages 12.2. At a glance, you'd say Wisconsin is better at protecting the ball. However, Wisconsin games average 64.3 possessions per game and Marquette averages 75.5 possessions. Using those numbers, it turns out Wisconsin turns it over on 17.3% of their possessions while Marquette turns it over on 16.2% of their possessions. When you take the pace out of the game, Wisconsin is 1.1% more likely to turn the ball over than Marquette is on any given possession. Both teams are good at protecting the ball, but the basic numbers tell you Wisconsin is better when the reverse is actually true.

Offensive Rebounds: This looks at percentages. Instead of looking at how many rebounds per game a team gets, it looks at what percentage they get. Marquette isn't very good on the offensive glass, getting to 26.7% of their own misses (257th in the country). Defensively, they are much better, only allowing their opponents to get to 25.1% (41st in the country). For reference, the best teams in the country get to over 40% of their own offensive boards while the best defensive teams limit teams to around 20% or below (and the converse is about the same).

Free Throws: This one is a bit tricky, because it looks at the quantity of free throws a team gets compared to the number of field goals they attempt, not percentage. Why? Because individual and team free throw percentages are virtually meaningless stats. That's a difficult concept for old-school statniks, so it bears repeating. Free throw percentage is virtually meaningless. I'll get back to that in another post, but what this looks at is how often a team gets to the line, again in terms of percentage. For Marquette, the number of free throws we take is 31.6% of the number of field goals we take. That's not a good number, 251st in the country. However, we still have a wildly efficient offense (17th nationally) because this really isn't that important of a metric. In terms of numbers, if you take 30 shots from the field and get to the line 15 times, it would be a 50% ratio. The best teams in the country get to the line at around a 50% rate while the worst are 25% or below.

[Alan Bykowski, "brewcity77"]

Why free throw percentage doesn't matter

Relatively speaking, free throw percentage is not important. Why? Let's break it down by the numbers.

Marquette averages 19.3 free throw attempts per game and is 5th in the nation at 81.3% from the line. So in an average game, Marquette gets 15.7 points on 19.3 free throws. That's pretty good, but consider if they were even better. Say we were we really good, at 91.3%. Generally, 10 percentage points is about a 125-150 point ranking difference. If Marquette shot 91.3% at the line, they would average 17.3 points on 19.3 free throws. Ten percentage points and it only makes a difference of 1.6 points per game. If we were 20% worse, we'd be at 11.8 ppg from the line. That's a difference of 5.5 points. Those are drastic swings in free throw percentage, but only make a difference of a handful of points per game.

The math:
19.3 x .813 = 15.7
19.3 x .913 = 17.6
19.3 x .613 = 11.8

Now consider what a difference eFG% makes. Marquette averages 61.1 field goal attempts per game and is 10th at 57.9 eFG%. So each field goal MU attempts is worth 1.158 points (percentage doubled because of the baseline 2 points per made basket). In an average game, Marquette gets 70.8 points on field goals. What if they could increase their eFG% by 2%? Bring it to 59.9 and you get FG attempts worth 1.198 points. At the same number of attempts per game, we would average 73.2 points on field goals. Marquette would get a 2.4 point benefit per game. 2% of eFG% is worth 150% what a 10% increase in FT% would be worth. To match the downward shift of 20%, it would take an eFG% drop of 4.45%.

The math:
61.1 x 1.158 = 70.8 (Marquette FGA/game x 2(eFG%)
61.1 x 1.198 = 73.2 (Marquette FGA/game x 2(eFG% + 2%)
61.1 x 1.069 = 65.3 (Marquette FGA/game x 2(eFG% - 4.45%)

Bottom line, eFG% is king, and free throw percentage is virtually meaningless. Huge swings in free throw percentage are worth a handful of points, while small swings in eFG% are worth significantly more. This is because of the sheer volume of field goals compared to free throws (double even for the best teams) and the increased value you will get taking two or three point shots as opposed to one one efforts.

To expand on that, efficiency is measured by points per possession. Generally, the average team scores 1.00 point per possession. It wavers marginally season to season, but by and large, that's the standard. When you go to the line, a team that makes 65% of their free throws is scoring 1.30 ppp when they get two attempts at the line. That would be better than the 2014 Wisconsin Badgers, which was the most efficient offense of the last 16 years. So essentially, getting to the line even for a bad free throw shooting team is going to be incredibly efficient. Bad at free throws still means good at offense when you get to the line.

[Nukem2]

Still good to make the most on your FTs however.

[Alan Bykowski, "brewcity77"]

A few other statistical meanings:

Adjusted Tempo: This is the number of possessions a team plays per game. Fast teams play in the 75+ range, slow teams play in the 65- range. Doesn't tell you whether the team has a good offense or defense, but does tell you their stylistic tendencies as far as if they want to score on the fast break or out of a half-court set, as well as the difference of teams that pressure the opposition into turnovers (like a WVU) or simply play suffocating man defense and try to force shot clock violations (like Wisconsin).

3P%/2P%/FT%: On Pomeroy's site, these are just the raw metrics. Primarily useful for calculating eFG% and offensive efficiency.

Point Distribution: This tells you where a team's points come from. Do they score most of their points on the three? From inside the arc? Again underscoring how meaningless free throw percentage is, only one team has ever gained over 30% of their points at the line. The 2014 North Carolina A&T Aggies were 332nd in offensive efficiency that year. Good teams focus on being efficient from the field, not the line.

[Alan Bykowski, "brewcity77"]

Still good to make the most on your FTs however.

You'd rather make them than miss them, but at the end of the day, there is very little difference between being a terrible free throw shooting team and an excellent one. It's virtually inconsequential. I know that goes against logic, and in a tight game you'd rather have a good shooter at the line than a poor one, but in terms of how good your offense will be and your likelihood of winning, free throw percentage isn't all that important.

Far more important is how often you get to the line. Because it's an efficient manner of scoring (whether you are a good or bad free throw shooting team) getting to the line more often will help you win. But compared to eFG%, it's nothing.

[TheSultan]

eFG% does for basketball what slugging percentage does for baseball. Instead of treating every basket (hit) the same, it weights those baskets (hits) that are more valuable.

Example. Traci Carter has a FG% of .565. Markus is .484. So Carter is the better shooter right?

However if you use eFG%, Carter is .630 and Markus is .648. It's because Markus' ability to shoot from deep is more valuable.

[Nukem2]

In the end, ya gotta make your shots. Hard to teach eFG%. That's just a resultant stat. Practice, hard work and coaching.

[TheSultan]

In the end, ya gotta make your shots. Hard to teach eFG%. That's just a resultant stat. Practice, hard work and coaching.

But it is a valuable statistic that can help you judge player v. player and how a basketball team should be put together. You can especially see this in the NBA where 3 point shooting is emphasized more than it ever has.