Prior to the website being legitimately backed up to a server, we’ve had 2-3 different versions of this article. The purpose was to have a post where people can look back to when certain words and terminologies of advanced stats become a little bit hazy. This will be a permanent post and we’ll link it up in the navigation bar above the logo so you can bookmark it and pull it up anytime!

#### Effective Field Goal Percentage(eFG%) / Points Per Shot (PPS)

Or simply,

This is a way for people to compare players regardless of the type of shot they take. A great example I always like to use is this.

FGM | FGA | FG% | eFG% | |
---|---|---|---|---|

Player A | 6 | 12 | 50% | 50% |

Player B | 4 | 12 | 33% | 50% |

Player C | 5 | 12 | 41.6% | 50% |

As you can see by the antiquated definition of “field goal percentage,” Player A would actually be praised more — he made 50 percent of the shots he took. But all six of Player A’s makes are all two-point shots — which means he only scored 12 points while taking 12 shots (which translates to a points per shot of {1}). If we now consider Player B, who *only *made one-third of his attempts but did make four triples, the antiquated definition of “field goal percentage” actually has him as the *worst *of the group. But he actually scored 12 points on 12 shots (which translates to a points per shot of one {1}), which is the same as Player A’s! Player C on the other hand has mixed attempts — three 2-PT makes and two 3-PT makes. That also translates to a field goal percentage that’s still behind Player A’s 50% (it’s actually 41.6%). But again, he scored 12 points on 12 shots (which translates to a points per shot of one {1}).

**What is the relationship between eFG and PPS?** Well if you find the formula above a tad too confusing, eFG is just *half (1/2) *of the player’s PPS from field goal attempts (this is important). So if a player scores 20 points from 9 field goal attempts, that’s an eFG of 111%.

Note, eFG’s maximum number is *NOT *100% — it’s actually 150% (i.e. if a player makes all of his 3PT shots). Some people have a hard time grasping this concept and have actually suggested adjusting IT to make it easier to understand (i.e. 100% is the maximum). But that moves away from the base concept that eFG/PPS tries to correct — that is a 3PT make is 50% more valuable than a 2PT make (hence the +0.5 in the formula).

#### True Shooting Percentage (TS%)

We’ve considered the difference in value between a 3PT make and a 2PT make. However, those two aren’t the ones that contribute to a player’s “Points per Game” statistic. With True Shooting Percentage (TS%), we’ll try to adjust not just for 3PT shots but also for free throws.

Some people are inclined to think that a free throw is no more valuable than the shot it was born out of – if it was born out of a 2PT attempt, then its value is no more valuable than that. To a degree, it is correct. Making 50% of your 2PT shots is no more valuable than making 50% of your free throw shots. That’s assuming those free throws were born out of fouls on field goals that missed. But what about free throw attempts that were born out of makes (And-1)? or what about technical free throws? Flagrant free throws?

The value of a free throw therefore is greater than the value of attempts.

As an added trivia, if you remove the “free throw variable” in that TS% formula, you’ll come up with the eFG formula which is simply PPS/2. It’s awesome how it all ties together, right?

#### Points per Possession (PPP) / Points per Play (PPP)

If you’ve read enough NBA material, you’ll come across writers talking about a certain players “points per possession” on certain play types (or points per play). This is a new way of breaking down scoring to micro-level analysis. Simply put, it’s the total points scored by a player divided by the total possessions he used. In here, possessions is defined as *scoring possessions* or which is simply the addition of field goal attempts, turnovers and the possession equivalent of free throw attempts (which is estimated to be around . This estimate may not apply if you’re actually doing play-by-play analysis). So if for example a player scored 12 points on plays defined as “spotup” while taking 5 shots, drawing 1 foul and taking 2 free throws and turning the ball over 2 times, his PPP on that particular possession would be 12 /(5+1+2) = 1.4 PPP.

This can also be used in a broader sense as “possession” or loosely defined as each time a team has control of a possession. Without play-by-play, this is estimated as:

### Pace and Volume Adjusted Stats

These are the statistics that tries to put most box score statistics in the right context. Namely: rebounds, assists, steals, blocks and turnovers. In actuality, the points per possession statistic above is another pace adjusted statistic.

#### Rebounds Per Game versus Rebound Rate

The normal measure of how good a player is at getting a rebound is just sheer volume per game i.e. rebounds per game. But a better way is to adjust for the total amount of rebounding opportunities per game.

In order to understand this, you have to understand where rebounds come from. Rebounds come off misses. But remember, these misses aren’t exclusive to field goal misses – there are also terminal free throw misses, which is different from free throw misses (which is a combination of misses on first attempts and on second attempts and on third attempts, if any). That would be a very difficult job. What would be better is if we consider the fact that misses (whether field goal or from a terminal free throw) that your team didn’t rebound becomes a rebound to the other team. Which means defensive rebounds by a team are a subset of your team’s defensive rebounds and your opponent’s offensive rebounds or vice versa. Which means a team’s rebounding rate is simply:

To adjust for a player’s ORB%/DRB%, you just have to adjust for the amount of rebounding opportunities while he was on the court. Without play-by-play numbers, this is usually done by ratio and proportion. If he plays 24 minutes in a 48 minute game and there are 100 total rebounding opportunities, then there were approximately 50 rebounding opportunities when he was on the court.

