Reviewed October 2026 · football, hockey, basketball, baseball
21 stats explained in plain English: what each one means, how to read it, and where it appears on the site. Football, hockey, basketball and baseball, with the numbers that separate good from bad.
Quick reference: average, good and poor
Typical ranges for a full season. Click a stat for what it means.
What it meansEvery spot on the field has an expected point value: the points a team typically scores from that down, distance and yard line. EPA is how much one play changed that number. A 20-yard gain on third-and-8 adds a lot. A sack on first down takes points away.
How to read it
EPA per play is close to 0.0 for the league as a whole.
Offense: above +0.10 is a good offense, above +0.20 is elite, below −0.10 is poor.
Defense: lower is better, so −0.10 or lower is a strong defense and anything above +0.10 is a weak one.
Sample matters: through six or seven games, one big play can move a team 0.03 or more.
Success rateShare of plays that keep the offense on schedule
What it meansA play is a success if it gains at least 40% of the yards needed on first down, 60% on second down, or all of them on third or fourth down. Success rate is the share of plays that clear that bar. It rewards steady, repeatable offense and ignores one-off big plays.
How to read it
The NFL average is around 44% to 46%.
Offense: 48% or higher is strong, under 41% is poor.
Defense: lower is better, because the opponent is succeeding less often.
Read it next to EPA: a high EPA with a low success rate usually means a team lives on big plays.
What it meansPROE compares how often a team actually throws to how often a typical team would throw in the same situation: the down, the distance, the score, the clock and the spot on the field. It separates a pass-happy coach from a team that is simply trailing.
How to read it
0 means a team calls plays like a typical team.
+5 percentage points or more is a pass-heavy approach, −5 or lower is run-heavy.
Good for projecting volume: a high PROE offense gives its quarterback and receivers more chances, whatever the score.
Use a window of several games; one blowout can swing a single week.
What it meansCPOE takes a quarterback's completion percentage and subtracts the completion rate a typical passer would have on the same throws, given how far and where the ball went and how much pressure there was. It credits accuracy on hard throws and does not reward dump-offs.
How to read it
0 is an average passer.
+2 or better is good, +4 or better is elite.
−2 or worse is a quarterback leaving completions on the field.
It needs volume: a few hundred dropbacks before it settles down.
What it meansEvery shot has a chance of scoring based on where it came from and what kind of shot it was. GSAx adds up the goals an average goalie would have allowed on the shots a goalie faced, then subtracts the goals he actually allowed. Save percentage treats every shot the same. GSAx does not.
How to read it
0 is average. Positive means the goalie stopped more than expected.
Over a full season, +10 is a good starter and +20 is elite. −10 or worse is a goalie who is hurting his team.
Early in a season, swings of five goals either way are mostly noise. Check the number of starts.
GSAx per 60 minutes compares goalies who play different workloads.
What it meansExpected goals (xG) give every unblocked shot attempt a value equal to its chance of becoming a goal. xGF% is a team's share of all the expected goals in its games. It measures who is creating the better chances, whether or not the puck went in.
How to read it
50% is even.
54% or higher over a full season is an elite team, 52% to 54% is good.
46% or lower is a team getting outchanced badly.
Usually quoted at 5-on-5, so power plays do not hide a weak even-strength team.
What it meansCorsi counts every shot attempt: shots on goal, misses and blocked shots. CF% is a team's attempts divided by all attempts in its games. It is a quick read on who has the puck and who is shooting more.
How to read it
50% is even.
52% or higher is strong, 48% or lower is weak.
Because every attempt counts the same, Corsi can be fooled by teams that fire a lot of low-quality shots. Pair it with xGF%.
What it meansFenwick is Corsi without the blocked shots: shots on goal plus misses. The idea is that a blocked shot is a defensive play, so removing it gives a cleaner picture of how many real chances a team creates. FF% is a team's unblocked attempts divided by all unblocked attempts.
How to read it
50% is even, and the same ranges as Corsi apply: 52% or higher is strong, 48% or lower is weak.
A big gap between a team's Corsi and Fenwick means it blocks (or gets blocked) more than most.
What it meansPDO adds a team's shooting percentage and its save percentage at 5-on-5, usually written so that average is 100 (or 1000). Over long stretches almost every team ends up close to 100, so a PDO far from it is a sign of good or bad luck rather than good or bad play.
How to read it
100 is average.
Above 102 usually means a team is running hot: pucks are going in and saves are being made at an unsustainable clip.
Below 98 usually means the opposite.
Over a full season, 99 to 101 is normal. Early in a season, a PDO swing of five points is common.
What it meansHigh-danger chances are the shot attempts with the highest chance of scoring, the ones from right in front of the net. HDCF% is a team's share of them. Goalies are also judged on their save rate against these chances.
How to read it
50% is even for HDCF%.
53% or higher over a full season is a team that owns the front of the net, 47% or lower is a team that concedes it.
For goalies, a high-danger save percentage in the low .800s is about average and above .850 is excellent.
High-danger counts are small, so they swing more than shot totals.
What it meansNet rating is points scored minus points allowed per 100 possessions. Counting per possession rather than per game strips out fast and slow styles, so a run-and-gun team and a slow team can be compared on equal terms.
How to read it
0 is an average team.
+5 or better is a contender, +8 or better is elite.
−5 or worse is a team that loses most nights.
Over a full NBA season net rating tracks wins better than points per game.
What it meansOffensive rating (ORtg) is points scored per 100 possessions. Defensive rating (DRtg) is points allowed per 100 possessions. The gap between them is net rating.