#### Assists Per Game versus Assist Rate

In a similar fashion, a better way of accounting for assists is to adjust it to where assists are typically born out of — field goal attempts. Put simply, Assist Rate is calculated as:

If there were 70 field goal attempts and a team has 25 assists, this means that a team’s assist rate is simply 35%. Again, for individual Ast%, we merely adjust the “team field goal attempts” by ratio and proportion (or if available, by play-by-play).

#### Steals per Game versus Steal Rate

Likewise, steals are born out of possessions – specifically scoring possessions (field goals, turnovers or possessions from free throws). All of those three *could *have been turnovers (in fact, some of those turnovers are actually steals). Therefore,

Without play by play, the total number of scoring possessions can be estimated as,

#### Blocks per Game versus Block Rate

Blocks are a tricky statistic to record. Blocks are technically born out of field goal attempts. Which means if we wanted to know a player’s propensity to block shots, we merely adjust for the total field goals attempted while he was on the court. However, that would skew the results because most of the shot blocks occur on 2-point attempts — rarely do you see a 3-point attempt get blocked. Which is why, we remove the 3-point field goal attempt part and we’re left with Block Rate — a much better measurement of a player’s propensity to block shots (compared to Blocks Per Game).

#### Turnovers per Game versus Turnover Rate

Turnovers, like steals, are born out of “scoring possessions”. Unlike steals (which are born from an *opponent’s *scoring possessions), turnovers will be compared to a team’s scoring possessions. So to calculate a player’s turnover rate,

Without play by play, the total number of scoring possessions can be estimated as,

#### Usage Rate

The archaic box score doesn’t have a measure to compare this to. Basically, it’s a measure of how many *scoring *possessions a player used compared to a team’s total *scoring *possessions when he was on the court.

### All-in-One Statistics

#### Player Efficiency Rating

This statistic has its share of detractors. The main argument against it is that it doesn’t account for defense (even steals and blocks are counted only as a means to an offensive end). However, like most all-in-one statistics, it has its uses as a “snapshot” tool that allows us to see how the best players are in terms of impact.

The idea behind PER is to add all good contributions (field goals made, free throws made, rebounds, blocks, etc…) and subtract all bad contributions in the context of the value of a possession.

Just a simple example: if the league wide points per 100

#### Offensive Rating

This is a metric developed by one of the forefathers of advanced statistics, Dean Oliver. The process is a very tedious one so I won’t bother explaining that. It’s similar to points per possession in that it tries to contextualize points with possessions. The difference lies in the definition of points and possessions. For Dean Oliver, points are not just the points scored by the player but rather the points he also produces from offensive rebounds and from assists. To put simply, one possession sometimes means more than one *scoring *possession. What Offensive Rating tries to capture is how one player helps in the *overall* possession rather than on just *scoring *possession. Theoretically speaking, if a player never gets assisted on his shots, he never assists anyone and he never gets offensive rebound, then his Offensive Rating (or Points Produced per Possession) will be close to his Points Scored per Possession (which is the one we discussed above). That is never the case, however.

Personally, Offensive Rating is a pretty good indication of how good a player is at helping a team get points. The great thing about it is that you should be able to connect it with Team Offensive Ratings.

### Plus/Minus and Lineup Combination Statistics

Although this type of statistic has been ubiquitous in NBA blogosphere, it’s almost non-existent in the PH basketball blogosphere. That’s because this type of statistic actually requires play-by-play data — it’s a very meticulous statistic to collect.

Plus/Minus (and all its derivative forms) are calculated by considering how many points were scored or allowed while a player was on the court. Example, if Gabe Norwood had a +10, this means Rain or Shine outscored its opponent by 10 points when he was on the court (this can sometimes be on a per 100 possession basis). This doesn’t mean Gabe Norwood was the *sole *reason why ROS outscored its opponent by 10, but he was on the court when they did.

The problem with plus/minus is the same thing we’re trying to fix in the other metrics above — it lacks context. What this means is that, plus/minuses can deceiving without considering who that player is playing *with *and playing *against. *

**Enter Lineup Combinations.**

We’ve yet to introduce this formally in an article (hopefully soon) but the basic premise is that lineup combinations take into account who the player plays with. If San Mig Coffee allows 96 points per 100 when Yap is on the court without Pingris, allows just 89 points per 100 when Yap is on the court with Pingris and allows just 86 points per 100 when Pingris is on the court without Yap, then it *may (note: MAY) *mean that Pingris is a net positive on defense and Yap is a net negative on defense. You can expand this to include combinations of 3,4 or 5-man lineups. It still hasn’t considered who they’re playing *against* and this type of analysis gets tedious when you move past 2-man combinations (since there are 3003 combinations of 5-man teams in a 15-roster team).

To help with this, some smart guys introduced a few other plus/minus statistics such as APM (Adjusted Plus/Minus), RAPM (Regularized Adjusted Plus/Minus) and SPM (Statistical Plus/Minus) just to name a few. Those are beyond the scope of this article so I won’t go into great detail on them.

This will be updated as we introduce more statistics.

###### Note: Offensive Rating is simply points per 100 possession (whether that’s points produced [Dean Oliver] or points scored).