How to read it
The league average moves a little each season and sits around 114 to 116 in recent NBA seasons. In college basketball it is lower, near the low 100s.
Offense: higher is better. Defense: lower is better.
A team 5 points above average on offense or below average on defense is a top-five unit.
Check the league average first: a 115 offense is average now and was elite a decade ago.
What it meansIn basketball, pace is the number of possessions a team uses per game, usually per 48 minutes in the NBA and per 40 in college. In hockey, our Labs show pace as the shot attempts by both teams per 60 minutes. It describes a style, not quality.
How to read it
NBA pace has been around 98 to 101 possessions in recent seasons, so 103 or more is fast and 96 or less is slow.
Pace moves totals: two fast teams play a game with more possessions, and more possessions mean more points.
Use per-possession ratings, not points per game, to judge a team that plays fast or slow.
What it meanswOBA gives every way of reaching base credit in proportion to how much it helps score runs, so a home run counts for more than a double and a double for more than a walk. It is scaled to look like on-base percentage.
How to read it
The league average is about .315 to .320.
.340 is above average, .370 or higher is star level, and .300 or lower is poor.
For a pitcher, wOBA allowed works the same way and lower is better.
xwOBA is the same idea built from exit speed and launch angle, so it shows what the contact deserved.
What it meansFIP estimates a pitcher's ERA from the things he controls: strikeouts, walks, hit batters and home runs. It ignores balls in play, because defense and luck have a big say in whether those become hits. It is put on the ERA scale so the numbers look familiar.
How to read it
The league average is around 4.00, close to league ERA.
Under 3.50 is a very good pitcher, above 4.50 is a struggling one.
A pitcher whose ERA is well below his FIP has been getting help and often gives some back.
What it meansxFIP is FIP with one change: a pitcher's home runs are replaced by the number a typical pitcher would allow on the same fly balls. Home run rates bounce around from year to year, so xFIP is a steadier read on talent than FIP.
How to read it
Same scale as ERA: around 4.00 is average, under 3.50 is very good, above 4.50 is poor.
When xFIP is much lower than FIP, a pitcher has been giving up more home runs than his fly balls usually produce.
What it meansSIERA is another ERA estimator. It uses strikeouts, walks and the kind of contact a pitcher allows (ground balls versus fly balls), and it lets those pieces interact. A strikeout pitcher benefits more from his walk rate than a contact pitcher does.
How to read it
Same ERA scale: around 4.00 is average, under 3.50 is very good.
It was designed to predict future ERA better than FIP does.
We show FIP and xFIP in the MLB Lab; SIERA sits close to xFIP for most pitchers.
What it meansExpected wins estimate how many games a team should have won from the points (goals, runs) it scored and allowed. A team that scores far more than it allows should win most of its games, and the formula turns that margin into a win percentage.
How to read it
Compare it with the real record. A gap of one win or less is normal.
A team well above its expected wins has been winning close games, which is a weak guide to what comes next.
A team well below has been losing close ones and is often better than its record.
In the NFL a 1.5-win gap after seven games is large. In baseball and basketball it takes more games to mean the same thing.
What it meansOn our ratings tables, luck is a team's actual win percentage minus its expected win percentage. Positive luck means a team has won more games than its scoring suggests. Negative means fewer.
How to read it
0.0 means the record matches the scoring.
Plus or minus 0.10 (ten percentage points) is a clear signal; beyond 0.20 is rare over a full season.
It does not predict by itself. It tells you how much of a record to trust.
What it meansA power rating states how many points better or worse a team is than an average team on a neutral field. The difference between two teams' ratings, plus a home-field bump, is a quick fair spread to compare with the market.
How to read it
0 is an average team. Positive is better.
The scale is the sport's own: points in the NFL and college football, goals in the NHL, runs in MLB.
A 3-point rating gap is about a field goal in the NFL.
Use it to find games where the market and the ratings disagree, not as a pick on its own.
What it meansA team's ATS record counts how often it covered the point spread, written wins-losses-pushes. A team can win the game and still lose against the spread if it wins by fewer points than the line.
How to read it
At the standard −110 price you need to win 52.4% of bets to break even.
50% is a coin flip, and over a full season most teams land between 45% and 55%.
A team's ATS record is mostly noise by itself. A 5-1 start is not a trend.
Look at it with the team's rating: a team covering a lot while rated poorly is often being priced too low.
For an NFL offense, above +0.10 per play is good and above +0.20 is elite. For a defense, lower is better, so −0.10 or lower is strong. The league as a whole is close to 0.0.
What is a good PDO in hockey?
100 is average. Over a full season most teams finish between 99 and 101. A team above 102 or below 98 is usually running hot or cold, and its results tend to drift back toward 100 as games pile up.
What is a good GSAx for an NHL goalie?
GSAx (goals saved above expected) of 0 is average. Over a full season, +10 is a good starter and +20 is elite, while −10 or worse is a struggling goalie. Early in the season, a swing of five goals either way is mostly noise.
What is the difference between Corsi and Fenwick?
Corsi counts every shot attempt: shots on goal, misses and blocked shots. Fenwick leaves out blocked shots. Both are shown as a team's share of all attempts, with 50% as even.
What is the difference between FIP and xFIP?
Both estimate a pitcher's ERA from strikeouts, walks and home runs. FIP uses the home runs he actually allowed. xFIP replaces them with the number a typical pitcher would allow on the same fly balls, which makes it steadier from year to year.
What does it mean when a team's record beats its expected wins?
It means the team has won more games than its scoring margin suggests, usually by winning close games. That kind of record is a weak guide to future results, so it is worth checking a team's rating before trusting a hot start.